ApoE4-carrying Human Astrocytes Oversupply Cholesterol into Neurons and Promote Aβ Generation

Background: The onset of Alzheimer’s disease (AD) typically occurs later in life. Recent genetic analysis of patients and unaffected individuals revealed multiple genetic variants associated with late-onset AD. One of the strongest genetic risk factors for AD is 𝜀 4 allele of APOE encoding apolipoprotein (ApoE), which is predominantly expressed in astrocytes. The role and mechanism of ApoE in initiating AD-associated pathologies, including amyloid-β (Aβ) accumulation and neurodegeneration in neurons, remains to be elucidated. Methods: Human induced pluripotent stem cells (hiPSCs) from healthy individuals and isogenic cells in which the ApoE 𝜀 3 allele was replaced with an 𝜀 4 allele were selected to generate human neurons and astrocytes. To investigate the effect of astrocytic ApoE4 on neuronal Aβ production, iPSC-derived neurons carrying the ApoE 𝜀 3 allele were cultured in conditioned media from healthy iPSC-derived astrocytes (ApoE3/E4 heterozygote) for ve weeks. Then, the media were replaced with either ApoE3 or ApoE4 astrocyte conditioned media (ACM), cultured for four days, and neuronal amyloid precursor protein (APP) expression and Aβ production were measured. To determine potential mechanisms for upregulation of APP in neurons by ApoE4 ACM, changes in plasma membrane lipid rafts were investigated by staining for cholera toxin B. Methyl-b-cyclodextrin (MβCD) was applied to deplete cholesterol in ApoE4 ACM. Results: Secretory factors in conditioned media from hiPSC-derived astrocytes carrying APOE4 signicantly increased the levels of APP and Aβ secretion in hiPSC-derived neurons. Increasing cholesterol levels in culture media mimicked the effects of ApoE4 ACM by inducing the formation of lipid rafts that potentially provide a physical platform for APP localization on the membrane. We further found that reducing cholesterol levels in ApoE4 ACM with MβCD abolished its effects on neuronal lipid raft expansion and Aβ generation. Conclusions: Our study suggests that ApoE4 astrocytes contribute to amyloidosis by the expansion of lipid

variants are located in genes known to be enriched in glial cells [4,7]. The precise AD-related phenotypes induced by these variants, and the mechanisms by which they arise, remain to be elucidated. Moreover, the reason as to why the timing of disease onset by these genetic factors is later than that induced by familial mutations located in APP, PSEN1, or PSEN2 remains uncertain.
APOE4 is one of the strongest genetic risk factors for LOAD [6]. ApoE is an apolipoprotein encoded by the APOE gene on chromosome 19, and is well known for its function in lipid transport by formation of lipoprotein complexes. In the central nervous system, ApoE is produced primarily by astrocytes, and its expression is upregulated in microglia under neurodegenerative conditions [7,8]. There are three genotypes for ApoE in humans, including ApoE2, ApoE3, and ApoE4. Each genotype produces proteins that are considered to have structural differences according to amino acid sequences at 112 and/or 158 (ApoE2 -Cys112, Cys158; ApoE3 -Cys112, Arg158; ApoE4 -Arg112, Arg158) [9]. Although the difference in their sequence appears subtle, the translated proteins result in a signi cant difference in the risk for AD.
While the 2 allele is known to be protective, bearing 4 increases the risk of AD [9,10].
A previous study using human induced pluripotent stem cells (iPSCs) from ApoE4 carriers suggested that ApoE4 contributes to amyloidosis by increasing Aβ secretion in ApoE4 neurons and decreasing Aβ clearance in ApoE4 astrocytes [11,12]. However, the involvement and mechanisms by which ApoE4 astrocytes contribute to neuronal Aβ production remains to be determined. Here, using hiPSC-derived astrocytes and neurons carrying ApoE3 or ApoE4, we aimed to investigate whether ApoE4 astrocytes regulate neuronal Aβ production and if this regulation is mediated by abnormal cholesterol supply by ApoE4 astrocytes.

