Abstract
Alzheimer's disease (AD) is a debilitating dementia characterized by progressive memory loss and aggregation of amyloid-β (Aβ) protein into amyloid plaques in patient brains. Mutations in presenilin (PS) lead to abnormal generation of Aβ, which is the major cause of familial AD (FAD), and apolipoprotein E4 (ApoE4) is the major genetic risk factor for sporadic AD (SAD) onset. However, whether dysfunction of PS is involved in the pathogenesis of SAD is largely unknown. We found that ApoE secretion was completely abolished in PS-deficient cells and markedly decreased by inhibition of γ-secretase activity. Blockade of γ-secretase activity by a γ-secretase inhibitor, DAPT, decreased ApoE secretion, suggesting an important role of γ-secretase activity in ApoE secretion. Reduced ApoE secretion is also observed in nicastrin-deficient cells with reduced γ-secretase activity. PS deficiency enhanced nuclear translocation of ApoE and binding of ApoE to importin α4, a nuclear transport receptor. Moreover, the expression of PS mutants in PS-deficient cells suppressed the restoration effects on ApoE secretion compared with the expression of wild-type PS. Plasma ApoE levels were lower in FAD patients carrying PS1 mutations compared with normal control subjects. Our findings suggest a novel role of PS contributing to the pathogenesis of SAD by regulating ApoE secretion.
SIGNIFICANCE STATEMENT Familial AD (FAD) typically results from mutations in the genes encoding amyloid precursor protein, presenilin 1 (PS1), or PS2. Many PS mutants have been found to exert impaired γ-secretase activity and increased amyloid-β 42 (Aβ42)/Aβ40 ratio, which induce early amyloid deposition and FAD. On the other hand, apolipoprotein E4 (ApoE4) is the major genetic risk factor for sporadic AD (SAD) and contributes to AD pathogenesis because it has reduced Aβ clearance capability compared with ApoE3 and ApoE2. FAD and SAD have long been considered to be caused by these two independent mechanisms; however, for the first time, we demonstrated that PS is essential for ApoE secretion and PS mutants affected ApoE secretion in vitro and in human samples, suggesting a novel mechanism by which PS is also involved in SAD pathogenesis.
Introduction
Alzheimer's disease (AD) is the most common form of dementia, and its pathologic hallmarks include the deposition of extracellular amyloid-β (Aβ) protein aggregates as amyloid plaques in patient brain (Scheltens et al., 2016). The intramembrane protease γ-secretase proteolytically process amyloid precursor protein (APP), to generate a series of amyloidogenic Aβ with 39–43 aa in the brains of AD patients (Golde et al., 1992; De Strooper et al., 2012). The mature γ-secretase protease consists of four components: presenilin (PS), nicastrin (NCT), anterior pharynx-defective 1 (APH-1), and PS enhancer 2 (PEN-2; Kimberly et al., 2003; Bai et al., 2015). Clinically, AD is divided into sporadic AD (SAD) and familial AD (FAD). FAD cases are caused by dominantly inherited genetic mutations in an individual carrying a mutation that confers disease in an autosomal-dominant inheritance pattern. While it is hypothesized that SAD arises because of reduced clearance of Aβ, this has not been definitively proved (Kanekiyo et al., 2014). Reliable estimates of the prevalence of FAD are 1–5%, and the prevalence of SAD accounts for >95% of all AD cases (Scheltens et al., 2016). FAD was found to be genetically linked to missense mutations either in PS or APP, and consistently increase the relative ratio between the long (Aβ42) and short (Aβ40) amyloid peptides (Scheuner et al., 1996; Huang and Mucke, 2012). Most mutations in the PS1 and PS2 genes lead to elevated molar ratios of Aβ42/Aβ40 (Borchelt et al., 1996; Sun et al., 2017), which is responsible for the early onset of FAD (Scheuner et al., 1996; Huang and Mucke, 2012; Watanabe and Shen, 2017).
Polymorphism in the apolipoprotein E (ApoE) gene is a major genetic risk determinant of SAD, with the ApoE4 allele conferring a stronger genetic risk factor for AD than the more common ApoE3 allele, whereas the presence of the ApoE2 allele conferring a decreased risk for AD (Strittmatter et al., 1993; Yamazaki et al., 2019). ApoE is one of the major apolipoproteins that regulate lipid and cholesterol transport and delivery through cell surface ApoE receptors (Mahley, 1988; Michikawa et al., 2000; Kanekiyo et al., 2014). Mounting evidence suggests that the ApoE4 protein isoform drives amyloid pathology through the disturbance of Aβ clearance machinery, leading to Aβ accumulation in the brain and an increased risk for the pathogenesis of AD (Liu et al., 2013). Much evidence from in vivo studies has shown that Aβ clearance is slower in ApoE4-TR mice compared with ApoE3-TR mice (Castellano et al., 2011). Indisputable evidence demonstrated that ApoE4 knock-in (KI) mice have greater Aβ deposition compared with ApoE3 knock-in mice (Holtzman et al., 2000; Li, 2019); whereas human ApoE2-expressing mice showed the dramatic reduction of Aβ deposition (Fagan et al., 2002; Li, 2019). These findings suggest that clearance of Aβ by ApoE is important for preventing Aβ deposition for AD onset and that the regulation of ApoE metabolism by other molecules can affect Aβ clearance.
Together, PS and ApoE have crucial and diverse roles, respectively, in FAD and SAD pathogenesis. Analysis of 138 distinct FAD-derived PS1 mutations revealed that 90% of the analyzed mutations impaired γ-secretase-dependent cleavage of APP, as shown by reduced production of Aβ40 and increased Aβ42/Aβ40 ratio. However, some of the PS mutations lead to reduced Aβ42/Aβ40 ratio (Sun et al., 2017). These findings suggest that PS may be involved in AD pathogenesis by different mechanisms other than regulating Aβ generation. PS is a multifunctional molecule and has several functions in addition to serving as γ-secretase. The γ-secretase-independent functions of PSs include stabilizing β-catenin in Wnt signaling pathway; regulating calcium homeostasis; and regulating autophagy, protein trafficking, and apoptosis, and their interaction with synaptic transmission (De Strooper et al., 2012; Zhang et al., 2013; Duggan and McCarthy, 2016). We hypothesized that PS or PS functions may be involved in ApoE metabolism or secretion. Herein, we found a novel role of PS in regulating ApoE secretion: that is, PS is essential for ApoE secretion and PS mutations exert impaired ApoE secretion in vitro and in vivo. For the first time, our findings suggest a novel mechanism by which PS is also involved in AD pathogenesis via regulating ApoE secretion.
Materials and Methods
Cell culture.
Wild-type (WT), PS double-knock-out (PS1/2−/−), PS1 knock-out (PS1−/−), and PS2 knock-out (PS2−/−) mouse embryonic fibroblast (MEF) cells were provided by Bart De Strooper, Department of Neurosciences, Leuven Brain Institute, Leuven 3000, Belgium (Herreman et al., 1999). NCT knock-out (NCT−/−) MEF cells were previously generated from NCT-deficient mouse embryos (Li et al., 2003a). Cell lines were maintained in DMEM (Thermo Fisher Scientific) containing 10% fetal bovine serum (FBS). WT, PS1/2−/−, PS1−/−, PS2−/−, and NCT−/− cells were cultured for 48 h in serum-free DMEM to analyze ApoE secretion. A total of 14 ml of serum-free DMEM culture supernatant was centrifuged at 5000 × g at 4°C using a centrifugal filter unit (Amicon Ultra-15; catalog #UFC901024, Millipore) to a final volume of 200 μl. The amount of sample used for the examination of secreted ApoE levels was normalized based on the total protein level and was subjected to SDS-PAGE and Western blotting.
