X11 Interaction with β-Amyloid Precursor Protein Modulates Its Cellular Stabilization and Reduces Amyloid β-Protein Secretion*

The protein interaction domain of the neuronal protein X11 binds to the YENPTY motif within the cytoplasmic domain of β-amyloid precursor protein (βAPP). Amyloid-β protein (Aβ), the major constituent of the amyloid deposited in brain of Alzheimer’s disease patients, is generated by proteolytic processing of βAPP, which occurs in part following βAPP internalization. Because the YENPTY motif has a role in the internalization of βAPP, the effect of X11 binding on βAPP processing was studied in mouse neuroblastoma N2a, human embryonic kidney 293, monkey kidney COS-1, and human glial U251 cell lines transfected with wild type or mutated βAPP cDNAs. Secretion of soluble βAPP via α-secretase activity increased significantly in cells transfected with βAPP variants containing mutations that impair interaction with X11 when compared with cells transfected with wild type cDNA. Cotransfection of βAPP and X11 caused retention of cellular βAPP, decreased secretion of sβAPPα, and decreased Aβ secretion. Thus, βAPP interaction with the protein interaction domain of X11 stabilizes cellular βAPP and thereby participates in the regulation of βAPP processing pathways.

The protein interaction domain of the neuronal protein X11 binds to the YENPTY motif within the cytoplasmic domain of ␤-amyloid precursor protein (␤APP). Amyloid-␤ protein (A␤), the major constituent of the amyloid deposited in brain of Alzheimer's disease patients, is generated by proteolytic processing of ␤APP, which occurs in part following ␤APP internalization. Because the YENPTY motif has a role in the internalization of ␤APP, the effect of X11 binding on ␤APP processing was studied in mouse neuroblastoma N2a, human embryonic kidney 293, monkey kidney COS-1, and human glial U251 cell lines transfected with wild type or mutated ␤APP cDNAs. Secretion of soluble ␤APP via ␣-secretase activity increased significantly in cells transfected with ␤APP variants containing mutations that impair interaction with X11 when compared with cells transfected with wild type cDNA. Cotransfection of ␤APP and X11 caused retention of cellular ␤APP, decreased secretion of s␤APP␣, and decreased A␤ secretion. Thus, ␤APP interaction with the protein interaction domain of X11 stabilizes cellular ␤APP and thereby participates in the regulation of ␤APP processing pathways.
Amyloid-␤ protein (A␤), 1 deposited in the brain of patients with Alzheimer's disease (AD), Down's syndrome, and sporadic and hereditary cerebral amyloid angiopathy and in the brain of elderly individuals is a proteolytic peptide of a larger ␤-amyloid precursor protein (␤APP). A␤ has been reported to have heterogeneous carboxyl termini, and A␤ 1-40 and A␤  appear to be the major species in the parenchymal deposition. A␤ is found in normal cerebrospinal fluid and in conditioned media from various tissue culture cell lines (1)(2)(3), suggesting that it is produced and secreted constitutively. Alternative processing pathways of ␤APP have been described. Cleavage at position 597 of ␤APP 695 by a ␤-secretase results in the generation of A␤. Alternatively, processing by an ␣-secretase between positions 612 and 613 of ␤APP 695 (positions 16 and 17 of A␤) precludes the release of intact A␤ (4,5). Both pathways result in the release of amino-terminal soluble ␤APP, s␤APP␤ and s␤APP␣, into the extracellular compartment. The carboxyl-terminal processing products may undergo an additional cleavage by a protease displaying a ␥-secretase activity leading to the formation of A␤ or P3.
