Inhibition of amyloid precursor protein processing by β-secretase through site-directed antibodies
- Department of Molecular Microbiology and Biotechnology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel
-
Communicated by Ephraim Katchalski-Katzir, Weizmann Institute of Science, Rehovot, Israel, March 23, 2005 (received for review September 26, 2004)
Abstract
Amyloid-β peptide (AβP) that accumulates in the Alzheimer's diseased brain is derived from proteolytic processing of the amyloid precursor protein (APP) by means of β- and γ-secretases. The β-secretase APP cleaving enzyme (BACE), which generates the N terminus of AβP, has become a target of intense research aimed at blocking the enzyme activity, thus reducing AβP and, subsequently, plaque formation. The search for specific inhibitors of β-secretase activity as a possible treatment for Alzheimer's disease intensified with the discovery that BACE may be involved in processing other non-APP substrates. The presence of the APP–BACE complex in early endosomes highlights the cell surface as a potential therapeutic target, suggesting that interference in APP–BACE interaction at the cell surface may affect amyloid-β production. We present here a unique approach to inhibit AβP production by means of antibodies against the β-secretase cleavage site of APP. These antibodies were found to bind human APP overexpressed by CHO cells, and the formed immunocomplex was visualized in the early endosomes. Indeed, blocking of the β-secretase site by these antibodies interfered with BACE activity and inhibited both intracellular and extracellular AβP formation in these cells.
Alzheimer's disease (AD) is characterized by the accumulation of senile plaques in the brain extracellular space and by intraneuronal accumulation of neurofibrillary tangles. The senile plaques are composed of deposited amyloid-β peptides (AβPs), which are derived from the enzymatic processing of a type I transmembrane protein called amyloid precursor protein (APP) (1). The β-secretase APP cleaving enzyme (BACE) generates the N terminus of the AβP peptide and produces a membrane-bound C-terminal fragment (CTF), C99. This membrane-bound product serves as a substrate for γ-secretase complex processing, which releases amyloid peptides of 40 or 42 aa. Pharmacologic and cell biology studies demonstrated that the three major enzymatic activities involved in APP processing, α-β- and γ-secretases, are distinct in their subcellular localization and in their respective cleavage products (2, 3). It was shown that β-secretase activity must reside both in the endosomes (4) and in the secretory pathway (5). Antibody uptake and biotinylation studies showed that most cell surface-located BACE is reinternalized into the early endosomal compartments, from where it can recycle back to the cell surface or can later be retrieved to endosomal/lysosomal compartments and/or to the trans-Golgi network (6, 7). The endocytic pathway, responsible for internalization and initial processing of cell surface APP in endosomes, is well established (4, 8–10). Indeed, the mutagenesis of the APP internalization signal (11) and expression of the dominant-negative dynamin mutant that prevents endocytosis in the transfected cells (10) reduced both AβP 40 and AβP 42 secretion levels. Recent morphological evidence from living cells tests the hypothesis that APP–BACE interactions occur at the cell surface and in early endosomes. Free resonance energy transfer analysis showed that there is a strong and previously unemphasized interaction at the cell surface where APP and BACE dramatically colocalize and appear to be internalized together after 15 min into early endosomes (12). Colocalization of APP and BACE in early endosomes highlights the cell surface as an additional potential site for APP–BACE interaction. All these studies suggest that AβP is derived through processing of APP endocytosed from the cell surface in addition to the secretory pathway.
We report here the preparation of a specific mAb against the β-secretase cleavage site of APP (blocking β-site 1, mAb BBS1). BBS1 antibody was found to reduce the extra- and intracellular AβP levels in CHO cells overexpressing human APP, whereas unrelated antibodies directed to the N-terminal of APP had no effect.
Materials and Methods
All animals were treated according to the regulations of the Animal Care and Use Committee of Tel Aviv University.
Antigen Preparation. The antigen used in this study for immunization mimics the β-secretase cleavage site of APP. The β-secretase cleavage site, which resides between amino acids 663 and 671, is highly conserved through evolution, whereas the double Swedish mutation localized at the same site exhibits the sequence NL instead of KM (amino acids 670–671). To overcome the poor immunogenicity of short peptides and the tolerance against self-antigens, we prepared multiple antigenic peptide (MAP) expressing eight copies of peptides that mimic the WT APP β-secretase cleavage site (MAP-[ISEVKMDA]8) as well as the half-Swedish mutation (MAP-[ISEVKLDA]8).
