Impaired Autophagy in APOE4 Astrocytes

Materials

Antibodies were obtained from the following sources: monoclonal mouse anti-actin (MP Biomedicals, Santa Ana, CA; 691001), polyclonal rabbit anti-LC3B (Sigma-Aldrich, St. Louis, MO; L7543), polyclonal rabbit anti p62 (MBL International, Woburn, MA; PM045), polyclonal rabbit anti phospho-Thr389-S6 kinase (Sigma, S6311), polyclonal rabbit anti S6 kinase (Sigma-Aldrich, S4047) and monoclonal mouse anti-Beta Amyloid (6E10) (Covance, Emeryville, CA; SIG-39300). Reagents are as follows: Earle’s Balanced Salt Solution (EBSS; Sigma-Aldrich, E3024) chloroquine (CQ; Sigma-Aldrich, C6628), Rapamycin (Rapa; Cayman Chemical, Ann Arbor, MI; 13346) and Beta Amyloid _1 - 42, HiLyte Fluor^TM 488-label (Anaspec, Campus Drive Fremont, CA; 60479).

Cell lines

The human ApoE3 and ApoE4 expressing astrocytes were previously described [22]. Cells were grown in Dulbeco’s Modified Eagle’s Medium/ Nutrient Mixture F-12 HAM (Sigma-Aldrich, D6421) supplemented with antibiotics and 10% heat-inactivated fetal bovine serum (FBS, Hyclone, CH30160.03). Cells were incubated at 37°C in 5% CO₂ in air, and the medium was changed every 3-4 days. Cells were passaged when 70% confluent using trypsin/Di-sodium ethylenediaminetetra-acetic acid (Biological Industries, 03-045-1). Cells were cultured at 30% confluence in growth medium, 1-2 days before each experiment.

Primary astrocytes cultures

The study was conducted according to the NIH Guidelines for Use and Care of Laboratory Animals and following the approval by Animal Care Committee of the Tel Aviv University, #L-12-020. Mouse primary astrocytes were prepared as previously described [23 –25], with minor modifications. Cerebral cortices from 1- to 3-day-old neonatal ApoE3 and ApoE4 target replacement mice were dissected in ice-cold Hanks’ Balanced Salt Solution (HBSS), carefully stripped of their meninges, and digested with 0.25% trypsin for 10 min at room temperature. Trypsinization was stopped by addition of an equal volume of Dulbeco’s Modified Eagle’s Medium (Sigma-Aldrich, D5796) to which we added 0.02% deoxyribonuclease I. The cells were dispersed into a single-cell level by repeated pipetting, then pelleted in centrifuge for 7 min at 700 g, 4°C and resuspended in culture medium supplemented with 10% FCS and antibiotics. Cells were seeded at a density of 300,000 cells/ml (equivalent to 62,500 cells/cm²) in flasks and cultured at 37°C in humidified 5% CO₂–95% air. A day after the preparation, all flasks were washed twice with 5 ml PBS, to allow washing of preparation remnants. Medium was replaced every 4 days. These cultures (“mixed glia”) reached confluence after 7–10 days and were used between 15 and 20 days after preparation. Astrocytes were isolated from the mixed glia cultures by mild trypsinization (0.06% trypsin) resulting in detachment of an intact layer of cells containing almost all the astrocytes. After reaching confluence (∼7–10 days), the astrocytes were treated with 10μM Ara-C for 4 days to eliminate proliferating cells. At this stage the culture contains 95% astrocytes [25].

Transfections

Transfection was performed using the Lipofectamine^® 2000 reagent (Invitrogen, Carlsbad, CA) as follows: the cells were incubated with 1.4μg plasmid DNA and 2.1μl transfection reagent in 6-well plates containing 1.5 ml medium supplemented with 10% Opti-MEM^® for 5 h, after which another 1.5 ml of medium was added. Medium was replaced after 24 h.

Lysate preparation and immunoblotting

Cells were grown in medium with or without the indicated stimuli. After treatment, cells were solubilized in lysis buffer (50 mM HEPES pH = 7.5, 150 mM NaCl, 10% glycerol, 1% triton X, 1 mM EDTA pH = 8, 1 mM EGTA pH = 8, 1.5 mM MgCl₂, 200μM Na₃VO₄, 150 nM aprotinin, 1μM leupeptin, 500μM AEBSF). Lysates were cleared by boiling in gel sample buffer followed by centrifugation as described [26, 27]. Lysates were resolved by SDS-polyacrylamide gel electrophoresis through 10–12.5% gels and electrophoretically transferred to nitrocellulose membranes. Membranes were blocked for 1 h in TBST buffer (0.05 M Tris HCl pH 7.5, 0.15 M NaCl and 0.1% Tween 20) containing 6% milk, blotted with primary antibodies for 2 h, followed by secondary antibody linked to horseradish peroxidase for 1 h. Immunoreactive bands were detected with the enhanced chemiluminescence reagent. Densitometric analysis of the results was performed using the ImageJ program.