Highly astrocyte-rich cultures were prepared according to a method described previously (Isobe et al., 1999). In brief, the brain was removed from 18-d-old WT C57BL/6J mouse embryos under anesthesia. The cerebral cortex was dissected, freed from the meninges, and minced into small pieces. The cortical fragments were incubated in 0.25% trypsin and 20 mg/ml DNase I in PBS (8.1 mm Na2HPO4, 1.5 mm KH2PO4, 137 mm NaCl, and 2.7 mm KCl, pH 7.4) at 37°C for 20 min. The fragments were then dissociated into single cells by repeated pipetting. The dissociated cells were seeded in 75 cm2 dishes at a density of 1 × 107 cells in DMEM containing 10% FBS. After 7 d of incubation in vitro, astrocytes in the monolayer were trypsinized (0.1%) and reseeded. The astrocyte-rich cultures were maintained in DMEM containing 10% FBS until use. Primary cultured MEFs were prepared according to a method described previously, with some modifications (Li et al., 2003a). Briefly, skin from 18-d-old embryos of WT C57BL/6J mice or postnatal day 1 ApoE3-KI C57BL/6NCrSlc mice was minced, suspended in 0.25% trypsin and 20 mg/ml DNase I in PBS, incubated at 37°C for 20 min, and then seeded in 75 cm2 flasks. The cultures were maintained in DMEM containing penicillin and streptomycin (50 μg/ml) and 10% FBS until use.
Western blotting.
Cultured cells were rinsed twice with ice-cold PBS and homogenized in lysis buffer (25 mm Tris-HCl, pH 7.6, 150 nm NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS) containing a protease inhibitor cocktail (Roche) for 30 min at 4°C, then centrifuged at 14,000 rpm for 5 min at 4°C. Equal amounts of protein from the cell lysate were dissolved in 4× SDS sample buffer (250 mm Tris-HCl, pH 6.8, 40% glycerol, 8% SDS, and 0.01% bromophenol blue) and heated for 5 min at 95°C for denaturation. Proteins were separated by SDS-PAGE and blotted onto polyvinylidene difluoride membranes (Sigma-Aldrich). The membranes were then probed overnight with specific primary antibodies at 4°C. Appropriate peroxidase-conjugated secondary antibodies were applied, and the blots were visualized using Super Signal Chemiluminescence solution (Wako) according to the manufacturer instructions. Membranes were then stripped and reprobed with anti-α-tubulin antibody. Quantification was performed using ImageJ software (National Institutes of Health). Detailed information regarding antibodies is provided in Table 1. Antibody validation profiles are available on the websites of the companies from which the antibodies were sourced.
Information regarding antibodies, reagents, and assay kits
Immunostaining.
Cells were seeded and incubated at 37°C for 24 h, then fixed in cold 4% paraformaldehyde for 30 min at room temperature. The cells were then permeabilized with 0.1% Triton X-100 and blocked for 60 min in 10% normal donkey serum in Tris-buffered saline (TBS) with Tween 20 cells were then incubated with the proper primary antibodies for 12 h at 4°C. Immunostaining for ApoE and importin a4 colocalization in WT and PS1/2−/− cells was accomplished using anti-ApoE (F-9) and anti-importin α4 primary antibodies. Immunofluorescent labeling was accomplished using Alexa Flour 488-tagged or Alexa Fluor 568-tagged secondary antibodies (Thermo Fisher Scientific). Before mounting onto slides, cells were stained with 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI; Prolong Gold Antifade Reagent, Thermo Fisher Scientific) and subjected to confocal microscopy. Images were captured using an Olympus FV 3000 confocal microscope (with an oil-immersion plan Apo 60× A/1.40 numerical aperture objective lens. Detailed information regarding antibodies is provided in Table 1.
Confocal imaging and analyses.
Confocal images were acquired as described in the previous section. Laser and detector settings were maintained properly for the acquisition of each target molecule. For all analyses, images were acquired from different cells and slides using a 60× oil differential interference contrast objective at 2048 × 2048 or 1024 × 1024 pixel resolution, respectively. For quantification of cytosolic and nuclear ApoE pixel intensity, acquired images were imported to ImageJ software. Nuclear boundaries were determined by staining with 4′,6-diamidino-2-phenylindole (DAPI; catalog #P36931, Thermo Fisher Scientific). Mean staining intensity of cytosolic and nuclear ApoE was determined for each region of interest. Colocalization of ApoE with importin α4 was quantified using ImageJ software. Threshold intensity was determined using the thresholding function and preset for both fluorescent signals. Pixel intensity for colocalized images was quantified and scored automatically when above the threshold intensity, which was expressed as colocalized mean intensity.
Subcellular fractionation.
Cell-free cytoplasmic and nuclear extracts were prepared as described previously (Theendakara et al., 2016), with some modifications. Briefly, cells were lysed in ice-cold fractionation buffer (20 mm HEPES, pH 7.4, 10 mm KCl, 250 mm sucrose, 1.5 mm MgCl2, 1 mm EDTA, 1 mm EGTA, 1 mm dithiothreitol, and protease inhibitor mixture) and sonicated briefly, and then the nuclei were pelleted by centrifugation for 10 min at 750 × g. The supernatant was then centrifuged at 100,000 × g for 30 min. The resulting supernatant contained the soluble cytosolic fraction. The pellets were resuspended in ice-cold fractionation buffer and sonicated three times for 30 s each, and then the debris was removed by centrifugation for 10 min at 4°C and 20,000 × g. Equal amounts of protein (120 μg) were subjected to SDS-PAGE and Western blotting analysis. Membranes were probed with anti-ApoE antibody. The purity of the nuclear fraction was assessed by the presence of poly(ADP-ribose) polymerase (PARP).
Blue native PAGE.
WT, PS1/2−/−, PS1−/−, PS2−/−, and NCT−/− MEFs were lysed in native sample buffer (Thermo Fisher Scientific) containing 1% digitonin and a protease inhibitor mixture. Lysates were then briefly sonicated and centrifuged at 20,000 × g at 4°C for 5 min; the supernatant was separated on a 4–16% Bis-Tris gel (Thermo Fisher Scientific) according to the instructions of the Novex Bis-Tris gel system (Thermo Fisher Scientific). The γ-secretase complex was probed using antibodies against PS1-C-terminal fragment (CTF), PS2-CTF, APH-1, and NCT.
Infection of WT PS and PS mutants.