iPSCs culture
The use of human iPSCs was approved by the Institutional Review Board (IRB) of Daegu Gyeongbuk Institute for Science and Technology (Permit Number: DGIST-190829-BR-071-01). ApoE3 iPSC line is generated from the Coriell Institute's broblast line derived from healthy individuals (age 75, female; #AG09173) by Dr. Yankner Laboratory at Harvard Medical School [13]. ApoE4 isogenic line is generated from this ApoE3 iPSC line as previously described [11]. For ApoE3/E4 heterozygous astrocyte conditioned media (ACM), an iPSC line derived from a healthy individual (age 22, female; #GM23720) was obtained from the Coriell Institute. iPSCs were maintained on matrigel (Corning 354277)-coated plate in mTeSR1 media (Stemcell) at 37 °C with 5% CO 2 conditioned incubator.
Astrocyte differentiation hiPSC-derived astrocytes were generated as described previously [14]. Brie y, NPCs were seeded at 1.5 × 10 5 cells/well in a 6-well plate. The next day, neuronal induction media were replaced with Astrocyte Media (Sciencell). AM was changed every 2 days. After 4 weeks of differentiation, cells were suspended with TrypLE (Gibco) and sorted with an anti-GLAST-PE antibody (Miltenyl Biotec) using a ow cytometer (Sony SH800). The identity of sorted astrocytes was con rmed by immunocytochemistry with anti-GFAP and anti-AQP4 antibodies. ACM were prepared by culture of astrocytes with Neurobasal, 1 x GlutaMAX, 0.5x N-2, 0.5x B-27 (without vitamin A), BSA (Thermo Fisher Scienti c).

Western blotting
Proteins were extracted from ACM-treated neurons with RIPA buffer (50 mM Tris pH 8.0, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS). Concentration of proteins was measured by the Bradford assay kit (Bio-Rad). Lysates were subjected to SDS-PAGE and transferred to polyvinylidene di uoride membrane (Bio-Rad), and proved with the indicated antibodies.

Filipin III staining
Neurons were xed with 4% PFA (Biosesang), then cholesterol staining kit (Biovision) was used for lipin III staining. Based on the manufacturer's instruction, neurons were washed with assay buffer, and incubated with lipin staining solution (1:100 diluted in assay buffer) for 1 hr at room temperature. Then, cells were washed with assay buffer three times and mounted for imaging.

Immunocytochemistry
Cells were brie y washed with cold DPBS and xed with 4% PFA solution at room temperature for 15 minutes. After xation, cells were washed with cold DPBS three times and incubated with CTX-B(Invitrogen) for 30 minutes in dark at room temperature. Cells were washed with DPBS three times and permeabilized with 0.1% Triton-X (Thermo Fisher Scienti c) in PBS for 10 min. Then, cells were incubated in blocking buffer [0.1% Tween 20 (Promega), 10% normal donkey serum (Merck), 2% BSA (Gemini Bio), 1M glycine (Sigma-Aldrich)] for 1hr at room temperature. Primary antibodies were diluted in blocking buffer, and cells were incubated at 37˚C for 1 hr. After 3 times of washing, secondary antibodies diluted in blocking buffer were treated to cells for 1 hr at room temperature. Washed cells were further incubated with Hoechst 33342 for nucleus staining and mounted for imaging.

Microscopy
The LSM 800 confocal microscope (Zeiss) was used for imaging, and the Zen software (Zeiss) and Image J (NIH) were used for data analysis.

Statistical analysis
The Prism 8 (GraphPad) was used for statistical analysis. Unpaired Student's t-test or one-way ANOVA test with Dunnett's post hoc analysis was used for normal distribution, while the Kruskal-Wallis test was performed for non-normal distribution.