Electron microscopy

Cells were grown in Matek dishes with coverslip in the bottom. Cells were first fixed in 4% paraformaldehyde followed by 2% paraformaldehyde and 1% glutaraldehyde, then fixed in osmium tetroxide and embedded in epon araldite, sectioned at 90 nm thickness and analyzed with an EM10 Zeiss electron microscope as previously described [28]. Electron micrographs were obtained at a magnification of ×5000 and ×25,000.

In situ Aβ plaque degradation in the presence of ApoE3 and ApoE4 astrocytes

This assay is based on measurement of the ability of the astrocytes to clear Aβ plaques following their co-incubation with Aβ plaques containing brain sections from 5XFAD mice [29, 30]. Accordingly, 20μm coronal frozen brain sections were prepared from 9-13-month-old 5XFAD mice using cryostat and placed on glass slides, 3-4 consecutive sections in each slide. The slides were kept at –80°C until being used.

At the beginning of the experiment, the slices were incubated for 48 h at 37°C with ApoE3 and ApoE4 astrocytes (10⁵ cells in 100μl medium per slice). Medium without astrocytes was used as control. The treated slides were gently placed in round petri dishes with thick wet paper, on the placement of the dishes. PapPen was used to draw liquid repellent borders of the sections to avoid spillage of coating solution out of the slides. In the pharmacological experiments, the cells were treated with either rapamycin (100 nM) or with chloroquine (10μM). After 48 h the slides were gently washed of all containing remains by placing them in PBS for 10 min. Then, the slides were fixed with 4% PFA for 10 min, followed by 0.1% Triton solution for 10 min after which they were incubated with blocking solution (8% Horse Serum, 0.3% Triton, 1 gr/ml BSA, 0.02% Sodium Azide in PBS) for 30 min, followed by incubation with primary anti-Beta Amyloid (6E10) antibody at a dilution of 1:750 in blocking solution for 1 h at room temperature. The slides were then washed 5 times with PBST. Secondary antibody (Alexa Flour^© 488 donkey α-mouse IgG, A-21202) was added at dilution of 1:250 and incubated at room temperature for 45 min, followed by 5 washes with PBST. The slides were then stained with DAPI (Vektashield, Burlingame, CA; H-1200). Analysis of the Aβ plaque burden was performed as percentage of the area covered in plaques using the ImageJ software.

Quantification of fluorescent Aβ uptake by flow cytometry

Cells were incubated with HiLyte Fluor^TM 488-labeled Aβ_1 - 42 at the indicated concentration in serum-free medium at 37°C for 2 h. The cells were then washed with PBS three times, trypsinized, centrifuged at 1,500 g for 5 min and resuspended in PBS. Cell fluorescence was analyzed in a fluorescence-activated cell sorter (FACScan; Becton Dickinson, Franklin Lakes, NJ) within 30 min.

Statistical analysis

All experiments were performed at least three times. Results are presented as mean±SE. One-tailed Student’s t-test was used to assess the differences between means. Results were considered statistically significant when p <  0.05. In the in situ plaques elimination assays and the electron microscopy results were analyzed using One Way ANOVA for group analysis followed by the indicated post hoc analysis.

