The deposition of amyloid-β (Aβ) peptide, which is generated from amyloid precursor protein (APP), is the pathological hallmark of Alzheimer's disease (AD). Three APP familial AD mutations (D678H, D678N, and H677R) located at the sixth and seventh amino acid of Aβ have distinct effect on Aβ aggregation, but their influence on the physiological and pathological roles of APP remain unclear. We found that the D678H mutation strongly enhances amyloidogenic cleavage of APP, thus increasing the production of Aβ. This enhancement of amyloidogenic cleavage is likely because of the acceleration of APPD678H sorting into the endosomal-lysosomal pathway. In contrast, the APPD678N and APPH677R mutants do not cause the same effects. Therefore, this study indicates a regulatory role of D678H in APP sorting and processing, and provides genetic evidence for the importance of APP sorting in AD pathogenesis.
The internalization of amyloid precursor protein (APP) increases its opportunity to be processed by β-secretase and to produce Amyloid-β (Aβ) that causes Alzheimer's disease (AD). We report a pathogenic APPD678H mutant that enhances APP internalization into the endosomal-lysosomal pathway and thus promotes the β-secretase cleavage and Aβ production. This study provides genetic evidence for the importance of APP sorting in AD pathogenesis.
Alzheimer's disease (AD) is the most common form of age-dependent neurodegenerative disorder (Wimo et al. 2007). Abnormal accumulation of amyloid-β (Aβ) is one of the pathological hallmarks of AD and plays an important role in pathogenesis (Selkoe 2002; Haass and Selkoe 2007). Aβ is generated from a single-transmembrane protein named amyloid precursor protein (APP). Matured APP can be cleaved by multiple secretases in two distinct pathways (Fig. 1a) (Haass 2004). In the non-amyloidogenic pathway, α-secretase cleavage of APP within Aβ region generates sAPPα and C83 fragments, thus precluding Aβ formation. In the amyloidogenic pathway, β-secretase cleavage of APP generates sAPPβ and C99 fragments (Vassar et al. 1999; Wilson et al. 1999; Dingwall 2001). Subsequent γ-secretase cleavage of C99 produces multiple varieties of Aβ peptides, the two most common containing 40 amino acids (Aβ40) and 42 amino acids (Aβ42) (Wolfe et al. 1999; Sastre et al. 2001). Compared to Aβ40, Aβ42 is more aggregate-prone and is more pathogenic (Jarrett et al. 1993).
One of the factors to regulate Aβ production is the sorting of surface APP into different cellular compartments containing different secretases (Small and Gandy 2006; Thinakaran and Koo 2008). APP is trafficked through the constitutive secretary pathway, where it undergoes post-translational modifications (Rajendran and Annaert 2012). Once reaching the cell surface, mature APP usually internalizes rapidly via clathrin-dependent endocytosis (Xiao et al. 2012). Internalized APP is capable of transporting to endosomes, recycle back to Golgi and cell surface, or undergo degradation in the lysosomes (Thinakaran and Koo 2008; Lorenzen et al. 2010). Higher α-secretase activity is found primarily at the cell surface, whereas higher β-secretase activity is found in the Golgi and endosomes (Thinakaran and Koo 2008; O'Brien and Wong 2011). Sorting of APP into different subcellular compartments alters the possibility of APP to encounter different secretases.
Although familial AD (FAD) individuals represent < 5% of all AD cases, studying FAD mutations may provide critical insight into AD pathogenesis. Most FAD-causing APP mutations are located at either flanking or center regions (Aβ21–23) of the Aβ, and can interfere with Aβ production, aggregation or removal (Tanzi 2012). In addition to these regions, three FAD APP mutations have been identified – the D678H, D678N, and H677R substitution (Fig. 1a, APP770 numbering) (Janssen et al. 2003; Wakutani et al. 2004; Chen et al. 2012; Lan et al. 2014) – which are located at the Aβ-N-terminal (sixth and seventh amino acid). According to previous reports, D678N and H677R mutations do not affect Aβ production, with D678H mutation increasing Aβ production, suggesting that the Aβ-N-terminal mutations interfere with Aβ production differently (Hori et al. 2007; Chen et al. 2012). In this study, we aim to investigate the mechanism of these three mutations in regulating APP processing.