For the expression of human WT PS1, WT PS2, or PS mutants (PS1ΔE9, PS1L166P, PS1G384A, and PS2N141I), cDNAs were inserted into the EcoRI site of pMXs-puro (Kitamura et al., 2003). The retroviral plasmids were transfected into platinum-E cells using FuGENE (catalog #E231A, Promega) for retroviral packaging (Kitamura et al., 2003; Watanabe et al., 2005; Futai et al., 2016). After 24 h, the conditioned medium was collected and used as viral stock. For highly efficient retroviral infection, cells were cultured with viral stock containing 5 μg/ml polybrene (catalog #sc-134220, Santa Cruz Biotechnology) and were allowed to proceed for PS expression. To determine the efficacy of retroviral infection of cells, virus stock was added to new medium containing 10% FBS and puromycin (1 μg/ml) immediately before use. After 24 h, the virus stock was replaced with fresh serum-free medium (lacking virus) and allowed to proceed for PS expression.
Human plasma.
Plasma was collected from living FAD patients and normal age-matched control subjects using standard methods (Abdullah et al., 2019). All relevant guidelines and regulations were strictly followed for all methods. For plasma collection, 12 FAD patients carrying PS1 mutations who visited Niigata University Hospital were enrolled. Plasma samples were also obtained from 12 normal control subjects who visited Oita University Hospital. Detailed information about the sex, age, age of AD onset, age of blood taking, and ApoE genotype of these patients is provided in Table 2. The samples were aliquoted after collection and immediately stored at −80°C until use. The average age of the subjects of the normal control group was 55.58 ± 1.96 years, and that of the subjects of the FAD group carrying PS1 mutations was 51.04 ± 2.30 years (data represent the mean ± SEM; p = 0.1469, unpaired two-tailed t test). The clinical diagnosis of AD was made according to the criteria of the National Institute of Neurologic and Communicative Diseases and Stroke/Alzheimer's Disease and Related Disorders Association. All subjects underwent clinical examination that included biochemical, molecular biological, and genomic testing. Experiments using human plasma were performed after obtaining informed consent from each subject. ApoE levels in the plasma of FAD patients and normal age-matched control subjects were determined using a human ApoE (AD2) ELISA kit (EHAPOE, Thermo Fisher Scientific). The ethics committees of Oita University Hospital, Niigata University Hospital, and Nagoya City University Hospital approved the study.
Information of normal controls and FAD patients with various PS1 mutations
Mice.
WT C57BL/6J mice were obtained from SLC. All mice were bred on a 12 h light/dark schedule with ad libitum access to standard chow (CE-2, CLEA) and tap water. ApoE3-KI C57BL/6NCrSlc mice expressing human ApoE3 instead of mouse ApoE were generated using gene-targeting techniques, taking advantage of homologous recombination in embryonic stem cells (KI), as reported previously (Gong et al., 2002). Postnatal day 1 mice harboring the homozygous E3 (3:3) allele and correctly expressing human ApoE3 protein were used in this study for generating primary fibroblasts. Ten-week-old WT C57BL/6J mice (n = 12; 6 male and 6 female) were divided into two groups (each group containing three males and three females) and administered vehicle (DMSO) or DAPT (100 mg/kg) for 48 h via intraperitoneal injection. The dose of DAPT for in vivo experiments was determined according to previous studies (Comery et al., 2005). At the end of the study, the mice were killed by intraperitoneal administration of a combination of anesthetics: medetomidine (0.3 mg/kg), midazolam (4 mg/kg), and butorphanol (5 mg/kg). CSF was then collected via the cisterna magna, the opening between the cerebellum and dorsal surface of the medulla oblongata. Mice were then transcardially perfused with cold PBS, and the brain was immediately collected from each mouse and frozen at −80°C until use. The experiments in this study were performed in strict accordance with the recommendations of the Fundamental Guidelines for Proper Conduct of Animal Experiments and Related Activities in Academic Research Institutions, under the jurisdiction of the Ministry of Education, Culture, Sports, Science and Technology, Japan.
ELISA.
Levels of Aβ40 and Aβ42 peptides were measured using a sandwich ELISA according to the manufacturer instructions (Wako Pure Chemical Industries). The cortices of DMSO- and DAPT-treated mice were homogenized in 19 volumes of ice-cold TBS (10 mm Tris-HCl and 150 mm NaCl, pH 7.6) containing a protease inhibitor cocktail (Roche). The resulting homogenates were centrifuged at 100,000 rpm for 20 min at 4°C. The pellets were washed with ice-cold TBS, and then 10 volumes of 5 m guanidine hydrochloride were added to the pellets. The samples were sonicated and incubated at room temperature for 1 h. The homogenates were centrifuged at 100,000 rpm for 20 min at 4°C, and the resulting supernatants were transferred to new tubes and used for Aβ determination. The amounts of Aβ40 and Aβ42 were determined using ELISA kits (Wako Pure Chemical Industries). The brain homogenates were further diluted 1:8 for Aβ40 and 1:15 for Aβ42 using the dilution buffer provided with the ELISA kit (Wako). Levels of Aβ42 were determined using a highly sensitive ELISA because of the low level of Aβ42 in WT C57BL/6 mice. Aβ levels were normalized to brain tissue weight. Fibroblasts stably expressing human APP695 (Zou et al., 2007) were grown in DMEM (Thermo Fisher Scientific) containing 10% FBS. The medium was changed when the cells reached 100% confluence, and the fibroblasts were treated with DAPT or with DMSO as a vehicle control. The levels of Aβ40 and Aβ42 in the medium were determined 48 h after treatment of fibroblasts with DAPT. All samples were analyzed in quadruplicate.
The sandwich ELISA for ApoE was performed according to the instruction manual of the Human ApoE (AD2) ELISA kit (EHAPOE, Thermo Fisher Scientific). Briefly, all reagents and samples were brought to room temperature (18–25°C) before use. Human plasma samples were diluted 1000-fold using 1× assay diluent. A total of 100 µl of standard or sample was added to appropriate wells, and the assay plate was covered and incubated for 2.5 h. After incubation, the solution was discarded, and the wells were washed four times with 300 µl of 1× wash buffer using a multichannel pipette. After the last wash, the plate was inverted and blotted against clean paper towels for the complete removal of any remaining wash buffer by aspiration or decanting. Next, 100 µl of 1× prepared biotinylated antibody was added to each well and incubated for 1 h, after which the solution was discarded and the washing step was repeated. A total of 100 µl of prepared streptavidin-HRP solution was added to each well and incubated for 45 min, after which the solution was discarded. The plate was then washed, and 100 µl of TMB substrate was added to each well and incubated for 30 min in the dark. Finally, 50 µl of stop solution was added to each well, and the absorbance was measured at 450 nm within 30 min of stopping the reaction. All incubations were performed at room temperature with gentle shaking, and all standards and samples were analyzed at least in duplicate.
Immunoprecipitation.
Immunoprecipitation (IP) experiments were conducted according to a method reported previously (Bergman et al., 2004). WT and PS1/2−/− MEFs were seeded in 10 cm2 dishes for 48 h and then lysed in 0.5 ml of coimmunoprecipitation buffer [50 mm HEPES, pH 7.4, 150 mm NaCl, 2 mm EDTA, Protease Inhibitor Mixture (catalog #11697498001, Roche), and 1% CHAPSO] followed by ultracentrifugation at 100,000 rpm for 20 min at 4°C. The samples were kept on ice and briefly sonicated before ultracentrifugation. The supernatants were then precleaned using protein G-Sepharose (Thermo Fisher Scientific) and subsequently immunoprecipitated overnight with goat anti-ApoE or normal goat IgG (dilution, 1:67), followed by end-over-end rotation. The following day, a mixture of protein G-Sepharose was added to the supernatants, and the incubation was continued for 3 h. The immunoprecipitates were then washed four times with coimmunoprecipitation buffer containing 0.5% CHAPSO and once in PBS, eluted using 4× SDS sample buffer (250 mm Tris-HCl, pH 6.8, 40% glycerol, 8% SDS, and 0.01% bromophenol blue), heated at 95°C for 5 min, and analyzed by immunoblotting. For immunoblot blocking and antibody incubation, 5% nonfat dry milk in TBS with Tween 20 (25 mm Tris-HCl, pH 8.0, 125 mm NaCl, and 0.5% Tween 20) was used. For immunoprecipitation, all steps were performed at 4°C.
