Cerebrospinal fluid lipoprotein-mediated cholesterol delivery to neurons is impaired in Alzheimer's disease and involves APOE4

Abstract

In the central nervous system, apolipoprotein (APO)E-containing lipoprotein particles mediate the transport of glial-derived cholesterol to neurons, which is essential for neuronal membrane remodeling and maintenance of the myelin sheath. We aimed to examine cholesterol transport via lipoprotein particles in cerebrospinal fluid (CSF) of Alzheimer's disease (AD) patients compared to control individuals. Additionally, we explored the ability of reconstituted HDL containing different APOE isoforms to regulate cholesterol transport. We evaluated the capacity of CSF lipoprotein particles to facilitate radiolabeled unesterified cholesterol efflux from A172 human glioblastoma astrocytes and to deliver cholesterol to SH-SY5Y human neuronal cells. The CSF lipoprotein proteome was analyzed by LC-MS/MS. Reconstituted HDL nanoparticles were prepared by combining phospholipids and cholesterol with human APOE3 or APOE4, followed by radiolabeling with unesterified cholesterol. Our results showed that cholesterol efflux from astrocytes to CSF were similar between AD patients and controls, both under baseline conditions and after activation of ABCA1 and ABCG1. However, CSF lipoprotein-mediated neuronal cholesterol uptake was significantly reduced in the AD group. LC-MS/MS analysis identified 239 proteins associated with CSF lipoproteins in both groups, with no major alterations in proteins linked to cholesterol metabolism. However, 27 proteins involved in noncholesterol-related processes were differentially expressed. Notably, synthetic reconstituted HDL particles containing APOE4 exhibited reduced capacity to deliver cholesterol to neurons compared to those with APOE3. These findings indicate that CSF lipoproteins from patients with AD demonstrate impaired cholesterol delivery to neurons. Our study highlights APOE4 as a critical contributor to abnormal neuronal cholesterol uptake in AD pathophysiology.

Keywords: APOE; Alzheimer's disease; CSF lipoprotein; cholesterol.

Fig. 1

Astrocyte cholesterol efflux to CSF remains similar in AD and control groups, whereas CSF HDL-like-mediated cholesterol delivery to neurons is impaired in AD. (A) Astrocyte cholesterol efflux assay. Human glioblastoma astrocytes were cultured for 24 h, followed by a 48-h incubation with radiolabeled cholesterol. Cells were then treated for 18 h with or without T0901317 to activate ABCA1/G1 pathways. Serum-free medium containing CSF was added for 4 h. Both the medium and cell fractions were processed to quantify radiolabeled cholesterol. (B) Astrocyte cholesterol efflux results: Left panel-Cholesterol efflux from human glioblastoma astrocytes to CSF (30% v/v) is shown for both control and AD samples, under baseline conditions and following T0901317 pretreatment. Right panel-specific ABCA1/G1-dependent cholesterol efflux was calculated subtracting baseline levels from those observed in ABCA1/G1-expressing cells. (C) CSF HDL-like-mediated cholesterol uptake assay: Human neuroblastoma cells were seeded and differentiated into neurons in a low-serum medium containing retinoic acid. After 24 h, radiolabeled CSF HDL-like particles containing unesterified cholesterol were added (10% v/v). The cells were incubated with serum-free medium containing CSF for 4 h, and both the medium and cell lysates were processed for radiolabeled cholesterol quantification. (D) Cholesterol uptake results: CSF HDL-like-mediated cholesterol uptake was measured in human neurons exposed to CSF from both control individuals and patients with AD. Among AD subjects, APOE4 carriers are indicated with a different color. (E and F) Influence of Tau and Aβ_1-42 on neuronal cholesterol uptake: Cholesterol uptake in SH-SY5Y neurons mediated by control CSF HDL-like particles was assessed in the presence of Tau or Aβ_1-42 added to the culture media at concentrations up to 1,000 and 1,500 pg/ml, respectively. Values are shown as the mean ± SD for 10 subjects per group in panels (B) and (D) The Student’s t test was used to compare the CSF HDL-like-mediated cholesterol uptake by neurons, as well as the astrocyte cholesterol efflux under various conditions between the control and AD groups. One-way ANOVA with a post test for linear trend was used in panels (E) and (F). Three separate experiments were carried out for each condition. Aβ, amyloid beta; AD, Alzheimer’s disease; APO, apolipoprotein; CSF, cerebrospinal fluid.