Materials and methods
A cDNA encoding human APP cDNA was subcloned into an entry vector pDONR221 (Chen et al. 2012). A Quickchange II mutagenesis kit (200519-4; Stratagene, La Jolla, CA, USA) was used to generate APPH677R, APPD678H and APPD678N mutants. Entry clones (APP-pDONR221) were transferred into cytomegalovirus promoter/enhancer-driven expression vector (pDEST26) via site-specific recombinations according to the Gateway cloning protocol (11791-019; Invitrogen, Carlsbad, CA, USA).
HEK293 (Human embryonic kidney 293) cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 2 mM l-glutamine at 37°C with 5% CO2. Plasmids carrying different APP cDNAs were transfected by Lipofectamine 2000 (11668-019; Invitrogen). All experiments were performed at 48 h after transfection unless specified. To monitor APP degradation, transfected cells were treated with 30 μg/mL cycloheximide (C7698; Sigma, St Louis, MO, USA) with or without 20 mM NH4Cl for 0, 3, and 6 h. Treated cells were lysed by incubating in lysis buffer (50 mM Tris–HCl, 150 mM NaCl, 5 mM EDTA, 0.5% Triton-X 100) containing protease inhibitor cocktail (04693116001; Roche, Basel, Switzerland) at 4°C with 10 min agitation, were centrifuged at 1550 g to remove debris. To confirm the C-terminal fragments (CTFs), 1 μM γ-secretase inhibitor L-685,458 (L1790; Sigma) was added into media 24 h after transfection.
APP fragments and Aβ measurement
To measure Aβ levels, 100 μL of media from APP expression cells culture were quantitated, using human β-Amyloid total and 42 enzyme-linked immunosorbent assay kits (27729 and 27711; IBL, Hamburger, Germany). To determine the levels of APP and its cleavage fragments, 30 μg of total protein from cell lysates were separated by 10% Tris-glycine or 15% Tris-Tricine sodium dodecyl sulfate–polyacrylamide gel electrophoresis, and analyzed with anti-APP N-terminus antibody 22C11 (MAB348; Millipore Corporation, Bedford, MA, USA) or anti-APP C-terminus antibody (AB5352; Millipore). To measure sAPPβ, media of APP expression cells, the culture was separated by 8% Tris-glycine sodium dodecyl sulfate–polyacrylamide gel electrophoresis and analyzed with anti-sAPPβ antibody (SIG-39138; Covance, Princeton, NJ, USA). All antibodies do not recognize Aβ-His6 and Aβ-Asp7 as an epitope.
Immunocytochemistry assay for APP internalization
Transfected cells were cultured on 24-well plates (2 × 104 cells/well) with poly-d-lysine-coated 12 mm glass coverslips. To determine APP internalization, unfixed cells were labeled with 22C11 antibody on ice for 1 h, allowed to internalize at 37°C for 15 min in the presence of 20 mM NH4Cl, and fixed by 4% paraformaldehyde. To determine transfection efficiency, cells were fixed immediately at 48 h after transfection and stained with antibody 22C11. These cells were blocked with 10% fetal bovine serum/0.6% Triton-X 100 in phosphate-buffered saline, incubated with Cy3-conjugated anti-mouse antibody (115165003; Jackson's Laboratory, Mount Desert Island, ME, USA), and stained by mounting gel containing 4',6-diamidino-2-phenylindole (H1200, VECTASHIELD+4',6-diamidino-2-phenylindole, H-1200; Vector Laboratories, Burlingame, CA, USA). All images were acquired by confocal microscope (FV-10i; Olympus, Tokyo, Japan) and analyzed by Metamorgh (Molecular Devices, Palo Alto, CA, USA).