Statistical analyses.
Prism 7.0 software (GraphPad Software) was used for statistical analyses. All data are shown as the mean ± SEM of at least three independent experiments with p values. p < 0.05 was considered statistically significant. Student's t tests were used to determine the significance of differences between two groups. Group differences were analyzed by one-way ANOVA followed by Tukey's multiple-comparison tests for multiple groups against the control group. All experiments produced similar results under identical or similar conditions, and the distribution of data were assumed to be normal.
Results
Lack of ApoE secretion and enhanced cytosol-to-nuclear translocation of ApoE in PS-deficient fibroblasts
Because not all PS mutations found in FAD can be interpreted by abnormal Aβ generation, we hypothesized that PS may be involved in the pathogenesis of SAD and sought to identity its function in regulating ApoE secretion. Notably, we found that ApoE secretion was completely abolished in PS1/2−/− fibroblasts. PS1−/− cells exhibited markedly reduced ApoE secretion, whereas PS2−/− fibroblasts exhibited slightly reduced ApoE secretion compared with WT fibroblasts. Intracellular ApoE levels were not altered in these PS-deficient fibroblasts compared with WT fibroblasts (Fig. 1A–C). These results suggest that PS is essential for ApoE secretion but does not affect the levels of intracellular ApoE. The PS2-CTF levels in PS1−/− cells were 1.Ninefold of those in WT cells (Fig. 1A,E). Formation of the PS2 γ-secretase complex was also dramatically upregulated in PS1−/− fibroblasts compared with WT cells (Extended Data Fig. 1-1A,C); however, secreted ApoE was almost undetectable in the culture medium of PS1−/− cells (Fig. 1A). These results suggest that PS1 or the PS1 γ-secretase complex is more crucial for ApoE secretion than PS2 or the PS2 γ-secretase complex.
Lack of ApoE secretion and increased cytosol-to-nuclear translocation of ApoE in PS-deficient fibroblasts. A, Immunoblot analysis of ApoE secretion and intracellular ApoE, PS1-CTF, and PS2-CTF expression in WT, PS1/2−/−, PS1−/−, and PS2−/− fibroblasts cultured for 48 h in serum-free medium. B–E, Quantification of ApoE secretion and intracellular ApoE, PS1-CTF, and PS2-CTF expression levels. Intracellular ApoE, PS1-CTF, and PS2-CTF expression levels were normalized to α-tubulin protein levels. n = 3. **p < 0.01, ***p < 0.001, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. F, Immunostaining for ApoE (red) and nuclear staining with DAPI (blue) in WT, PS1/2−/−, PS1−/−, and PS2−/− fibroblasts. Scale bars, 5 μm. G, H, Quantification of cytosolic and nuclear ApoE intensity in WT, PS1/2−/−, PS1−/−, and PS2−/− fibroblasts. n ≥ 40 different stained cells/group, nuclear boundaries determined based on DAPI. ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. I, J, Cytosolic and nuclear distribution of ApoE in WT and PS1/2−/− fibroblasts. CF, Cytosolic fraction; NF, nuclear fraction. K, Quantification of ApoE nuclear/cytosolic ratio showed a significant increase in PS1/2−/− fibroblasts compared with WT fibroblasts. n = 3. **p < 0.01, by unpaired two-tailed t test. All quantitative data are shown as the mean ± SEM. Formation of γ-secretase complex and cellular levels of NCT, PEN-2, and APH-1 in WT, PS1/2−/−, PS1−/−, and PS2−/− fibroblasts are shown in Extended Data Figure 1-1.
Figure 1-1
PS γ-secretase complex formation and expression levels of NCT, PEN-2, and APH-1 in WT, PS1/2–/–, PS1–/–, and PS2–/– fibroblasts. A, A total of 5 μg of protein from cell lysates of WT, PS1/2–/–, PS1–/–, and PS2–/– fibroblasts was subjected to Blue native PAGE (BN-PAGE) and analyzed by immunoblotting using PS1-CTF and PS2-CTF antibodies. B, C, Quantification of PS1 and PS2 γ-secretase complex levels from the experiment shown in A. B, PS2–/– fibroblasts showed a significant decrease in mature PS1 γ-secretase complex levels. C, PS1–/– cells showed a significant increase in PS2 γ-secretase complex levels. n = 3. *p < 0.05, ****p < 0.0001, by one-way ANOVA followed by Tukey's multiple-comparison tests. D, Expression levels of NCT, PEN-2, and APH-1 in WT, PS1/2–/–, PS1–/–, and PS2–/– fibroblasts were determined by immunoblotting. Tubulin served as a loading control. E, F, Quantification of mature and immature NCT. Mature NCT was completely abolished in PS1/2–/– fibroblasts and significantly decreased in PS1–/– fibroblasts compared with WT fibroblasts. The level of immature NCT was significantly increased in PS1/2–/– fibroblasts compared with WT fibroblasts. n = 3. *p < 0.05, ***p < 0.001, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple comparison tests. G, H, Quantification of PEN-2 and APH-1 expression levels from the experiment shown in D. PEN-2 expression was completely abolished in PS1/2–/– fibroblasts and significantly decreased in PS1–/– fibroblasts compared with WT fibroblasts. APH-1 expression was significantly decreased in PS2–/– fibroblasts compared with WT fibroblasts. n = 3. *p < 0.05, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple comparison tests. All data are shown as the mean ± SEM. Download Figure 1-1, EPS file.
To gain mechanistic insights into the lack of ApoE secretion by PS-deficient cells, we investigated whether the subcellular distribution of ApoE changed in these cells. We found that nuclear localization of ApoE was markedly enhanced by PS1 deficiency or PS1/2 deficiency. Correspondingly, cytosolic ApoE levels were dramatically lower in PS1/2−/− and PS1−/− fibroblasts compared with WT fibroblasts (Fig. 1F–H). We corroborated these findings through subcellular fractionation experiments. Nuclear fractionation was assessed by the presence of PARP, a nuclear marker. The nuclear/cytosolic ApoE ratio was 2.8-fold greater in PS1/2−/− fibroblasts than in WT fibroblasts (Fig. 1I–K). These results suggest that abnormal nuclear translocation of ApoE in PS1/2−/− and PS1−/− cells may impair ApoE secretion.
We also measured the cellular levels of other γ-secretase components. As has been reported, depletion of both PS1 and PS2 completely abolished NCT maturation and PEN-2 cellular levels, whereas depletion of PS2 did not affect NCT maturation or PEN-2 levels, and PS2 deficiency caused only a strong reduction in APH-1 levels (Steiner et al., 2002; Zhang et al., 2005; Extended Data Fig. 1-1D–H).