␣-Secretase processing of ␤APP occurs in a late compartment of the constitutive secretory pathway (6,7), probably in a late trans-Golgi compartment (8,9). However, ␤APP can elude the intracellular cleavage and reach the cell surface as a full-length mature product, and ␣-secretase activity may occur also at the plasma membrane in several cell systems (10,11). Immunolabeling of cell surface ␤APP in living cells demonstrated that cell surface ␤APP is either rapidly released or internalized via clathrin-coated vesicles, such that the duration at the cell surface is very short (10,12,13). Whereas most A␤ is produced by a ␤-secretase at the cell surface (14,15), via the endosomal/ lysosomal pathway, a small fraction of the normally produced A␤ 1-42 is generated in the endoplasmic reticulum/intermediate compartment, and the trans-Golgi network is a site for A␤  generation (14,16,17). ␥-Secretase activity occurs during recycling of endosomes to the cell surface (1,11).
The ␤APP cytoplasmic domain has a sequence motif, YENPTY, that is involved in protein internalization (11,18,19). Deletion of either the cytoplasmic domain or the YENPTY sequence resulted in reduced ␤APP internalization, increased secretion of s␤APP␣ and P3, and significantly diminished A␤ release (6,11,20). No clear effect was observed on the level of intracellular A␤ (21).
␤APP is a cell surface protein with a large extracellular amino-terminal domain, a single transmembrane segment, and a short cytoplasmic tail (22). Its location and structural features are characteristic of a receptor for signal transduction. Screening for potential proteins capable of interacting with its cytoplasmic domain led to the identification of proteins containing phosphotyrosine interaction/phosphotyrosine binding (PI/PTB) domains: Fe65, X11, and their homologues (23)(24)(25)(26)(27)(28). The PI/PTB domain was first identified as the component of the adaptor protein Shc (Src homology 2/collagen homology) that binds to activated and tyrosine-phosphorylated receptors (29,30). This domain was further found in several unrelated regulatory proteins (31), suggesting a general role for this domain in protein-protein interactions and signal transduction. Fe65 was originally proposed as a transcriptional activator (32). However, the presence of a WW domain (a variant of Src homology 3 domains) (33), as well as two PI/PTBs (31), suggests that it is likely to be involved in signal transduction. The expression of Fe65 appears to be enriched in brain in human, rat, and mouse (25,32,34). The X11 gene was originally isolated as a candidate gene for Friedreich ataxia, an autosomal recessive degenerative disorder that affects the cerebellum, spinal cord, and peripheral nerves (35). The X11 protein is a neuron-specific protein of unknown function that contains two postsynaptic density protein, disc-large, zo-1 (PDZ), domains at its carboxyl terminus in addition to a PI/PTB domain (31).
Biochemical characterization of the interaction of X11 and Fe65 with ␤APP indicated that the YENPTY motif located at the cytoplasmic carboxyl terminus of ␤APP is essential for its association with the PI/PTB domains (23,24). The interaction of ␤APP and the PI/PTB domains of X11 and Fe65 is phosphorylation-independent (23,24,27,28), suggesting that, in contrast to the SH2 domains, the PI/PTB domains are primarily peptide binding domains that have in some cases acquired specificity for phosphorylated tyrosines (36). The binding site of the Fe65 PI/PTB domain appears to be different from that of X11 as mutations within the YENPTY motif differentially affect the binding of X11 and Fe65 (24). The crystal structures of the X11 PI/PTB domain bound to the ␤APP peptide (QNGEY-ENPTYKFFEQ) revealed that the sequence-specific recognition extends to peptide residues that are carboxyl-terminal to the YENPTY motif (36).
Association of PI/PTB domain-containing proteins with the coated pit-mediated internalization signal of ␤APP may affect the patterns of ␤APP trafficking, generation of A␤ as well as of soluble ␤APP, and normal physiologic function. The effect of X11 binding on ␤APP processing was studied in different cell lines transiently or permanently transfected with the wild type or mutated ␤APP cDNAs. We demonstrate that mutations in ␤APP that impair interactions with X11 result in increased proteolysis by an ␣-secretase. Cotransfection of ␤APP with X11 produces the opposite effect. Furthermore, overexpression of both X11 and ␤APP results in accumulation of cell-associated ␤APP and decreased secretion of A␤.