Autophagy and autophagic flux is inhibited in ApoE4 expressing astrocytes

To explore the role of ApoE in autophagy, we examined the extent to which autophagy is differentially regulated by ApoE alleles, utilizing immortalized mouse astrocytes cell lines derived from ApoE knock-in mice expressing either human ApoE4 or ApoE3 [22]. Accordingly, we first used a known trigger of autophagy, namely, amino acid deprivation (incubation in EBSS), for the indicated time periods to determine its effects on the expression of the autophagic proteins, LC3 and p62/SQSTM1 (presently abbreviated as p62). Cytosolic LC3-I is lipidated (LC3-II) during autophagy, becomes associated with the autophagosomal membrane, and serves as a marker for autophagy [31]. p62, a scaffold protein that binds LC3 and ubiquitinated protein aggregates, is degraded during autophagy [32]. Figure 1A shows that the LC3-II/LC3-I ratio, which is a marker of autophagic activation, was significantly higher in ApoE3 astrocytes compared to ApoE4 astrocytes, after 1, 2, and 4 h of starvation (6.5±1.14, 6.2±0.35, and 5.6±0.71 compared to 2.4±0.36, 2.58±0.50, and 1.82±0.02, respectively; p <  0.05). In addition, the levels of p62, which is associated with the degradation steps and is degraded during autophagy, decreased more rapidly following amino acid starvation in ApoE3 astrocytes compared to ApoE4 astrocytes (0.23±0.01, 0.18±0.03, and 0.24±0.03 compared to 0.43±0.04, 0.52±0.13, and 0.28±0.06, respectively; p <  0.05). Taken together, these results imply that induction of autophagy is impaired in ApoE4 astrocytes. Next, we examined autophagy levels in ApoE3 and ApoE4 astrocytes following treatment with rapamycin, an inhibitor of mTOR, which is known to induce autophagy. Utilizing the LC3-II/LC3-I ratio and p62 levels as markers, we found that, similar to the EBSS treated cells, autophagy was significantly lower in ApoE4 astrocytes following treatment with 100 or 150 nM rapamycin. Accordingly, the LC3II/LC3I ratio under these conditions was 2.9±0.40 and 3.3±0.57 in ApoE3 compared to 1.49±0.21 and 1.34±0.23 for the corresponding ApoE4 cells (p <  0.05; Fig. 1B) whereas the corresponding p62 levels were 0.31±0.07 and 0.28±0.04 in ApoE3 compared to 0.49±0.04 and 0.53±0.09 in ApoE4, (p <  0.05; Fig. 1B). In addition, we found that the EBSS/rapamycin treatments inhibited mTOR activity and decreased the levels of its downstream effector pS6K (Fig. 1A, B). Interestingly, LC3-II/LC3-I ratio was significantly higher under basal untreated conditions in ApoE3 cells compared with ApoE4 cells (Fig. 1A, B), suggesting that the basal autophagy levels in untreated cells are also higher in ApoE3 than in the ApoE4 astrocytes.

Autophagy is a dynamic process, in which LC3-II generation is followed by its degradation in the autolysosome; thus, an increase in LC3-II levels can be derived either from de novo autophagy induction or from inhibition of protein degradation by autophagy. These possibilities can be resolved by measurements of autophagic flux, which are performed by comparing LC3-II levels in the presence or in the absence of a drug that blocks the fusion of the autophagosome with the lysosome [33]. Accordingly, the autophagic flux was determined in the ApoE-expressing astrocytes, using amino acid starvation as autophagy inducer and chloroquine, which inhibits the fusion of autophagosome with the lysosome. As can be seen in Fig. 2, chloroquine led to LC3-II accumulation following starvation for 2 h, in ApoE3 and ApoE4 astrocytes. Quantification of these results revealed that LC3-II/actin levels were significantly higher in ApoE3 astrocytes compared with ApoE4 (9.26±1.75 and 4.5±0.16, respectively; p <  0.05), demonstrating that the autophagic flux is lower in ApoE4 astrocytes compared to ApoE3 astrocytes.

Autophagy can be also monitored microscopically by tagging the LC3 protein with GFP [33]. Since lipidated LC3 is attached to the autophagosome membrane, this tag allows monitoring the formation of autophagosomes in the cells (also termed LC3-GFP puncta). Thus, to further confirm the relationship between ApoE isoforms and autophagy levels, we transfected ApoE astrocytes with LC3-GFP vector and measured the autophagosome formation after 1 h amino acids starvation. Representative images are shown in Fig. 3. Using quantitative analysis, we found that the area of LC3-GFP puncta as the percentage of total cell area was significantly higher in ApoE3 compared to ApoE4 astrocytes. The observed differences are in accordance with the immunoblot results and suggest that the induction of autophagy is more prominent in ApoE3 than in ApoE4 astrocytes. These findings were also confirmed by electron microscopy analysis, which revealed that ApoE3 astrocytes are more responsive to induction of autophagy by amino acids starvation than the ApoE4 astrocytes and that the ApoE3 astrocytes contained less electrodense aggregates than the corresponding ApoE4 astrocytes (Fig. 4). Taken together, these results show that ApoE4 mitigates both basal and induced autophagy in mouse astrocytes.