Inhibition of γ-secretase activity reduces ApoE secretion in vitro and in vivo
Consistent with the findings in PS-deficient fibroblasts, we also found that blockade of γ-secretase activity using DAPT at 10 μm significantly decreased ApoE secretion by both primary cultured embryonic WT astrocytes (Fig. 2A,B) and fibroblasts (Extended Data Fig. 2-1A,B). In contrast to PS1/2−/− fibroblasts, the cellular levels of the γ-secretase components PS1-CTF, mature NCT, PEN-2, and APH-1 were significantly increased in DAPT-treated cells (Sogorb-Esteve et al., 2018; Fig. 2A,D,F–I, Extended Data Fig. 2-1A,D,F–I). Blockade of γ-secretase activity was confirmed by measuring Aβ levels and notch intracellular domain (NICD) generation (Fig. 2F,J, Extended Data Fig. 2-1F,J,N,O). Similar to PS1/2−/− fibroblasts, the inhibition of γ-secretase activity significantly reduced cytosolic ApoE levels and increased those of nuclear ApoE (Fig. 2K–M, Extended Data Fig. 2-1K–M). These results strongly suggest that ApoE secretion depends on γ-secretase activity, regardless of the levels of the proteins that constitute this activity. The in vivo effects of γ-secretase inhibition were examined by treating WT mice with DAPT via intraperitoneal injection. ApoE levels in CSF were significantly decreased in DAPT-treated mice compared with control mice (Fig. 2N,O). Total protein in CSF did not show any change in DAPT-treated mice compared with control mice (Fig. 2P). Aβ40 and Aβ42 levels in the cortex of DAPT-treated mice were also significantly decreased compared with control mice (Fig. 2Q,R). These results suggest that decreased γ-secretase activity reduces ApoE secretion in vivo.
Inhibition of γ-secretase activity reduces ApoE secretion, promotes cytosol-to-nuclear translocation of ApoE in primary cultured embryonic WT astrocytes, and reduces ApoE levels in the CSF of DAPT-treated mice. A, Primary cultured embryonic WT astrocytes were treated with 1–10 μm of DAPT or DMSO vehicle control for 48 h in serum-free conditional medium. Levels of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF were determined by immunoblotting. B–E, Quantification of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF levels from the experiment shown in A. n = 3; *p < 0.05, ***p < 0.001, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. F, Immunoblot analysis of NCT, PEN-2, APH-1, and NICD in primary cultured embryonic WT astrocytes treated with DAPT. G–J, Quantification of mature NCT, PEN-2, APH-1, and NICD expression levels from the experiment shown in F, normalized to α-tubulin protein levels. J, Decreased NICD expression indicates inhibition of γ-secretase activity. n = 3; *p < 0.05, **p < 0.01. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. K, Immunostaining for cellular distribution of ApoE (red) and DAPI staining (nuclei, blue) in primary cultured embryonic WT astrocytes treated with DAPT for 24 h. Scale bars, 50 μm. L, M, Quantification of cytosolic and nuclear ApoE intensity from the experiment shown in K. Cytosolic ApoE was significantly decreased and nuclear ApoE was significantly increased in 10 μm of DAPT-treated astrocytes compared with control cells. n ≥ 24 different stained cells/group. **p < 0.01, ***p < 0.001, by unpaired two-tailed t tests. N, Western blot analysis of ApoE levels in CSF from WT C57BL/6J mice treated with DAPT or DMSO vehicle control. O, Quantification of CSF ApoE levels from the experiment shown in N. DAPT treatment significantly decreased CSF ApoE levels in WT C57BL/6J mice compared with control mice. n = 6 mice/group. *p < 0.05, by unpaired two-tailed t test. P, Quantification of total protein in CSF. n = 6 mice/group. NS, Not significant, by unpaired two-tailed t test. Q, R, Aβ40 and Aβ42 levels in the cortex of mice measured using an Aβ40 or Aβ42 ELISA kit. P, Q, DAPT treatment significantly decreased the levels of Aβ42 (P) and Aβ40 (Q). n = 6 mice/group. *p < 0.05, ****p < 0.0001, by unpaired two-tailed t test. All quantitative data shown as the mean ± SEM. Primary cultured embryonic WT fibroblasts were also treated with DAPT, and similar results are shown in Extended Data Figure 2-1.
Figure 2-1
Inhibition of γ-secretase activity reduces ApoE secretion and promotes cytosol-to-nuclear translocation of ApoE in primary cultured embryonic WT fibroblasts. A, Primary cultured embryonic WT fibroblasts were treated with 1–10 μm DAPT or DMSO vehicle control for 48 h in serum-free conditional medium. Levels of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF were determined by immunoblotting. B–E, Quantification of the levels of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF from the experiment shown in A. ApoE secretion was significantly decreased in primary cultured embryonic WT fibroblasts treated with 10 μm DAPT compared with control cells. n = 3. *p < 0.05. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. F, Immunoblot analysis of NCT, PEN-2, APH-1, and NICD in primary cultured embryonic WT fibroblasts treated with DAPT. G–J, Quantification of mature NCT, PEN-2, APH-1, and NICD from the experiment shown in F. J, Decreased level of NICD expression indicates inhibition of γ-secretase activity. n = 3. *p < 0.05, ***p < 0.001, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. K, Immunostaining for cellular distribution of ApoE (red) and DAPI staining (nuclei, blue) in primary cultured embryonic WT fibroblasts treated with 10 μm DAPT. Scale bars, 5 μm. L, M, Quantification of cytosolic and nuclear ApoE intensity from the experiment shown in K. Cytosolic ApoE was significantly decreased and nuclear ApoE was significantly increased in 10 μm DAPT-treated fibroblasts compared with control cells. n = 20 different stained cells/group. ***p < 0.001, by unpaired two-tailed t-tests. N, O, Quantification of Aβ42 (N) and Aβ40 (O) generation in APP-overexpressing fibroblasts treated with DAPT for 48 h. DAPT treatment significantly decreased Aβ40 and Aβ42 generation. n = 4. ****p < 0.0001, by unpaired two-tailed t tests. All data are shown as the mean ± SEM. Download Figure 2-1, EPS file.
Reduced ApoE secretion in NCT−/− fibroblasts
Because PS1 has multiple functions in addition to γ-secretase activity (Donoviel et al., 1999), we investigated whether the loss of γ-secretase activity impaired ApoE secretion in NCT−/− fibroblasts. We also found that ApoE secretion significantly decreased, but was not completely abolished, in NCT−/− fibroblasts (Fig. 3A,B), possibly because PS1-CTF and the γ-secretase complex (PS1 and APH-1) were still present (Fig. 3A,D,J,K,N) and γ-secretase activity was still detectable in these cells (Li et al., 2003b). NCT deficiency also resulted in a decrease in the molecular weight of both secreted and intracellular ApoE (Fig. 3A), suggesting that NCT is involved in O-linked glycosylation of ApoE (Wernette-Hammond et al., 1989; Flowers et al., 2020). We also confirmed that NCT deficiency dramatically reduced NICD generation and the levels of PEN-2 and APH-1 (Zhao et al., 2010; Fig. 3F–I). Similar to PS-deficient cells, cytosolic ApoE levels significantly decreased and nuclear ApoE levels significantly increased in NCT−/− fibroblasts compared with WT fibroblasts (Fig. 3O–Q). These results suggest that the loss of γ-secretase activity increases abnormal translocation of ApoE from the cytosol to the nucleus and impairs ApoE secretion.