Surface APP was labeled with sulfo-NHS biotin (89881, Pierce Cell surface protein isolation kit; Thermo, Waltham, MA, USA) for 30 min on ice, and allowed to internalize at 37°C for 30 min with 20 mM NH4Cl. Cells were stripped twice by fresh stripping buffer for 20 min (50 mM glutathione, 75 mM NaCl, 75 mM NaOH and 10 mM EDTA, pH = 7.5). After four washes, the cells were sonicated for five 1-s bursts at power 1.5 (Misonix Sonicator 3000; QSonica, Newtown, CT, USA) in the lysis buffer. After two 10 000 g centrifugations, half of the supernatant was collected as the total protein control and the other was passed through the NeutrAvidin agarose column to isolate biotin-labeled surface proteins.
The effect of Aβ-N-terminal mutations on secreted Aβ levels
Based on previous reports, only the APP D678H mutation, but not the D678N or H677R mutations, could increase the production of Aβ (Hori et al. 2007; Chen et al. 2012). Because these reports were performed by different groups, we first made side-by-side comparison of secreted Aβ levels from APPWT, APPD678H, APPD678N and APPH677R transfected cells. The total Aβ and Aβ42 levels in the media were examined by enzyme-linked immunosorbent assay, and the full-length APP level in the cell lysate was determined by western blot. The total secreted Aβ and Aβ42 levels were normalized to total APP, and the Aβ/APPWT ratios were arbitrarily set as one (Fig. 1b and c). Consistent with previous findings, only APPD678H expressing cell produced significantly more total Aβ and Aβ42 than APPWT among these three mutants.
The effect of Aβ-N-terminal mutations on APP cleavage
The higher Aβ level may be caused by the alteration of αβ-secretase cleavage ratio of APP (Fig. 1a). Thus, we examined the levels of CTFs, especially C83 and C99, in these cells. In western blot analysis, the cells expressing APPWT, APPD678N and APPH677R produced more C83 (~ 10 kDa), and the cells expressing APPD678H produced more C99 (~ 12 kDa) (Fig. 1d). To confirm that these bands were actually CTFs, we added 1 μM γ-secretase inhibitor, L-685,458, to APPWT and APPD678H-expressing cells. The levels of CTFs, especially C83, were increased in γ-secretase inhibitor treated cells relative to untreated cells (Fig. 1e). The C99/C83 ratio in APPD678H-expressing cells was significantly higher than in APPWT and in other APP mutant cells (Fig. 1f). However, the C99/C83 ratio in APPH677R and APPD678N-expressing cells had no statistical difference relative to APPWT. In addition, the level of the other β-secretase cleavage product, sAPPβ, was also significantly higher in the media of APPD678H-expressing cells than others (Fig. 1d and g). In conclusion, the D678H is the only mutation that can enhance amyloidogenic cleavage in this region.
The effect of Aβ-N-terminal mutations on APP internalization
In the amyloidogenic cleavage pathway, APP and β-secretase are internalized through different endocytic pathways, and sorted into the endosomes where the pH is conducive for β-secretase cleavage (Rajendran and Annaert 2012). Thus, we tested whether APPD678H has a higher tendency to internalize. The APP at the cell surface was labeled with antibody protein against the APP N-terminal for 1 h on ice, and then allowed to internalize at 37°C for 15 min (Fig. 2a, upper). The transfection efficiency of these cells had no significant difference among all APPs (Fig. 2a, lower). We counted the percent of APP-expressing cells with the majority APP internalized into the cells. APPWT, APPH677R and APPD678N mostly stay at the cell surface. However, the APPD678H had a higher tendency to internalize within 15 min than other APPs (Fig. 2b).
Moreover, we confirmed this finding using the surface protein biotinylation assay. Cell surface proteins on APPWT and APPD678H-expressing cells were labeled with biotin on ice and returned to 37°C to permit internalization. After stripping out the remaining biotins on the surface, the internalized proteins in the cell lysates were collected using the avidin column. Both total and eluted proteins were analyzed by western blot (Fig. 2c). We quantified the ratio of biotinylated APP to total APP and found that the APPD678H had a higher tendency to internalize as compared to APPWT (Fig. 2d). To confirm our stripping step, we removed all the surface biotinylated proteins, APPWT transfecting cells were labeled with biotin but did not permit internalization. No biotinylated APP could be detected after the stripping step (Fig. 2e).