EXPERIMENTAL PROCEDURES
Cell Culture-Mouse neuroblastoma N2a, human embryonic kidney 293 and African green monkey kidney COS-1 cells were cultured in Dulbecco's modified Eagle's medium, and human glial U251 cells were cultured in RPMI medium at 37°C in a 5% CO 2 atmosphere. The media were supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, and 100 g/ml streptomycin sulfate.
Transfection of Cell Lines-The wild type or mutated ␤APP 695 cDNAs and the X11 cDNA were cloned into pRK5 vector as described previously (24). The cDNAs were transiently transfected into N2a, 293, and COS-1 culture cell lines using calcium phosphate (37). U251 cells were stably transfected using LipofectAMINE transfection reagent (Life Technologies, Inc.) and selected with Geneticin (G418, Life Technologies, Inc.). N2a cells stably transfected with wild type ␤APP 695 cDNA (kindly provided by Dr. S. S. Sisodia) and 293 cells stably transfected with human wild type cystatin C gene (38) were transiently transfected with X11 or pRK5 vector cDNAs. Overexpression and secretion of ␤APP were confirmed by immunoblot analysis of cell lysates and medium proteins with anti-␤APP 66 -81 antibody (22C11) (Boehringer Mannheim) and with anti-A␤ 1-17 antibody (6E10) (Senetek). Overexpression of Myc-tagged X11 was confirmed by immunoblot analysis of cell lysates proteins with anti-myc antibody (9E10) (Santa Cruz Biotechnology).
Pulse-Chase Labeling of ␤APP-Transiently or stably transfected culture cells were labeled 24 h after transfection. Cells were incubated with 0.3 mCi/ml [ 35 S]methionine/cysteine EXPRE 35 S 35 S (DuPont) in methionine/cysteine-free medium with 5% dialyzed serum for a pulse of 20 min at 37°C. Following a wash with phosphate-buffered saline, pH 7.3, the cells were chased in complete medium with 150 g/ml methionine at 37°C for different periods. The media were collected and spun at 4500 ϫ g for 10 min at 4°C. The cells were harvested in phosphatebuffered saline, pH 7.3, and lysed in lysis buffer (1% Triton, 1% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 50 mM Tris-HCl, pH 7.2). Equal amounts of total proteins from cell lysates or equal volumes of media based upon the relative concentration of total proteins in cell lysates were immunoprecipitated with anti-A␤ 1-17 antibody (6E10) or anti-␤APP 66 -81 antibody (22C11) overnight at 4°C and with ␥-Bind Plus Sepharose beads (Amersham Pharmacia Biotech) for 2 h at 4°C. In control experiments, cell lysates and medium proteins were immunoprecipitated with anti-cystatin C antibody (Axell) and protein A-Sepharose (Amersham Pharmacia Biotech) for 4 h at 4°C. The immunopre-cipitated proteins were washed with lysis buffer and phosphatebuffered saline, pH 7.3, boiled in sample buffer (1% SDS, 3% glycerol, 1.5% ␤-mercaptoethanol, and 20 mM Tris-HCl, pH 6.8), and separated by 8% SDS-PAGE. The gels were enhanced with Amplify (Amersham Pharmacia Biotech) and exposed to x-ray films. The protein bands were scanned using Adobe PhotoShop and quantitated using the NIH Image program. Relative intensity of the bands was calculated as a percentage of the intensity of the protein band in cell lysates at time 0 of the chase. Results are expressed as means Ϯ S.E. Data were compared between experimental groups using one-way analysis of variance combined with Fisher's test. In all cases, p values Ͻ0.05 were considered to be statistically significant. Direct quantitation of incorporated radioactivity was performed in vials containing liquid scintillation mixture (Beckman) by liquid scintillation spectrometry (Beckman model LS6000IC). Significant positive correlation between quantification done by densitometry and direct measurements of incorporated radioactivity were obtained (r ϭ 0.952 Ϯ 0.008; p Ͻ 0.0002). The half-life of intracellular ␤APP was calculated from pulse-chase experiments performed for chase periods of up to 8 h.