Reduced ApoE secretion and increased cytosol-to-nuclear translocation of ApoE in NCT−/− fibroblasts. A, Immunoblot analysis of secreted and intracellular ApoE, NCT, PS1, and PS2 in WT and NCT−/− fibroblasts. B–E, Quantification of secreted ApoE, intracellular ApoE, PS1-CTF, and PS2-CTF expression levels shown in A. B, D, ApoE secretion and PS1-CTF levels were significantly decreased in NCT−/− fibroblasts. Intracellular ApoE, PS1-CTF, and PS2-CTF expression levels were normalized to α-tubulin protein levels. n = 3. **p < 0.01, ****p < 0.0001. NS, Not significant, by unpaired two-tailed t test. F, Immunoblot analysis of PEN-2, APH-1, and NICD levels in WT and NCT−/− fibroblasts. G–I, Quantification of the levels of PEN-2, APH-1, and NICD from the experiment shown in F. Levels of PEN-2, APH-1, and NICD were significantly decreased in NCT−/− fibroblasts compared with WT fibroblasts. PEN-2, APH-1, and NICD expression levels were normalized to α-tubulin protein levels. n = 3. **p < 0.01, ***p < 0.001, ****p < 0.0001, by unpaired two-tailed t test. J, A total of 5 μg of protein from cell lysates of WT and NCT −/− fibroblasts was subjected to Blue native PAGE (BN-PAGE) and analyzed by immunoblotting using PS1-CTF, PS2-CTF, PEN-2, and APH-1 antibodies. K–N, Quantification of PS1, PS2, PEN-2, and APH-1 γ-secretase complex formation from the experiment shown in J. PS1 and APH-1 γ-secretase complex formation was significantly decreased in NCT −/− fibroblasts compared with WT cells. n = 3. **p < 0.01, ***p < 0.001, ****p < 0.0001, by unpaired two-tailed t test. O, Immunostaining for cellular distribution of ApoE (red) and DAPI staining (nuclei, blue) in WT and NCT−/− fibroblasts. Scale bars, 10 μm. P, Q, Quantification of cytosolic and nuclear fluorescence intensity from the experiment shown in O. Cytosolic ApoE was significantly decreased and nuclear ApoE significantly increased in NCT−/− fibroblasts compared with WT fibroblasts. Nuclear boundaries were determined based on DAPI. n = 20 different stained cells/group, **p < 0.01, ***p < 0.001, by unpaired two-tailed t test. All data are shown as the mean ± SEM.
Fully restoration of ApoE secretion by WT PS1 but not by PS1 mutants in PS1/2−/− fibroblasts
We also examined whether PS1 expression restores ApoE secretion in PS1/2−/− fibroblasts. PS1/2−/− fibroblasts were infected with retrovirus vectors expressing human WT PS1 or human PS1 bearing FAD-associated mutations (PS1ΔE9, PS1L166P, or PS1G384A). WT PS1 restored ApoE secretion in PS1/2−/− fibroblasts to a level similar to that of WT fibroblasts, whereas PS1ΔE9, PS1L166P, and PS1G384A PS1 mutants restored ApoE secretion to a lesser degree than WT PS1 (Fig. 4A,B). After infection with the retrovirus vectors, the similarity of levels of WT PS1 and PS1 mutants in PS1/2−/− fibroblasts was confirmed (Fig. 4A). Instead of PS1-CTF, full-length PS1 was detected in PS1/2−/− cells infected with PS1ΔE9, because PS1ΔE9 lacks an exon-containing endocleavage site (Steiner et al., 1999). WT PS1 dramatically restored the cytosolic localization of ApoE in PS1/2−/− fibroblasts, whereas the PS1ΔE9, PS1L166P, and PS1G384A PS1 mutants exhibited no significant effects (Fig. 4C,D). We also found that WT PS1 restored ApoE secretion and increased cytosolic localization of ApoE in PS1/2−/− fibroblasts in a dose-dependent manner accompanied by reduced nuclear localization of ApoE (Extended Data Fig. 4-1), suggesting that ApoE secretion is positively correlated with WT PS1 levels and cytosolic localization of ApoE.
Human PS1 bearing FAD mutations restored from ApoE secretion to a lesser degree than WT PS1. A, Immunoblot analysis of secreted and intracellular ApoE, PS1, and PS2 in WT fibroblasts and PS1/2−/− fibroblasts infected with retrovirus vectors bearing human WT PS1 or PS1 with FAD mutations (PS1ΔE9, PS1L166P, and PS1G384A). B, Quantification of secreted ApoE levels. The level of ApoE secreted from WT fibroblasts was adjusted to 1. n = 3. *p < 0.05, ***p < 0.001, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. C, Confocal microscopy analysis of cytosolic and nuclear localization of ApoE. Nuclei were stained with DAPI. Scale bars, 2 μm. D, Quantification of cytosolic intensity of ApoE shown in C. WT PS significantly restored the cytosolic distribution of ApoE in PS1/2−/− fibroblasts, whereas PS1ΔE9, PS1L166P, and PS1G384A did not show significant restoration. n = 20 different stained cells/ group. **p < 0.01. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. All quantitative data are shown as the mean ± SEM. Overexpression of WT PS1 enhanced ApoE secretion in PS1/2−/− fibroblasts and primary cultured embryonic ApoE3-KI fibroblasts (Extended Data Figures 4-1, 4-2).
Figure 4-1
ApoE secretion and cytosolic distribution are positively correlated with WT PS1 expression levels. A, PS1/2–/– fibroblasts were transiently infected with 0.5, 1, or 2.5 ml of retrovirus vector expressing human WT PS1 or 2.5 ml of retroviral vector control (Mock). Levels of secreted and intracellular ApoE, PS1, PS2, PEN-2, and APH-1 were determined by immunoblotting. B–F, Quantification of secreted and intracellular ApoE, PS1-CTF, PEN-2, and APH-1 expression levels from the experiment shown in A. WT PS1 dose-dependently increased ApoE secretion in PS1/2–/– fibroblasts. Intracellular ApoE, PS1-CTF, PEN-2, and APH-1 expression levels were normalized to α-tubulin protein levels. n = 3. *p < 0.05, **p < 0.01. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. G, Immunostaining for cellular distribution of ApoE (red) and DAPI staining (nuclei, blue) after the expression of WT PS1 in PS1/2–/– fibroblasts. Scale bars, 2 μm. H, I, Quantification of cytosolic and nuclear ApoE intensity from the experiment shown in G. Cytosolic ApoE intensity is positively, and nuclear ApoE intensity negatively, related to the expression level of WT PS1 in PS1/2–/– fibroblasts. n ≥ 20 different stained fibroblasts/group. ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. All data are shown as the mean ± SEM. Download Figure 4-1, EPS file.