Immunization Protocol. BALB/c mice, 8 weeks old and weighing 20–30 g, were challenged with MAP displaying eight copies of either the half-Swedish mutation or the WT sequence of the β-secretase cleavage site. Mice were immunized with 100 μg of immunogen emulsified with Freund's adjuvant (Difco) five times at 14-day intervals. Blood samples were drawn before the first immunization and 7 days after each injection. Sera fractions were purified and analyzed for IgG levels against MAP-[ISEVKMDA]8 expressing the WT sequences of the β-secretase cleavage site of APP.
Evaluation of IgG Levels by ELISA. Microtiter plates (Nalge Nunc) were coated with 0.1 μg per well of MAP-[ISEVKMDA]8, which mimics the WT β-secretase cleavage site of APP [diluted in 0.1 M Na2CO3 (pH 9.6)] and incubated overnight at 4°C. The plates were washed with PBS (0.05% Tween 20) and then blocked with 3% BSA/PBS for 1 h at 37°C. Serial sera dilutions were added for an additional hour at 37°C, and the sera end-point titer, defined as the maximal sera dilution in which antibodies can still be detected in the assay, was determined by using an excess of horseradish peroxidase-conjugated secondary antibody (8 ng per well, Jackson ImmunoResearch) followed by O-phenylenediamine (Sigma). After the redox reaction was stopped by using 4 M HCl, substrate degradation was monitored by absorption at 492 nm.
Generation of a mAb Against the β-Secretase Cleavage Site on APP. Immunized mice with the highest titer against MAP conjugated to the WT sequence of the β-secretase cleavage site (MAP-[ISEVKMDA]8) were killed, and lymphocytes isolated from their spleens were fused with NSO myeloma cells as described for hybridoma technique (13). The generated clones were screened for MAP-[ISEVKMDA]8 binding by ELISA as described above. Positive clones were isolated and further analyzed for binding of full-length WT APP as expressed by CHOhAPP751 cells by using immunofluorescence techniques as detailed below. The selected mAb, designated BBS1, demonstrated a strong and specific labeling both on the cell surface and inside the cell and was chosen for further analysis.
Specificity of mAb BBS1 Binding. mAb BBS1 purification from ascetic fluids was performed by protein G affinity chromatography by using the AKTA Prime continuous flow chromatography system. The antibody ability to bind MAP-[ISEVKMDA]8 at different concentrations was analyzed by using ELISA as described above.
Antibody specificity was analyzed by a competitive ELISA. Decreasing concentrations of MAP-[ISEVKMDA]8 (0.125 mM to 29 pM) were preincubated for 1 h at room temperature with mAb BBS1 (13.3 nM). The immunocomplex was then added to a microtiter plate precoated with 0.1 μg per well of MAP-[ISEVKMDA]8. The residual amount of immunocomplex adsorbed to the plate was evaluated as described above. As a negative control, mAb BSS1 was incubated with nonrelevant MAP expressing eight copies of a prion antigen.
Stability of the Immunocomplex at Different pH Values. The stability of the mAb BBS1 and MAP-[ISEVKMDA]8 immunocomplex was evaluated at different pH values in the range of 3–7. Microtiter plates were coated with 0.1 μg per well of MAP-[ISEVKMDA]8 as described above. After blocking, mAb BBS1 (13.3 nM) diluted in buffer in the pH range of 3–7 was added to the antigen-coated plates in triplicate for an additional hour at 37°C. Antibody ability to bind the antigen at different pH values was evaluated as described above by using ELISA. Release of the coated antigen from the plate at low pH conditions was tested by incubation of the antigen-coated wells with buffers at different pH values.
Antibody Recognition of the β-Secretase Cleavage Site. Cell line. CHO cells stably transfected with WT human APP 751 isoform (CHOhAPP751) were kindly provided by D. Selkoe (Harvard Medical School, Boston). Cells were grown in DMEM (F-12) containing 10% FCS and 2.5 mM l-glutamine. hAPP 751-expressing cells were selected by using 1 mg/ml G-418 (Calbiochem) in the cells growing medium.