Metabolic Labeling of A␤-N2a cells, stably transfected with ␤APP cDNA, were labeled 24 h after transient transfection with vector or X11 cDNAs. Cells were incubated with 0.5 mCi/ml [ 35 S]methionine/cysteine EXPRE 35 S 35 S (DuPont) in methionine/cysteine-free medium with 5% dialyzed serum for 4 h at 37°C. The media were collected and spun at 4500 ϫ g for 10 min at 4°C. Equal volumes of media, based upon the relative concentration of total proteins in cell lysates, were immunoprecipitated with anti-A␤ 1-17 antibody (6E10) and anti-A␤ 17-24 antibody (4G8) (Senetek) overnight at 4°C and with protein A-Sepharose beads (Amersham Pharmacia Biotech) for 5 h at 4°C. The immunoprecipitated proteins were washed with lysis buffer and phosphate-buffered saline, pH 7.3, boiled in sample buffer, and separated by 16.5% Tris-Tricine PAGE. The gels were enhanced with Amplify (Amersham Pharmacia Biotech) and exposed to x-ray films.

Mutations That Inhibit
Binding of X11 to the YENPTY Motif of ␤APP Affect the Secretion of Soluble ␤APP-We have previously analyzed the binding sites on ␤APP for the X11 PI/PTB domain, using site-directed mutagenesis of the cytoplasmic domain of ␤APP (24). The different mutations were placed into full-length ␤APP 695 (see Fig. 1 for the sequence of the carboxylterminal domain of ␤APP). We found that deletion of the last 18 amino acids severely inhibited the ability of the X11 PI/PTB domain to bind ␤APP. Similarly, two mutations within the YENPTY motif, Y682G and N684A, impaired the binding of the protein interaction domain to ␤APP (24).
Pulse-chase experiments were performed to study the effect of these mutations on ␤APP processing. Mouse neuroblastoma N2a, human embryonic kidney (293), and monkey kidney (COS-1) cells that were transiently transfected and human glial (U251) cells that were stably transfected with wild type or mutated ␤APP cDNAs were labeled with [ 35 S]methionine/cysteine for 20 min and chased for different time periods. Results of the SDS-PAGE temporal profile of ␤APP expression and secretion are shown in Fig. 2. Immunoprecipitation of cell lysates proteins revealed that mutations within the X11 inter-active domain of ␤APP do not significantly affect the decrease of intracellular ␤APP (Fig. 3, A, C, E, and G). Immunoprecipitation of medium proteins with anti-A␤ 1-17 antibody (6E10) revealed that the secretion of soluble ␤APP via the ␣-secretase pathway (s␤APP␣) was increased in all of the cell lines transfected with ␤APP lacking the 18 carboxyl-terminal amino acids, which include the YENPTY motif, when compared with cells transfected with wild type cDNA (data not shown). These results are compatible with previous reports of the effect of deletion of either the cytoplasmic domain or the YENPTY sequence of ␤APP on its processing (6,11,20). We demonstrate that mutation of the amino-terminal tyrosine (Y682G) or asparagine (N684A) within the YENPTY motif also caused increased secretion of s␤APP␣ (Fig. 3, B, D, F, and H).