Results
ApoE4 astrocyte-conditioned media increased APP expression and Aβ 42 secretion in hiPSC-derived neurons To address the effect of ApoE4 astrocytes on neuronal APP expression and Aβ generation, we utilized an iPSC line derived from healthy individuals carrying the ApoE3 allele and its isogenic line in which ApoE3 is converted to ApoE4 [11]. Both ApoE3 and ApoE4 iPSC lines were differentiated in astrocytes or excitatory neurons as described previously [14,15] with some modi cations (Fig. 1A), and immunostaining con rmed the identity of these cells (Supplementary Fig. 1). In order to investigate whether secretory factors from ApoE4 astrocytes could affect neuronal Aβ production, we generated healthy hiPSC-derived neurons carrying ApoE3 and cultured them in conditioned media from other healthy iPSC-derived astrocytes (ApoE3/E4 heterozygote) for ve weeks. The media were then replaced with either ApoE3 or ApoE4 astrocyte conditioned media (ACM) and cultured for four days (Fig. 1B). We found that ApoE4 ACM positively regulated the expression of APP in hiPSC-derived neurons carrying ApoE3 (Fig. 1C, D). Immunoblotting from neuronal lysates revealed that APP levels were signi cantly increased by ApoE4 ACM (Fig. 1E). We further measured the secreted levels of Aβ 40 and Aβ 42 and found that Aβ 42 secretion was signi cantly increased by ApoE4 ACM. These data show that secretory factors from ApoE4 astrocytes positively regulate neuronal APP expression and Aβ secretion.
Cholesterol positively regulated the formation of lipid rafts and APP expression in hiPSC-derived neurons Accumulation of intracellular cholesterol in hiPSC-derived ApoE4 astrocytes compared to isogenic ApoE3 astrocytes has been recently reported [11,16], and Lin et al. further reported increased cholesterol secretion from ApoE4 astrocytes. Astrocytes supply cholesterol to neurons to support synapse formation and regulate membrane uidity [9,17]. Moreover, cholesterol, along with ganglioside and triglyceride, is a critical component of membrane lipid rafts, which provide a suitable platform for various membranebound proteins, including glutamate receptors. APP and its processing secretases, β-and γ-secretase, are also known to be located in lipid rafts, while α-secretase is mainly expressed in non-lipid rafts [18,19].
Previous studies have shown that increasing cholesterol in the membrane induced the formation of lipid rafts and increased Aβ production [20,21]. Therefore, we hypothesized that increased levels of cholesterol in ApoE4 ACM could be a key factor in upregulating APP and its processing by facilitating the formation of lipid rafts. First, the conditions were optimized to regulate environmental cholesterol levels by treating neurons with cholesterol or methyl β-cyclodextrin (MβCD; to deplete cholesterol) in a rat primary neuron culture system (Supplementary Fig. 2A). We found that treatment with 20 µM of cholesterol for four days was su cient to increase neuronal cholesterol, as visualized by lipin III staining, which is naturally uorescent upon cholesterol binding ( Supplementary Fig. 2B, C). We then measured the levels of lipid rafts by staining for cholera toxin B (CTX-B), a well-known lipid raft marker, and found a signi cant increase in lipid rafts following cholesterol treatment ( Supplementary Fig. 2D-E).
APP expression was also increased in cholesterol-treated rat primary neurons (Supplementary Fig. 2F). There was no reduction in both neuronal cholesterol and lipid rafts following MβCD treatment in rat primary neurons (Supplementary Fig. 2B-E), which could be due to the homeostatic mechanism of neurons to compensate for the loss of intracellular cholesterol. To address whether cholesterol treatment is su cient to mimic the effect of ApoE4 ACM on neuronal APP expression and Aβ secretion, as shown in Fig. 1C-F, we treated hiPSC-derived neurons with cholesterol or MβCD (Fig. 2A). The upregulation of cholesterol in neurons by exogenous cholesterol treatment was con rmed by lipin III staining (Fig. 2B,  C). Consistent with the observation in rat primary neurons, MβCD treatment did not alter the levels of cholesterol in hiPSC-derived neurons (Fig. 2B, C). We then measured levels of lipid rafts in neurons and found that cholesterol treatment increased the area of CTX-B signals without affecting intensity, suggesting the expansion of lipid rafts (Fig. 2D, E). We also measured APP levels in neurons and found signi cant upregulation of APP expression by cholesterol treatment, due to increased area of APP signals rather than intensity (Fig. 2F), which is consistent with rat primary neurons treated with cholesterol ( Supplementary Fig. 2F). We further found that the co-localization of APP and CTX-B was signi cantly increased by cholesterol treatment in both hiPSC-derived neurons and rat primary neurons (Fig. 2G,   Supplementary Fig. 2G). To determine whether APP upregulation is caused simply by the expansion of lipid rafts or if more APP is recruited to the given area of lipid rafts, we measured the intensity of APP in the CTX-B/APP co-localized area. The data showed that there was no alteration in APP intensity in these regions (Fig. 2G, Supplementary Fig. 2G), suggesting that increased APP expression by extracellular cholesterol supply is mainly due to the increased area of lipid rafts.