Western blot. Lysates extracted from CHOhAPP751 cells were used for Western blot analysis. Cells were lysed with ice-cold Triton-doc lysis buffer (0.5% Triton X-100/0.25% Na-deoxycholate/150 mM NaCl/10 mM Tris·HCl, pH 7.5/10 mM EDTA) and then centrifuged at 21,000 × g for 1 min. Supernatants were collected, incubated for 20 min on ice, subjected to 10% SDS/PAGE, and then blotted onto nitrocellulose membrane (Schleicher & Schuell). The membrane, blocked with 4% milk in Tris-buffered saline (0.3% Tween 20), was further incubated overnight with different concentrations of mAb BBS1 (6.6–26.6 nM) and mAb AMY33 (20 nM, Zymed) that bind APP in the midregion of amyloid-β (Aβ). Anti-mouse IgG horseradish peroxidase-conjugated secondary antibody was added for 45 min after the membrane was thoroughly washed. Blots were developed by using the enhanced chemiluminescence system according to the manufacturer's instructions (Pierce).
Cell immunofluorescence. CHOhAPP751 cells (2 × 105) were seeded on coverslips in 24-well plates. At ≈80% confluence, cells were washed twice with PBS and fixed with 4% paraformaldehyde (in PBS) for 30 min at room temperature. Cells were washed four times with 1% NH4Cl (in PBS) and permeabilized by adding 0.1% Triton X-100 in PBS for 2 min. After washes with PBS, cells were blocked with 10% normal goat serum in 3% BSA for 30 min and incubated with mAb BBS1 (80 nM) for 1 h, followed by an additional hour of incubation with Cy2-conjugated goat anti-mouse IgG (Jackson ImmunoResearch). After being thoroughly washed with PBS, cells were mounted by using Prolong Antifade (Molecular Probes).
Antibody Internalization into the Cell. The antibody internalization assay is similar to the cell labeling described above, except that mAb BBS1 was administered in the cell medium before immunolabeling.
At 80% confluence, mAb BBS1 (13.3 nM) was added to the cell medium. Cells were fixed and permeabilized as mentioned above after 30, 60, or 90 min of incubation with the antibody. After cell blocking, rabbit anti-early endosome antigen 1 (EEA1) polyclonal antibodies (Calbiochem) were added to the cells for 1 h. mAb BBS1 and rabbit anti-EEA1 were visualized by the addition of both Cy2-conjugated goat anti-mouse IgG and Cy3-conjugated goat anti-rabbit IgG, respectively, for 45 min.
Antibody Interference with AβP Production. CHOhAPP751 cells (2.5–4 × 106) were seeded in six-well plates. At 100% confluence, cells were washed twice with PBS and administered with sera-free media consisting of mAb BBS1 (13.3 nM), rabbit anti APP N-terminal antibodies (residues APP 46–60) (13.3 nM, Sigma), and/or 100 μM chloroquine, which is known to inhibit cell endocytosis. The basal level of AβP was monitored in cells treated with sera-free media alone. The experiment was performed six times for each treatment. For extracellular AβP evaluation, media was collected after 3, 9, and 24 h of incubation, and cells were further incubated for an additional 4 days. Cells were then collected from each well by using a cell scraper, centrifuged at 3,000 × g for 2 min, washed with PBS, and resuspended in 100 μl of 70% formic acid, followed by 10-s sonication. The solution was then centrifuged at 100,000 × g for 20 min at 4°C to remove insoluble material, and supernatant was collected and neutralized with 1.9 ml of 1 M Tris (pH 9). All samples were analyzed for their protein concentration by using Bradford reagent (Bio-Rad) and aligned for their protein content before evaluation of AβP levels. One of six repeats from each treatment group was used for 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) viability assay (see below). AβP measurements. The levels of both secreted and intracellular AβP were quantified by using a sandwich ELISA. The monoclonal anti-AβP antibody AMY-33, used as the capture antibody, was added to ELISA plates [0.23 μg per well diluted in 0.1 M Na2CO3 (pH 9.6)] and incubated overnight at 4°C. The plates were washed with PBS (0.05% Tween 20) and blocked with 3% BSA/PBS for 3 h at 37°C. Media samples (40, 15, and 5 μl collected after 3, 9, and 24 h, respectively) and equilibrated protein extracts from each treatment were applied in triplicate and incubated overnight at 4°C (a 50-μl volume was adjusted in all treatment groups by using PBS). Biotinylated monoclonal anti-AβP antibody 196 (76 ng per well), which recognizes the sequence 3–6 of AβP, and biotinylated mAb 8G7 (75 ng per well, Calbiochem) were used for detection of total AβP and AβP42, respectively. Antibodies were diluted in 1% BSA/PBS and incubated for 3 h at 37°C. Plates were thoroughly washed with PBS (0.05% Tween 20) and administered with extraavidin-conjugated alkaline phosphatase (Sigma) for 1 h at 37°C. The substrate p-nitrophenyl phosphate (Sigma) was used to monitor AβP levels by absorption at 405 nm. Synthetic AβP40andAβP 42 (0.1–1 ng per well, Bachem) were used for construction of calibration curves.