Figure 4-2
Overexpression of human WT PS1 enhances ApoE secretion. Primary cultured embryonic ApoE3-KI fibroblasts (embryonic, 18 d old) were transiently infected with retrovirus vectors expressing human WT PS1 or WT PS2 and PS mutants (PS1ΔE9, PS1L166P, PS1G384A, and PS2N141I) as indicated or retroviral vector control (Mock). A, Levels of secreted and intracellular ApoE, PS1, and PS2 were determined by immunoblotting. B–E, Quantification of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF expression levels from the experiment shown in A. A, B, Overexpression of WT PS1 significantly enhanced ApoE secretion and intracellular ApoE expression, whereas PS1ΔE9, PS1L166P, and PS1G384A did not show any effects on ApoE secretion or expression levels compared with control vector. n = 3. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. F, Immunostaining for cellular distribution of ApoE (red) and DAPI staining (nuclei, blue) after overexpression of WT PS1 or PS1 mutants (PS1ΔE9, PS1L166P, and PS1G384A) in primary cultured embryonic ApoE3-KI fibroblasts. Scale bars, 2 μm. G, H, Quantification of cytosolic and nuclear ApoE intensity from the experiment shown in F. G, H, Cytosolic ApoE significantly increased and nuclear ApoE significantly decreased after overexpression of WT PS1, whereas overexpression of PS1ΔE9, PS1L166P, and PS1G384A did not affect cellular ApoE localization compared with Mock. n = 20 different stained cells/group. *p < 0.05, **p < 0.01. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. I, Immunoblotting of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF. J–M, Quantification of secreted and intracellular ApoE, PS1-CTF, and PS2-CTF expression levels from the experiment shown in I. Overexpression of WT PS2 and PS2N141I did not affect ApoE secretion or expression levels in primary cultured embryonic ApoE3-KI fibroblasts. n = 3. *p < 0.05. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. N, Immunostaining for cellular distribution of ApoE (red) and DAPI staining (nuclei, blue) after infection with WT PS2 and PS2N141I in primary cultured embryonic ApoE3-KI fibroblasts. Scale bars, 2 μm. O, P, Quantification of cytosolic and nuclear ApoE intensity from the experiment shown in N. WT PS2 and PS2N141I did not affect the cytosolic and nuclear distribution of ApoE. n = 20 different stained cells/group. NS, Not significant, by one-way ANOVA followed by Tukey's multiple-comparison tests. All data are shown as the mean ± SEM. Download Figure 4-2, EPS file.
We also examined whether PS1 or PS2 mutants exert a dominant-negative effect on ApoE secretion in primary cultured embryonic ApoE3-KI fibroblasts, which secrete more ApoE than normal WT fibroblasts. Overexpression of WT PS1 increased ApoE secretion and cytosolic ApoE localization, accompanied by decreased nuclear localization of ApoE (Extended Data Fig. 4-2A,B,F–H). However, PS1ΔE9, PS1L166P, PS1G384A, WT PS2, and PS2N141I exhibited no significant effect on ApoE secretion or the subcellular localization of ApoE in ApoE3-KI fibroblasts (Extended Data Fig. 4-2A,B,F–H,I,J,N–P). These results suggest that if WT PS or γ-secretase activity persists at normal levels, PS mutations would have no dominant-negative effect on ApoE secretion, and that WT PS1 has stronger effects in promoting ApoE secretion and its cytosolic localization than WT PS2.
Reduced plasma ApoE levels in FAD patients carrying PS1 mutations compared with normal control subjects
To determine whether reduced γ-secretase activity affects ApoE levels in vivo, we investigated ApoE levels in human plasma from FAD patients carrying PS1 mutations and age-matched healthy control subjects. Previous reports showed that these PS1 mutations result in decreased or altered γ-secretase activity (Jørgensen et al., 1996; Yasuda et al., 1999; Hamaguchi et al., 2009; Liu et al., 2017; Sun et al., 2017). There was no significant difference in mean age between the FAD patients (51.04 ± 2.30 years) and controls (55.58 ± 1.96 years; p = 0.1469). Plasma ApoE levels in FAD patients with PS1 mutations were significantly reduced compared with healthy control subjects (Fig. 5A). There was no significant correlation between plasma ApoE level and age among the control, PS1 mutation, and combined groups (Fig. 5B–D). These results suggest that decreased γ-secretase activity induced by PS1 mutations negatively regulates ApoE secretion.
Reduced ApoE levels in the plasma of FAD patients carrying pathogenic PS1 mutations. A, ApoE levels in plasma of control individuals and FAD patients carrying PS1 mutations were determined using an ELISA kit. Plasma ApoE levels were significantly lower in FAD patients with PS1 mutations than control subjects. Control subjects, n = 12; FAD patients, n = 12. *p < 0.05, by unpaired two-tailed t test. B–D, There was no statistically significant correlation between plasma ApoE level and age. p-Values were determined by Spearman rank test. All data are shown as the mean ± SEM.
PS deficiency increases binding of importin α4 to ApoE and enhances nuclear colocalization of importin α4/ApoE in PS1/2−/− fibroblasts
The importin α family of multipurpose nuclear transport receptors mediate the nuclear translocation of many proteins, including NICD, signal transducer and activator of transcription 1 (STAT1), and STAT2 (Goldfarb et al., 2004; Huenniger et al., 2010). Importin α4 exhibits the strongest affinity for NICD, whereas importin α3 exhibits moderate affinity for NICD (Huenniger et al., 2010). Interestingly, levels of importin α4, but not of importin α3, were significantly increased in PS1/2−/− fibroblasts in the present study, and NICD generation was completely abolished in PS1/2−/− fibroblasts (Fig. 6A–D). We then examined potential interactions between importin α3 or α4 and ApoE using IP assays and found that importin α4, but not importin α3, bound to ApoE. In addition, ApoE bound more importin α4 in PS1/2−/− fibroblasts than in WT fibroblasts (Fig. 6E,F). Confocal microscopy studies demonstrated that the colocalization of importin α4 and ApoE was significantly higher in the nucleus of PS1/2−/− fibroblasts (Fig. 6G,H). These results suggest that nuclear transport of ApoE is mediated by importin α4 and may be enhanced by the lack of NICD in PS1/2−/− cells.
Increased binding of importin α4 to ApoE and enhanced nuclear colocalization of importin α4/ApoE in PS1/2−/− fibroblasts. A, Immunoblot analysis of importin α3, importin α4, and NICD in WT and PS1/2−/− fibroblasts cultured for 48 h. B–D, Quantification of importin α3, importin α4, and NICD levels from the experiments shown in A, normalized to α-tubulin protein levels. n = 3. *p < 0.05. NS, Not significant, by unpaired two-tailed t test. E, F, Immunoprecipitation analysis of the binding of ApoE to importin α3 or importin α4. ApoE bound more importin α4 in PS1/2−/− fibroblasts than in WT fibroblasts. n = 3. **p < 0.01, by unpaired two-tailed t test. G, Immunostaining analysis showed that importin α4 colocalized with ApoE in the cytosol of WT fibroblasts, whereas the colocalization was found predominantly in the nucleus of PS1/2−/− fibroblasts. Scale bars, 50 μm. WT and PS1/2−/− fibroblasts were stained with secondary antibodies labeled with Alexa Fluor 488 and Alexa Fluor 568: ApoE (green), importin α4 (red). H, Quantification of fluorescence intensity of colocalized ApoE/importin α4 in nucleus from the experiment shown in G. n = 21 different stained cells/group. ****p < 0.0001, by unpaired two-tailed t test. All quantitative data are shown as the mean ± SEM.