Effect of Overexpression of X11 Together with ␤APP on Cellassociated ␤APP and on Soluble ␤APP Secretion-Conversely, pulse-chase experiments were performed with N2a, 293, and COS-1 cells transiently cotransfected with wild type ␤APP and either vector or X11 cDNAs. Immunoprecipitation of cell lysate proteins with anti-A␤ 1-17 antibody (6E10) revealed that cotransfection of ␤APP and X11 delayed intracellular ␤APP depletion (Fig. 4). The half-life of intracellular ␤APP in cells cotransfected with ␤APP and X11 was prolonged in comparison to cells transfected with ␤APP alone (Fig. 5). Half-life in this system refers to turnover rate, dependent on both degradation and secretion. Although X11 overexpression attenuated the depletion of intracellular ␤APP in all cell lines tested, the magnitude of the effect was different, with the greatest effect observed in 293 cells and the least in COS-1 cells. Delayed intracellular ␤APP depletion was also observed following immunoprecipitation of N2a cell lysates proteins with anti-␤APP 66 -81 antibody (22C11) (Fig. 6).
Furthermore, immunoprecipitation of conditioned medium proteins with anti-A␤ 1-17 antibody (6E10) (Fig. 4) revealed decreased secretion of s␤APP␣. In order to confirm the results obtained with anti-A␤ 1-17 antibody (6E10), cell lysates and medium proteins were immunoprecipitated with an antibody raised against the amino-terminal end of ␤APP, anti-␤APP 66 -81 antibody (22C11). This antibody immunoprecipitates full-length ␤APP as well as its amino-terminal fragments, s␤APP␣ and s␤APP␤. Decreased secretion of s␤APP was also observed in N2a cells cotransfected with ␤APP and X11 cDNAs, compared with cells cotransfected with ␤APP and vector, following immunoprecipitation with anti-␤APP 66 -81 antibody (22C11) (Fig. 6). The decrease in s␤APP immunoprecipitated with 22C11 antibody may be due to decreased ␣-secretase activity or may indicate a decrease also in ␤-secretase activity. The same results were obtained with N2a cells stably transfected with wild type ␤APP and transiently transfected with X11 (data not shown).
Coexpression of X11 with mutated ␤APP (Y682G or N684A) resulted in a decrease in s␤APP retrieved from the culture media, compared with s␤APP secreted by cells cotransfected with the respective mutated ␤APP and vector alone. Pulsechase experiment with COS-1 cells transiently transfected with wild type ␤APP or ␤APP N684A cDNAs and either vector or X11 cDNAs is presented in Fig. 7. However, cotransfection of X11 with mutated ␤APP had a lesser effect than its effect on wild type s␤APP secretion. These results are compatible with the finding that both mutant ␤APPs retain 5-10% of X11 binding activity (24).
To rule out the possibility that X11 affects protein secretion in a more generalized manner, pulse-chase experiments were performed with 293 cells stably overexpressing the human cystatin C gene, transiently transfected with X11 or vector cDNAs. X11 had no effect on the expression or secretion of cystatin C, a cysteine proteinase inhibitor, produced and secreted through the constitutive secretory pathway (Fig. 8).
These results suggest that ␤APP binding to X11 delays ␤APP turnover, resulting in retention of intracellular ␤APP and decreased ␤APP secretion.
Effect of Overexpression of X11 and ␤APP on A␤ Secretion-In order to test the effect of X11-␤APP binding on A␤ secretion, metabolic labeling experiments were performed. N2a cells permanently transfected with wild type ␤APP were transiently transfected with vector or with X11 cDNAs and labeled with [ 35 S]methionine/cysteine, and the conditioned medium proteins were immunoprecipitated with anti-A␤ 1-17 and anti-A␤ 17-27 antibodies (6E10 and 4G8) (Fig. 9A). X11 binding to wild type ␤APP diminished A␤ secretion. A decrease in A␤ secretion was also observed in media conditioned by COS-1 cells transiently cotransfected with X11 and wild type ␤APP, compared with cells transiently transfected with ␤APP cDNA alone (data not shown).