Cholesterol in ApoE4 ACM increased the formation of lipid rafts in hiPSC-derived neurons
To investigate whether the effects of ApoE4 ACM on neuronal cholesterol levels and lipid raft formation were due to cholesterol oversupply, we measured cholesterol levels in ApoE3 and ApoE4 ACM. Consistent with a previous report [11], we found higher cholesterol levels in ApoE4 ACM than in ApoE3 ACM (Fig. 3A).
We then measured lipin III in hiPSC-derived neurons cultured with conditioned media from either ApoE3 or ApoE4 astrocytes as described in Fig. 1B, and ApoE4 ACM-treated neurons displayed increased lipin III signals (Fig. 3B, C). We also found that the area and total levels of CTX-B were signi cantly increased compared to those of ApoE3 ACM-treated neurons (Fig. 3D-E). To determine whether the cholesterol in ApoE4 ACM is the major cause of the upregulation of lipid rafts and APP in neurons, we added MβCD to ApoE4 ACM during neuronal culture (Fig. 3F) and found that the upregulation of neuronal cholesterol by ApoE4 ACM was signi cantly attenuated by MβCD, potentially due to its scavenging effect toward exogenous cholesterol (Fig. 3G, H). The addition of MβCD to ApoE4 ACM abolished the ApoE4 ACMinduced increase in lipid raft expansion (Fig. 3I, J).

Reducing cholesterol attenuated ApoE4 ACM-induced APP upregulation and Aβ 42 secretion in hiPSC-derived neurons
To determine whether cholesterol in ApoE4 ACM was the major cause for the upregulation of APP and its metabolism to produce Aβ in hiPSC-derived neurons, we added MβCD to ApoE4 ACM during neuronal culture (Fig. 4A). As shown in Fig. 1E, neurons cultured with ApoE4 ACM showed increased expression of APP compared to those cultured with ApoE3 ACM. However, in the presence of MβCD, ApoE4 ACM was not able to induce signi cant upregulation of APP. Increased co-localization of lipid rafts and APP by ApoE4 ACM was also abolished in neurons treated with MβCD ( Fig. 4B-E). Furthermore, MβCD treatment inhibited the ApoE4 ACM-induced increase in Aβ 42 secretion in hiPSC-derived neurons (Fig. 4F). Taken together, these data suggest that an excess supply of cholesterol is responsible for ApoE4 ACM-mediated neuronal Aβ 42 overproduction.