C99 analysis by metabolic labeling. Because the levels of C99 produced by CHOhAPP751 are below detection limits, evaluation of treatment ability to reduce C99 levels was performed in 100-mmdiameter culture dishes where each treatment group consists of cells combined from four identical plates. CHOhAPP751 cells were grown to 100% confluence, washed with PBS, and administered with sera-free media in the absence of methionine. After 30-min starvation, cells were labeled with 125 μCi/ml (1 Ci = 37 GBq) [35S]methionine and administered with mAb BBS1 (13.3 nM) for 4 h. Labeled cells alone under the same conditions were used for evaluation of the basal level of C99. Cells were then harvested and lysed by using 0.05 M Tris·HCl, pH 8.0/0.15 M NaCl/0.005 M EDTA containing 1% Nonidet P-40 and 200 μg/ml PMSF. Extracted proteins were immunoprecipitated with anti-APP C-terminal antibody (Calbiochem) and 50 μl of protein G magnetic beads (Dynal, Great Neck, NY). The immunocomplexes were separated on a 15% Tris-Tricine gel. Radioactive enhancement was achieved by gel incubation with 2,5-diphenyloxazole (20% in acetic acid, Sigma) for 1 h, after which the gel was dried and the radioactive signal collected. C99 density was evaluated by using imagemaster densitometry software (Amersham Pharmacia).
sαAPP levels analysis by Western blot. Culture media (30 μl) was collected from cells treated with mAb BBS1, as well as from untreated cells after 24 and 48 h, and electrophoresed as described above for Western blot analysis. Visualization of sαAPP was achieved by using mAb 196, which binds amino acids 3–6 of Aβ (1 μg/ml) and thus binds sαAPP and not sβAPP.
MTT viability assay. Cell viability was measured by colorimetric MTT assay as described in ref. 14. One six-well plate containing cells from each treatment group after 5 days of incubation was washed twice with PBS and administered with fresh sera-free media containing 1 mg/ml MTT reagent (Sigma) for 2 h at 37°C. The dark-blue crystals formed were dissolved by addition of 100 μl per well of lysis solution [20% SDS (wt/vol) dissolved in 50% dimethylformamide (pH 4.7)] and an overnight incubation at 37°C. The colorimetric reaction was measured by absorption at 570 nm.
Results
Generation of Immune Response Against the β-Secretase Cleavage Site of APP. Mice were immunized with MAP displaying eight copies of WT β-secretase cleavage site. However, the antibody titer obtained was low after five immunizations, as can be expected for a self-antigen. To stimulate the mice's immune systems against the β-secretase cleavage site of WT APP (amino acids 663–671), which is highly conserved in mammals, we immunized an additional group of mice with MAP displaying eight copies of the partial Swedish mutation that resides at the same site (M670L). The immunoreactivity of sera isolated from the two groups of immunization was analyzed against MAP displaying the WT epitope by using ELISA. The one amino acid substitution, using MAP displaying the half-Swedish mutation, enabled the generation of a high antibody titer against both the WT and Swedish mutated β-secretase cleavage site in a short period. High levels of IgG were evident after the first immunization with MAP displaying half of the Swedish mutation, reaching an end-point titer, defined as the maximal sera dilution in which specific antibodies can still be detected, of 1:106 after the third immunization (Fig. 1A).