The effect of Aβ-N-terminal mutations on lysosomal degradation of APP
The enhanced internalization of APPD678H suggested that this mutation may increase the chance of APP sorting into the endosomal-lysosomal pathway. Thus, we compared the stabilities of these APP mutants in the absence or presence of the lysosome inhibitor NH4Cl. Their stabilities were assayed by treating APP-expressing cells with protein synthesis inhibitor cycloheximide for 0, 3, and 6 h, and detecting the APP levels by western blotting. Without NH4Cl, the APPD678H level was significantly lower than APPWT and APPD678N at the third and sixth hour after treatment (Fig. 3a and b). The APPD678N level was lower than APPD678H but higher than APPWT and APPD677R levels only at the sixth but not the third hour after treatment (Fig. 3a and b). After treating with NH4Cl, the degradation of APPD678H became the same rate as other APPs (Fig. 3c and d). Our results indicated that APPD678H has the highest tendency to be degraded via lysosome among these four APPs.
This study demonstrated that D678H mutation on APP increases amyloidogenic cleavage and Aβ production, whereas D678N and H677R mutations do not. The difference on APP cleavage may be due to the increase of APPD678H internalized into the endosomal-lysosomal pathway. Thus, these three mutations with differential regulation of APP sorting contribute to AD pathogenesis through different mechanisms.
APPD678N did not perform exactly as APPWT did in our assays. The internalization of APPD678N had a trend but did not internalize significantly faster than APPWT (Fig. 2). Likewise, the degradation rate of APPD678N was slower than APPD678H but faster than APPWT and APPD677R (Fig. 3a and b). Therefore, APPD678N may also have a minor defect in sorting, but not as strong as APPD678H. This minor defect could not be detected in the short time-span (15 min) of our internalization experiment but was found in the degradation assay with a longer duration (6 h). However, this minor defect may not be sufficient to alter the cleavage ratio (Fig. 1).
One of the potential factors promoting the amyloidogenic cleavage of APPD678H is the enhancement of APP internalization through clathrin-dependent endocytosis. Multiple domains in APP have been demonstrated to regulate APP sorting. Within the APP intracellular sequence, the YENPTY motif is the major regulatory site for APP internalization and Aβ production (Perez et al. 1999). In addition to that, phosphorylation at S655 and Y678 enhances APP secretary trafficking and endocytosis, thus contributing to amyloidogenic cleavage (Rebelo et al. 2007; Vieira et al. 2009).
Although most regulatory domains in APP are intracellular, D678H mutation is located in the extracellular part of APP. This, however, is not the only extracellular domain in APP that is involved in sorting. For example, the Kunitz protease inhibitor domain in the extracellular part of APP751 or APP770 promotes sorting of APP to the plasma membrane and decreases Aβ formation in comparison to the Kunitz protease inhibitor-absent APP695 isoform (Ben Khalifa et al. 2012). Thus, it is possible that this extracellular D678H mutation regulates APP sorting, but understanding the precise mechanism would require further study.
Several risk factors for late-onset AD also regulate APP sorting, including: SORL1, SORCS1, SORCS2, BIN1, PICALM-1, and CD2AP (Rogaeva et al. 2007; Harold et al. 2009; Naj et al. 2011). Here, D678H is the first early onset FAD mutation that can interfere with the APP sorting. These findings provide genetic evidence that aberrant regulation of APP sorting may be a major factor in the etiology of both early and late onset of AD.
Acknowledgments and conflict of interest disclosure
This work is supported by Taiwan Ministry of Science and Technology grant (NSC 102-2320-B-010-021-MY2), National Health Research Institute (NHRI-EX98-9816NC), Cheng Hsin General Hospital (102F218C05), Yen Tjing Ling Medical Foundation (CI-102-4), Taipei Veterans General Hospital grant (V103E4-002), and Taiwan Ministry of Education Aim for Top University Grant.
All experiments were conducted in compliance with the ARRIVE guidelines. The authors have no conflict of interest to declare.