Insoluble Aβ plaques elimination and soluble Aβ uptake are inhibited by ApoE4 and are affected by chloroquine and rapamycin

It has been suggested that autophagy plays a role in the degradation of senile plaques [14]. We therefore examined the extent to which the ApoE4-driven impaired autophagy is associated with reduction of Aβ plaques clearance. This was pursued utilizing an in situ plaque removal model, in which the ability of the astrocytes to clear senile plaques from brain sections of 5XFAD transgenic mice is determined. As shown in Fig. 5, the ApoE3 astrocytes reduced the levels of Aβ containing senile plaques to a residuary level of 52±11.9% (p <  0.001) of the total Aβ plaques area as compared with cell-free medium. In contrast, ApoE4 astrocytes had no effect under these conditions on the Aβ plaque area (Fig. 5A, B). Hence, ApoE3 astrocytes remove Aβ plaques more effectively than ApoE4 astrocytes. The possibility that the impaired ability of the ApoE4 astrocytes to eliminate Aβ plaques is related to its effects on autophagy was investigated next. To explore this effect, the cells were treated with either the autophagy inhibitor chloroquine or the autophagy inducer rapamycin and the resulting effects on the removal of Aβ plaques by the ApoE astrocytes were determined. As shown in Fig. 6, chloroquine completely abolished the ability of ApoE3 astrocytes to remove Aβ plaques. In contrast, rapamycin significantly increased the Aβ plaque removal capacity of ApoE4 cells (residuary of 31.5±3.1% in ApoE3 and 30.1±5.0% in ApoE4 compared with cell-free medium, p <  0.05) (Fig. 7).

Following the finding that ApoE3 astrocytes eliminate Aβ plaques more efficiently than ApoE4 astrocytes, we examined the possibility that ApoE4 also affects the uptake of soluble Aβ_1 - 42. This was performed by flow cytometry analysis of the amount of fluorescent Aβ, which accumulates within the ApoE3 and ApoE4 astrocytes following 2 h incubation. As shown in Fig. 8A, Aβ uptake by ApoE3 cells (2.94±0.07 and 4.33±0.06; fold induction at 0.025 and 0.05μM Aβ, respectively) was significantly higher than the corresponding uptake by ApoE4 cells (1.62±0.01 and 2.04±0.04; fold induction at 0.025 and 0.05μM Aβ, respectively; p <  0.01). Moreover, when cells were co-incubated with chloroquine to inhibit autophagy, the rate of Aβ uptake was significantly reduced in ApoE3 cells (from 3.2±0.32 to 2.2±0.12; p <  0.05) but not in the ApoE4 cells, whose basal rates of Aβ uptake were low and not affected by this treatment (Fig. 8B). Taken together, these findings show that ApoE4 inhibits Aβ plaques elimination and soluble Aβ uptake, and suggest that the impairment of autophagy is related to these effects.

Autophagy and autophagic flux is inhibited in ApoE4 expressing primary astrocytes

To exclude the possibility of a clonal effect of the cultured astrocytes, we prepared primary astrocytes cultures from ApoE3 and ApoE4 transgenic mice. As shown in Fig. 9A, following amino acids starvation, LC3-II levels decreased more rapidly in ApoE3 astrocytes compared to ApoE4 astrocytes (1.23±0.25, 0.71±0.11 and 0.38±0.03 compared to 1.63±0.33, 1.11±0.15 and 0.67±0.11, respectively; p <  0.05). If this decreased levels of LC3-II in ApoE3 astrocytes results from enhanced degradation by autophagy, it may indicate that autophagy is more efficient also in primary ApoE3 astrocytes. Indeed, when autophagic degradation was blocked using chloroquine (Fig. 9B), the reduction in LC3-II in ApoE3 astrocytes was prevented, indicating that the decreased levels resulted from enhanced autophagy. Moreover, the levels of LC3-I also decreased in ApoE3 but not in ApoE4 astrocytes following autophagy induction (1.15±0.47, 0.42±0.16, and 0.19±0.01 compared to 2.84±0.95, 1.82±0.58, and 1.20± 0.31, respectively; p <  0.05; Fig. 9A). These results may suggest that the lipidation of LC3-I is more efficient in ApoE3 astrocytes. In addition, the starvation resulted in a more dramatic decrease of p62 levels in ApoE3 astrocytes compared to ApoE4 astrocytes (0.34±0.06, 0.30±0.08, and 0.28±0.03 compared to 0.52±0.07, 0.55±0.09, and 0.96± 0.27, respectively; p <  0.05; Fig. 9A). Thus, although autophagy kinetics differs in the primary astrocytes compared to the astrocyte cell lines, in both cases the autophagy is more efficient in ApoE3 astrocytes.

Taken together, the results show that ApoE4 inhibits autophagy and the clearance of Aβ plaques and that it increases the extracellular accumulation of Aβ via a chloroquine sensitive mechanism.