Generation and in vitro characterization of mAb BBS1 raised against the β-secretase cleavage site of APP. (A) Immune response in BALB/c mice immunized with MAP expressing eight copies of the partially Swedish mutated BACE cleavage site of APP (MAP-[ISEVKLDA]8). Sera fractions of immunized mice were analyzed for their anti-MAP-[ISEVKMDA]8 IgG levels displaying the WT sequence of the β-secretase site of APP. Black bars, first immunization; gray bars, second immunization; white bars, third immunization. Spleens were isolated from mice with the higher titers, and mAb was prepared by the hybridoma technique. (B) mAb BBS1 binding properties. mAb BBS1's ability to bind MAP-[ISEVKMDA]8 was analyzed by ELISA. Percentage of the maximal antibody binding is presented in different antibody concentrations. (C) Competition between soluble antigen (MAP-[ISEVKMDA]8) and antigen adsorbed to ELISA plates for mAb BBS1 binding. Continuous line, binding of mAb BBS1 after incubation with increasing concentration of MAP-[ISEVKMDA]8; dashed line, antibody binding after incubation with nonrelevant antigen. (D) Stability of immunocomplex at various pH values. mAb BBS1's ability to bind its antigen (MAP-[ISEVKMDA]8) was tested by ELISA after 1 h of incubation at different pH values as described in Materials and Methods. Diamonds, pH effect on mAb BBS1 antigen binding; squares, pH effect on the coated antigen.
Generation of a mAb Against the β-Secretase Cleavage Site of APP. Mice with the highest antibody levels were killed, and their spleens were used for preparation of monoclonal antibodies by the hybridoma fusion technique. A series of mAbs was obtained by screening the hybridoma supernatant for their ability to bind MAP displaying the WT sequence of the β-secretase cleavage site. mAb BBS1, which showed the higher affinity in binding MAP expressing the WT sequence of the β-secretase cleavage site (MAP-[ISEVKMDA]8), was chosen for further analysis. The antibody ability to bind MAP-[ISEVKMDA]8 was analyzed by ELISA, showing an IC50 of 1.165 nM (Fig. 1B). mAb BBS1 was also analyzed for its ability to bind AβPs (AβP 1–16, AβP 40, and AβP 42) by using ELISA. In all cases described, no immunoreactivity was detected (data not shown).
Antibody specificity to MAP-[ISEVKMDA]8 was confirmed in a competitive ELISA performed after preincubation of increasing antigen concentration with a constant amount of mAb BBS1 (Fig. 1C). Although antibody ability to bind the precoated plate was reduced after incubation with the antigen in solution in a dose-dependent manner, preincubation with nonrelevant MAP failed to affect antibody binding.
Because optimal β-secretase activity requires a mildly acidic pH (≈5), we tested the immunocomplex stability under a pH range of 3–7. At pH 5, mAb BBS1 retains 95% of its ability to bind the antigen, and 20% of its binding abilities are still evident after 1 h of incubation at pH 3 (Fig. 1D).
mAb BBS1 Recognizes Full-Length APP. Because mAb BBS1 was generated against a small peptide from APP sequence, it was essential to assure that the antibody binds full-length APP. The ability of the antibody to bind the WT β-site of full-length APP was analyzed in the CHOhAPP751 cell line expressing high levels of human WT APP by Western blot (Fig. 2Ai) and cell immunofluorescence (Fig. 2Aii). In the Western blot analysis, mAb BBS1 specificity is demonstrated in a concentration-dependent manner. mAb BBS1, generated against half of the Swedish mutated β-site sequence, recognizes the double Swedish mutated APP in Western blot using protein extracts of CHO cells overexpressing the Swedish mutated APP [a kind gift from T. Gold and C. Eckman (both of the Mayo Clinic, Jacksonville, FL)] (data not shown). In the immunofluorescence assay, cells were fixed and permeabilized, after which mAb BBS1 was administered and visualized by using a Cy2-conjugated secondary antibody. Antibody reactivity was evident at the cell surface and in the perinuclear areas, such as the endoplasmic reticulum and Golgi apparatus, as well as in other cell compartments.