Discussion
Abnormal Aβ generation and an increased Aβ42/Aβ40 ratio induced by PS1 mutations have long been considered causative factors for FAD (Sun et al., 2017). Here, we demonstrated a novel function of PS; that is, PS is essential for ApoE secretion, and decreased γ-secretase activity induced by PS mutations reduces ApoE secretion. We demonstrated that PS-deficient cells completely abolished ApoE secretion. Furthermore, we showed that the blockade of γ-secretase activity by a γ-secretase inhibitor, DAPT, decreased ApoE secretion both in astrocytes and in CSF, suggesting that γ-secretase activity is crucial for ApoE secretion in brain. A mature γ-secretase consists of the following four components: PS, NCT, APH-1, and PEN-2. These four core components of the γ-secretase complex tightly regulate each other's expression, maturation, and assembly as well as γ-secretase activity (Takasugi et al., 2003; Capell et al., 2005). To corroborate the involvement of other γ-secretase molecule, we demonstrated that NCT deficiency reduced ApoE secretion, but its secretion was not completely abolished, possibly because PS function or γ-secretase activity was still present in NCT-deficient cells. These results suggest that NCT partially regulates ApoE secretion and PS plays the key role in ApoE secretion. We also observed an altered post-translational modification of ApoE in NCT-deficient cells (Fig. 3A). Relatedly, an aberrant migration of ApoE was also found in the intracellular fraction of PS1−/− cells (Fig. 1A). Previous study reported that NCT is required for processing and trafficking of APP in mammals and secretion of Aβ peptides is abolished in NCT-deficient fibroblasts, suggesting a role of NCT in secretory pathway (Li et al., 2003a). Glycosylation occurs in endoplasmic reticulum-to-Golgi apparatus very early in the secretory pathway. Nonetheless, the role of NCT and PS deficiency in ApoE glycosylation in the secretory pathway needs further investigation.
PS mutations lead to the loss of function of γ-secretase, which was confirmed by decreased cleavage of Notch, syndecan, and N-cadherin, and this can probably be extended to other substrates (De Strooper, 2007). Importantly, FAD patients carrying pathogenic PS1 mutations exhibited lower plasma ApoE levels than control subjects. In contrast to the changes in Aβ generation associated with FAD, decreased Aβ clearance is crucial for the pathogenesis of SAD, and ApoE plays a major role in brain Aβ clearance (Scheltens et al., 2016). For example, SAD patients carrying the ApoE4 allele show decreased Aβ clearance in the CSF compared with carriers of ApoE3 or ApoE2 allele (Castellano et al., 2011; Yamazaki et al., 2019). Together, these findings suggest that loss of PS function alters the generation of Aβ and impairs ApoE secretion, both of which may initiate and accelerate Aβ accumulation, ultimately leading to the development of AD.
PS associated with the trafficking of various proteins, including NCT, APP, triggering receptor expressed on myeloid cells 2, integrin β1, and ApoE receptor 2 (Zou et al., 2008; De Strooper et al., 2012). Our results suggest that PS or γ-secretase components may also be involved in ApoE trafficking; however, γ-secretase activity, regardless of the levels of γ-secretase components, is crucial for ApoE secretion (Fig. 2A,B,D,F–I, Extended Data Fig. 2-1A,B,D,F–I). ApoE is a 34 kDa glycoprotein that is highly expressed and secreted by various cell types (Hasel and Liddelow, 2021). Although ApoE is predominantly present in the cytosol, its presence in the nuclear compartment has also been reported (Theendakara et al., 2016). A recent study demonstrated that both ApoE3 and ApoE4 bound to SirT1 promoter sequence in the nucleus with similar binding affinities and downregulated the transcription of SirT1, a predominantly nuclear protein. In addition to the overall reduction in SirT1 protein expression, only ApoE4 caused the redistribution of SirT1 from the nucleus to the cytosol (Theendakara et al., 2016). The nuclear roles of ApoE are given additional significance with the demonstrations that ApoE4 specifically binds certain DNA sequences, namely “CLEAR” (coordinated lysosomal expression and regulation) elements. This connects the phenomenon of nuclear ApoE to negative outcomes, specifically the growing interest in autophagic failure in AD (Parcon et al., 2018; Lima et al., 2020). Furthermore, ApoEs are shown to be associated with the transcription of MADD (MAP kinase-activating death domain) and COMMD6 genes, which further bind to the nuclear factor-κB complex or the RelA subunit of the complex to regulate transcription (Del Villar and Miller, 2004; de Bie et al., 2006; Theendakara et al., 2016). These findings suggest that ApoE can translocate to the nucleus and recycle back to the cytosol. However, the mechanism by which ApoE escapes the secretory pathway and translocates to the nucleus from cytosol is currently unclear. Importin α family proteins are multipurpose nuclear transport receptors that bind to hundreds of proteins, including NICD, linking them to importin β to facilitate their translocation from the cytosol to the nucleus (Goldfarb et al., 2004; Huenniger et al., 2010). Here, we report that ApoE binds to importin α4 to facilitate the translocation of ApoE from the cytosol to the nucleus in PS-deficient cells, suggesting that abnormal translocation of ApoE from the cytosol to the nucleus attenuates ApoE secretion. ApoE has been shown to be a transcriptional regulator of many genes that are potentially involved in AD (Theendakara et al., 2016). Although the significance of the enhanced nuclear localization of ApoE in PS-deficient cells is not clear, it is possible that enhanced nuclear ApoE may suppress or activate the transcription of some genes that are regulated by PS or NICD.
Together, this study demonstrated a novel link of PS to ApoE, tying together the mechanisms associated with FAD and SAD. PS mutations and ApoE4 have long been considered to cause FAD and SAD by two independent mechanisms, abnormal Aβ generation and impaired Aβ clearance, respectively. For the first time, we demonstrated that PS is essential for the secretion of ApoE, suggesting that PS mutations or altered γ-secretase activity is also involved in Aβ clearance via regulation of ApoE secretion. Our findings may offer opportunities for the development of novel AD treatment targets.
Footnotes
This work was supported by the Grant-in-Aid for Scientific Research B 16H05559 (to M.M.), Challenging Exploratory Research Grant 15K15712 (to M.M.), and Grant-in-Aid for Scientific Research C 19K07846 (to K.Z.) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan. This work was also supported by “Project of Translational and Clinical Research Seed A” Grant A-128 (to M.M.); “Research and Development Grants for Dementia” Grants AMED 21dk0207050h0002 (to M.M.), AMED JP21dk0207049, and JP21dk0207045 (to T.I.); and Grants-in-aid from “The 24th General Assembly of the Japanese Association of Medical Sciences” (to K.Z.), the Daiko Foundation (to K.Z.), and the Hirose International Scholarship Foundation (to K.Z.). S.I. was supported by the Ministry of Education, Culture, Sports, Science, and Technology, Japan. We thank Dr. Bart De Stropper for providing wild-type, PS double knock-out (PS1/2–/–), PS1 knock-out (PS1–/–), and PS2 knock-out (PS2–/–) mouse embryonic fibroblast cells.
The authors declare no competing financial interests.
- Correspondence should be addressed to Kun Zou at kunzou{at}med.nagoya-cu.ac.jp or Makoto Michikawa at michi{at}med.nagoya-cu.ac.jp