Fig. 2

Altered proteome of CSF HDL-like particles in AD. (A) Native Gel Electrophoresis of CSF samples. Representative native polyacrylamide gel electrophoresis of CSF from patients with AD and control individuals. Left panel-Coomassie blue staining showing lipoprotein and albumin bands. Right panel-corresponding nitrocellulose blot probed with a polyclonal anti-APOE antibody, identifying the APOE-containing band using nondenaturing polyacrylamide gradient gel electrophoresis. (B) Native gel electrophoresis of isolated serum HDL, CSF samples, lipid-free APOA1, lipid-free APOE3 and lipid-free APOE4. (C) Proteins involved in cholesterol metabolism quantified in the APOE-containing lipoprotein band from CSF. Proteins with increased (red) or decreased (blue) abundance in AD versus control are shown according to the indicated Zq score scale. APOM: apolipoprotein M; APOA1: apolipoprotein A-I; APOD: apolipoprotein D; LRP1: low-density lipoprotein receptor-related protein 1; PLTP: phospholipid transfer protein; PON1: serum paraoxonase/arylesterase 1; APOH: Beta-2-glycoprotein 1; APOE: apolipoprotein E; APOA2: apolipoprotein A-II. (D) Differentially regulated proteins in HDL-like particles from the CSF of patients with AD. Heat map depicting significant protein abundance changes (P < 0.05, proteins quantified with more than one single peptide) in AD and control groups (10 independent samples per group). Increased (red) or decreased (blue) abundances are shown according to the indicated Zq scale. Differential protein expression analysis was performed using moderated t-statistics (limma test). ALDOA: fructose-bisphosphate aldolase A; ACHL1: neural cell adhesion molecule L1-like protein; ATRN: attractin; LTBP4: latent-transforming growth factor beta-binding protein 4; NrCAM: neuronal cell adhesion molecule; IGLC2: immunoglobulin lambda constant 2; GRIA4: glutamate receptor 4; SEZ6: seizure protein 6 homolog; IGLL5: immunoglobulin lambda-like polypeptide 5; C5: complement C5; GELS: gelsolin; B4GAT1: beta-1,4-glucoronyltransferase 1; CLEC11A: C-type lectin domain family 11 member A; FUCA2: plasma alpha-L-fucosidase; MCAM: cell surface glycoprotein MUC18; HPX: hemopexin; C8B: complement component C8 beta chain; ENPP2: ectonucleotide pyrophosphatase/phosphodiesterase family member 2; AGT: angiotensinogen; PLXDC2: plexin domain-containing protein 2; BCAM: basal cell adhesion molecule; LUM: lumican; OLFML3: isoform 2 of olfactomedin-like protein 3; VCAM1: vascular cell adhesion protein 1; CADM4: cell adhesion molecule 4; CTSD: cathepsin D; FBLN1: fibulin-1; AD, Alzheimer’s disease; APO, apolipoprotein; CSF, cerebrospinal fluid.

Fig. 3

Synthetic rHDL nanoparticles containing APOE4 exhibit impaired cholesterol delivery to neurons. (A) Schematic representation of the synthesis of rHDL-APOE nanoparticles: recombinant APOE3 or APOE4 were combined with DMPC and cholesterol in a molar ratio of 59:7:1. The mixture underwent three cycles of vortexing and temperature modulation, alternating between 37°C and 4°C, to optimize APOE-lipid interactions. (B) Native gel electrophoresis of synthetic rHDL-APOE3 and HDL-APOE4 nanoparticles: Representative native polyacrylamide gel electrophoresis image of synthetic rHDL-APOE3 and rHDL-APOE4 nanoparticles, lipid-free APOA1, lipid-free APOE3, and lipid-free APOE4, visualized with Coomassie Blue staining. (C) Characterization of synthetic rHDL-APOE3 and HDL-APOE4 nanoparticles: The particle size distributions of purified synthetic rHDL-APOE3 and rHDL-APOE4 nanoparticles were analyzed using dynamic light scattering. Representative images from also shown as insets. (D) Astrocyte cholesterol efflux to synthetic rHDL-APOE3 and rHDL-APOE4: Cholesterol efflux from human glioblastoma astrocytes to synthetic rHDL-APOE3 and rHDL-APOE4 (5 μg/ml) was measured under baseline conditions and after T0901317 pretreatment, as described in Fig. 1A. (E) Neuronal cholesterol uptake mediated by synthetic rHDL-APOE3 and rHDL-APOE4: synthetic rHDL-APOE3 and rHDL-APOE4 (5 μg/ml) were loaded with radiolabeled unesterified cholesterol, and their capacity to facilitate cholesterol uptake in human neuroblastoma cells was assessed as described in Fig. 1C. Mean ± SD is used to express values. Student t-tests were used to compare HDL-mediated neuronal cholesterol delivery between synthetic rHDL-APOE3 and rHDL-APOE4, as well as astrocyte cholesterol efflux under various conditions. Five separate experiments were conducted to evaluate each condition. APO, apolipoprotein; DMPC: 1,2-dimyristoyl-sn-glycero-3-phosphocoline; rHDL, reconstituted HDL.

Fig. 4

Reconstituted Oregon Green 488-Labeled HDL-APOE nanoparticles internalization in human neuroblastoma cells. The phospholipid 1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine was labeled with Oregon Green 488 and incorporated into synthetic rHDL-APOE3 and HDL-APOE4 nanoparticles. Human neuroblastoma cells were incubated with the labeled nanoparticles or left untreated as a control for 4 h, followed by confocal microscopy and flow cytometry analyses. Representative images of SH-SY5Y neurons and scatter plots showing side scatter (SSC-A) plotted against green fluorescence (GFP/FITC-A): (A and B) Representative control cells (not labeled); (C and D) Cells treated with synthetic rHDL-APOE3 nanoparticles; (E and F) Cells treated with synthetic rHDL-APOE4 nanoparticles. (G) The graph displays fluorescence phospholipid uptake into cells based on data obtained from flow cytometry. Average fluorescence intensity per cell is shown, representing the degree of nanoparticle uptake. Student t test was used to compare both groups. Three independent experiments were conducted to evaluate each condition. APO, apolipoprotein; rHDL, reconstituted HDL.