Antibody ability to bind full-length APP expressed by CHOhAPP751 cells. (A) Antibody ability to bind APP expressed by CHOhAPP751 is demonstrated by Western blot (i) and immunofluorescence (ii). (i) CHOhAPP751 cells were lysed, electrophoresed, and transferred to a nitrocellulose membrane. mAb BBS1 was applied at different concentrations, and mAb AMY-33, which binds the midregion of Aβ, was used as a positive control (left). (ii) Immunolabeling of cells with mAb BBS1. Antibody binding is detected by using a Cy2-conjugated secondary antibody and visualized by using an LSM-510 Zeiss confocal microscope. (Scale bar, 10 μm.) (B) Internalization assay of mAb BBS1. CHOhAPP751 cells were administered with mAb BBS1 in the growing media and incubated with the antibody for 60 min. After incubation with the antibody, cells were fixed and permeabilized, and antibody presence inside the cells was detected by using Cy2 secondary antibody (Left). Cells were counterstained for early endosomes by using rabbit anti-EEA1 after Cy3 secondary antibody (Center), and the superposition is demonstrated in Right. Upper and Lower represent two different fields at different magnitudes. (Scale bar, 10 μm.) Cell labeling was visualized by using an LSM-510 Zeiss confocal microscope.
Because most β-secretase activity is localized intracellularly, we tested antibody ability to cointernalize into the early endosomes with APP after binding at the cell surface. Antibody presence inside the cell was evident after 60 min (Fig. 2B). Fig. 2B Upper, which is at higher magnitude, shows very strong colocalization of EEA1 and mAb BBS1, as can be expected after APP internalization by means of clathrin-coated vesicles. However, at that time, mAb BBS1 is also evident in other intracellular areas.
mAb BBS1 Interferes with AβP Production. The effect of immunocomplexation of mAb BBS1 with APP on AβP production was measured by using 100% confluent CHOhAPP751 cells in the presence of mAb BBS1 in sera-free media. Anti-N-terminal polyclonal antibodies were used as a negative control and chloroquine, which is known to inhibit endocytosis and thus reduce AβP levels, served as a positive control. Untreated cells were used to measure the basal levels of AβP. Media were collected after 3, 9, and 24 h, as described in Materials and Methods. These time points were chosen, because the relatively high AβP concentration in growing media is measurable above detection limits. The percentage of secreted AβP in each treatment group compared with AβP secreted by the untreated group at the three time points measured is shown in Fig. 3A. An ≈13% reduction in secreted AβP levels was evident after 3 h of incubation with mAb BBS1, reaching 22% within 9 h. After 24 h, only 16% reduction was measured, probably a result of antibody consumption. Anti-N-terminal polyclonal antibodies had no significant effect on AβP production at any of the measured time points. Chloroquine, as can be expected, displayed the most profound effect on AβP levels; however, the reduction is not only a result of endocytosis inhibition, but also of massive cell death.
Inhibition of extra- and intracellular AβP accumulation. The ability of mAb BBS1 to interfere with the APP processing and thus reduce AβP levels was examined in the following way: CHOhAPP751 cells were administered with mAb BBS1, anti-APP N-terminal polyclonal antibodies, or chloroquine diluted in sera-free media. Untreated cells were used to measure AβP basal level. (A) Secreted AβP levels were measured from the growing media at different time points (white bars, 3 h; black bars, 9 h; gray bars, 24 h), and the ratio between secreted AβP in each treatment group and the untreated group was calculated and presented in percentage. The experiment was repeated six times for each group. (B) Intracellular AβP measurements were performed after 5 days of incubation. The ratio between the intracellular AβP levels in each treatment group and in the untreated group was calculated and presented in percentage (black bar, untreated cells; white bar, mAb BBS1-treated cells; gray bars, anti-N-terminal antibody and chloroquine-treated cells). The experiment was repeated five times for each group. (C) CTFβ levels were measured after 4 h of metabolic labeling and immunoprecipitation with anti-C-terminal antibody. CTFβ is shown at a molecular mass of 12 kDa in the untreated (–) and mAb BBS1-treated (+) cells, and densitometric analysis estimated a 20% reduction of CTFβ levels in mAb BBS1-treated cells. (D) MTT cell viability assay was performed after 5 days of incubation (black bar, untreated cells; white bar mAb BBS1-treated cells; gray bars, anti-N-terminal antibody and chloroquine-treated cells). (E) The levels of sαAPP released to the cells' media were analyzed after 24 and 48 h by Western blot using mAb 196, which binds Aβ in the N terminus. Soluble αAPP was detected at a molecular mass of ≈100 kDa in the supernatant of mAb BBS1-treated cells (+) as well as in that of untreated cells (–).