Discussion
Abnormal neuronal cholesterol levels have been linked to AD-related pathology both in vitro and in vivo. For example, increased levels of cholesterol were observed in AD brain samples, and the severity of pathology was correlated with cholesterol levels [22,23]. Inhibition of cholesterol e ux by reducing expression of CYP46A1, a cholesterol 24S-hydroxylase, in a neuron-speci c manner was shown to result in cognitive de cits and neuronal death in wild-type mice. The study further showed the recruitment of APP to lipid rafts and Aβ upregulation prior to neuronal death in both wild-type and APP23 mice [24].
Here, we showed that ApoE4 bearing astrocytes increased cholesterol supply to neurons, leading to the upregulation of APP and Aβ secretion in a paracrine manner. Application of cholesterol recapitulated the effects of ApoE4 ACM on neurons, and the effects of ApoE4 ACM were abolished by MβCD. These data revealed the major contribution of secretory cholesterol from ApoE4 astrocytes to lipid raft expansion and APP expression, which promotes Aβ 42 secretion in neurons.
A recent study showed that CHO cell lines expressing familial AD-associated PSEN1 ΔE9 display increased levels of cholesterol, which leads to the enrichment of APP in lipid rafts [25]. Similarly, a previous study suggested binding between cholesterol and β-secretase-derived APP C-terminal fragment (β-CTF) [26]. These data suggest the active role of cholesterol in recruiting APP to lipid rafts. Here, we found that cholesterol oversupply by ApoE4 astrocytes increases the levels of APP in lipid rafts. Although we did not nd a difference in APP intensity (local clustering density) in lipid rafts by extracellular cholesterol supply (Fig. 2G and Supplementary Fig. 2G), neurons cultured in ApoE4 ACM displayed increased APP intensity in lipid rafts, which was abolished by MβCD treatment (Fig. 4D). As cholesterol was shown to increase the proximity between APP and β-secretase 1 (BACE1) in the membrane [27], further studies are required to examine the expression pattern of β-and γ-secretases on lipid rafts in neurons when they are cultured with ApoE3 or ApoE4 ACM.
Cholesterol was also shown to increase Aβ production through the facilitation of APP endocytosis to endosomes [20]. The low pH environment in endosomes increases BACE1 activity to facilitate APP cleavage and produce β-CTF. Increased size and/or number of early endosomes were found in the brains of AD patients as well as in multiple AD model systems. Moreover, cholesterol loading to the neuronal plasma membrane was shown to result in enlarged endosomes [21]. In the current study, for the rst time, we revealed the impact of ApoE4 astrocytes on neuronal cholesterol and lipid rafts that affect Aβ production. Further studies are required to determine whether astrocytic ApoE4 regulates APP processing only in lipid rafts or in endosomal compartments, that is whether cholesterol only affects membraneoriginated Aβ production and/or facilitates APP endocytosis and subsequent Aβ generation in endosomes.
Lipid rafts provide a platform for various neuronal membrane proteins and are important for synaptic functions [28]. Previously, cholesterol was shown to promote synapse maturation, whereas depletion of cholesterol signi cantly reduced lipid raft domain and synapses [28,29]. The positive correlation between neuronal activity and Aβ production has been supported by multiple studies [30,31], and the expansion of lipid rafts by ApoE4 ACM could also contribute to the upregulation of Aβ by increasing neuronal activity.
In addition to Aβ production, Aβ clearance can be manipulated by the ApoE4 ACM. Cholesterol accumulation in cellular organelles such as lysosomes has been found to increase intracellular Aβ levels [32]. Therefore, it would be interesting to determine whether increased cholesterol supply from ApoE4 astrocytes induces lysosomal cholesterol accumulation and impairs lysosomal activity and Aβ degradation in neurons.

Conclusion
ApoE4, the strongest genetic risk factor for LOAD, has recently been shown to have detrimental effects on astrocytes, including endocytic defects and impaired homeostatic functions [11,33,34]. It is not clear, however, whether altered astrocytic properties could affect neighboring neurons and induce ADassociated pathology. Here, we revealed that ApoE4 astrocytes could regulate neuronal APP metabolism to initiate amyloidosis through cholesterol oversupply (Fig. 5). This study provides new insight into the contribution of ApoE4 and astrocytes to amyloidosis in AD, and the importance of regulating astrocytic ApoE isotypes.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.