After 5 days of incubation, cells were lysed with 70% formic acid and sonicated. Fig. 3B shows the percentage of intracellular AβP compared with that of the untreated group. A dramatic reduction of ≈50% is evident in cells treated with mAb BBS1 for 5 days. Chloroquine, as observed for secreted AβP, also showed a dramatic reduction of 60% in the intracellular levels; however, by the time of measurement (5 days), most of the chloroquine-treated cells had died (Fig. 3D). To assure that the dramatic reduction of intracellular AβP levels observed in the mAb BBS1 treatment group is not the result of a lower number of cells producing AβP, as evident in the chloroquine-treated cells, one repeat of the experiment (of six repeats) was applied to MTT viability assay (Fig. 3D). No significant change in viability is evident in the untreated and mAb BBS1-treated cells.
The sandwich ELISA described here measures total AβP levels for both secreted and intracellular AβP. Results obtained from detection of only AβP 42 (using biotinylated 8G7 antibody) in these samples showed that most of the secreted AβPisAβP 40, whereas intracellular AβP is predominantly AβP 42 (data not shown), as was reported by others (15, 16).
Inhibition of β-secretase activity results in a decrease in the levels of the corresponding membrane-bound CTF, C99. For that purpose, we analyzed the levels of C99 in mAb BBS1-treated cells versus untreated cells by means of metabolic labeling using [35S]methionine. After 4 h of treatment with the antibody, the decrease in C99 levels was estimated as 20% (Fig. 3C). Reduction in C99 levels at this point is in accordance with the 13% decrease in secreted AβP levels observed after 3 h of incubation with the antibody (Fig. 3A). The most profound effect on AβP levels was evident by examining intracellular AβP levels after 5 days of incubation with the antibody. However, evaluation of C99 by means of metabolic labeling could not be performed for such a long period because, at later time points (i.e., 12 h), cells were beginning to die in the two examined groups, probably due to lack of methionine.
To evaluate whether blocking the β-secretase site of APP, and thus inhibiting AβP production, enhances α-processing, sαAPP levels were analyzed. No significant change in sαAPP levels was observed between the untreated and mAb BBS1-treated cells after both 24 and 48 h of incubation (Fig. 3E).
Discussion
Altered processing of the APP is considered a major event in the pathogenesis of AD, but what accelerates amyloidogenesis in sporadic AD has not been identified. The AβP that accumulates in Alzheimer's diseased brain is derived from proteolytic processing of the APP by means of β- and γ-secretases through the secretory and endocytic pathways. BACE, which generates the N terminus of AβP, has become a main target of intense research aimed at blocking the enzyme activity and, thus, production of AβP toward inhibition of amyloid plaque formation (17–19). However, recent evidence has shown that BACE may process some non-APP substrates (20–22) and that other putative β-secretases may be involved in APP cleavage, mainly in sporadic AD (23, 24). In this study, we describe a unique approach to inhibit the mainly endocytic pathway of β-secretase activity based on blocking the β-secretase cleavage of APP by site-directed antibodies.
We generated a specific mAb against the β-secretase cleavage site of APP. Specificity of the antibody was established in a competitive ELISA, as well as by Western blot analysis and immunofluorescence, by using CHO cells expressing WT human APP. The immunocomplex was found to remain 95% stable within pH 5, the common pH in the β-secretase environment (e.g., endosomes). mAb BBS1, raised against the β-site of APP, cointernalized with APP into the early endosomes after APP binding at the cell surface and reduced the extra- and intracellular AβP levels in CHO cells overexpressing human APP. The intracellular levels of AβP were dramatically reduced compared with the basal level of AβP produced by these cells after 5 days of incubation with the antibody. The secreted AβP levels were reduced by 22% after 9 h. CTF-β levels, after 4 h of treatment with mAb BBS1, were reduced by 20%, in accordance with the 13% decrease observed in secreted AβP after 3 h. It is worth mentioning that antibodies directed to the N terminus of APP have failed to produce such an effect with both the extra- and intracellular levels of AβP.
The ability of BBS1 antibody to interfere with AβP production and thus with the disease progression in vivo is under investigation by using AD transgenic mice models expressing either the WT or Swedish mutated β-site of APP.
We performed active immunization feasibility studies in a small number of 8-month-old AD transgenic mice (Tg 2576, Taconic Farms). Mice were immunized with the described antigen exhibiting eight copies of half-Swedish mutated β-site (n = 5). The treated mice showed improved cognitive behavior in the Morris water maze test compared with saline-treated mice (n = 4). Moreover, brain biochemistry and immunohistochemistry analysis demonstrated a remarkable reduction in total AβP levels and plaque load, respectively (Fig. 4, which is published as supporting information on the PNAS web site). These data suggest that the generated antibodies against the β-secretase cleavage site do cross the blood–brain barrier, bind APP, and interfere with plaque formation by reducing the soluble AβP levels.
Experimental evidence that peripheral antibodies cross the blood–brain barrier are derived also from several recently reported data. Morgan and coworkers (25) show that in APP transgenic mice administered with anti-Aβ antibodies, and not in those administered with a control IgG, antibodies labeled the remaining congophilic amyloid plaques in the brain. Seabrook et al. (26) also reported immunohistochemical labeling of amyloid deposits for mouse IgG after passive immunization. These data, performed with paraffin-embedded tissues, confirm the observations of Bard et al. (27), who used cryosections in demonstrating the ability of anti-Aβ antibodies administered peripherally to enter the CNS, decorate preexisting amyloid plaques, and mediate their clearance (27).
Despite the fact that AD pathogenesis is classically characterized by the extracellular accumulation of Aβ peptides, the extent to which those insoluble fibrils contribute to the neuronal death is still unclear. Support for the intracellular origin of the extracellular Aβ deposits comes from several recent studies demonstrating that AβP accumulates in the endosomal/lysosomal system before it appears in the amyloid plaque formed in AD transgenic mice models (28).
Observation of early AβP 42 intraneuronal accumulation within those brain areas affected earliest in AD suggests a mechanism that may explain AD progression within the brain. Intracellular AβP42 may act as a nidus for Aβ deposition intra- and extracellularly in the cell body and along processes and terminals of affected neurons. The resultant accumulation of AβP in the parenchyma may hasten the pathological process, providing a potential mechanism for the “spread” of Aβ-related pathology (15).
All these studies emphasize the importance of limiting intracellular AβP 42 formation at early stages of AD development. To our knowledge, this study is the first attempt to inhibit β-secretase activity by blocking the BACE cleavage site of APP by means of site-directed antibodies rather than by a direct and total inhibition of BACE itself by various small synthetic molecules. Antibody-mediated blockage of the β-secretase site of APP by either mAb BBS1 in a cellular model or by active immunization of transgenic mice led to a considerable reduction in intracellular AβP, suggesting that this approach may be applicable as a prophylactic therapy for AD. Antibody internalization into the cell after APP binding, together with the fact that it does not bind AβP, may be beneficial in avoiding microglia and complement activation, as was reported for anti-Aβ antibodies that bind the senile plaques (reviewed in ref. 29).
The concept presented here for inhibiting β-secretase activity on APP primarily affects AβP generated by means of the endocytic pathway, which is responsible for the internalization and processing of cell surface APP. Because up-regulation of endocytosis may represent a possible basis for accelerated β-amyloidogenesis in >90% of all AD cases in the absence of a known gene mutation (23, 30), the immunological approach presented in this study may become a therapeutic strategy in AD treatment that is worthy of further investigation.
Acknowledgments
We thank Prof. D. Selkoe, Dr. T. Golde, and Dr. C. Eckman for providing the different cell lines; Drs. Vered Lavie and Maria Becker for analyzing amyloid burden in the transgenic mice experiment; F. Margolin for manuscript editing; D. Galitzki for help with animal care; and the members of our laboratory for helpful discussions.
Footnotes
-
↵ † To whom correspondence should be addressed. E-mail: beka{at}post.tau.ac.il.
-
↵ * M.A. and I.Y. contributed equally to this work.
-
Author contributions: B.S. designed research; M.A. and I.Y. performed research; M.A. and B.S. analyzed data; and M.A. and B.S. wrote the paper.
-
Abbreviations: Aβ, amyloid-β; AβP, Aβ peptide; AD, Alzheimer's disease; APP, amyloid precursor protein; BACE, β-secretase APP cleaving enzyme; BBS1, blocking β-site 1; EEA1, early endosome antigen 1; MAP, multiple antigenic peptide; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; CTF, C-terminal fragment.
- Copyright © 2005, The National Academy of Sciences