In order to confirm the finding that X11 binding to ␤APP causes decreased secretion of A␤ peptides, the levels of A␤ secreted into conditioned media were studied by ELISA. The levels of A␤ peptides secreted by N2a cells stably transfected with wild type ␤APP cDNA and transiently transfected with X11 cDNA were compared with cells transfected with the ␤APP cDNAs alone. The secretion of both A␤ 1-40 and A␤ 1-42 was lower in N2a cells transfected with X11 and wild type ␤APP as compared with cells transfected with ␤APP alone (Fig. 9B). These results demonstrate a decrease of 53 and 34% in the levels of A␤ 1-40 and A␤ 1-42 , respectively, secreted by cells transfected with wild type ␤APP and X11 in comparison to cells transfected only with wild type ␤APP.
These data suggest that X11 interaction with ␤APP participates in the regulation of ␤APP processing, affecting both s␤APP and A␤ secretion in neuronal and nonneuronal cells. DISCUSSION Our studies demonstrate that association of ␤APP with X11 in cultured cells affects ␤APP turnover and processing, which in turn shifts the balance toward decreased production of specific proteolytic peptides. The ␤APP binding site for X11 lies within the YENPTY motif present at the carboxyl terminus of ␤APP (23,24). Specific mutations within the binding motif inhibit interaction with X11 and result in increased secretion of soluble ␤APP via the ␣-secretase pathway. Cotransfection of , separated by 8% SDS-PAGE, and exposed to x-ray film. The protein bands were scanned and quantitated. Relative intensity of the bands was calculated as the percentage of the intensity of the protein band of cell lysates at the beginning of the chase (time 0). Mean and S.D. from 3-5 different experiments are presented. p values Ͻ0.05 for the difference between the experimental conditions were considered to be statistically significant (*). Symbols represent transfection with wild type ␤APP with vector (E) or wild type ␤APP together with X11 (q). wild type ␤APP and X11 cDNAs stabilizes cellular full-length ␤APP, down-regulates s␤APP secretion, and concomitantly triggers a decrease in A␤ secretion. These data suggest that ␤APP-X11 interaction plays a role in the regulation of ␤APP processing, differentially affecting s␤APP and A␤ secretion in neuronal and nonneuronal cells.
Numerous studies have established that stimulation of protein kinase C leads to cell type-specific enhancement of s␤APP with a concurrent suppression of A␤ production (40,41). The secretory processes of s␤APP and A␤ were also differentially affected by interleukin-1␤ stimulation (42). Thus, in certain cell systems s␤APP secretion and A␤ generation apparently are not mutually exclusive. These observations are the basis for the hypothesis that there is a surplus of full-length ␤APP, which gives rise to s␤APP, A␤, or lysosomal degradation products, depending on the enzymatic activities and regulatory mechanisms present (40). Because stimulation of protein kinase C causes a rapid reduction of cell surface ␤APP (43) and increases secretory vesicle formation (44), it was suggested that its effects are due primarily to a reduction in the transport of ␤APP to the cell surface. This alteration correlates with an acceleration of intracellular ␣-secretase ␤APP cleavage and of ␤APP trafficking in the exocytic pathway. In most cell cultures, displacement of cell surface ␤APP reduces the substrate available for endocytic processing with resultant inhibition of A␤ production (43).
The normal function of ␤APP within a signal transduction pathway may be elucidated by studying the interaction of ␤APP with PI/PTB-containing proteins. NPXY is a conserved protein binding motif that has been implicated in signal transduction (30,31) and internalization of ␤APP via clathrin-mediated endocytosis (11,18,19). X11 binds to the YENPTY motif within the cytoplasmic domain of ␤APP. Our data demonstrate that X11 modulates ␤APP processing via a direct protein/protein interaction. Because X11 has multiple protein interaction domains, other proteins may be complexed to the ␤APP-X11 moiety, affecting ␤APP processing or the normal physiologic function of ␤APP. X11 has two PDZ motifs implicated in the , separated by 8% SDS-PAGE, and exposed to x-ray film. The protein bands were scanned and quantitated. Relative intensity of the bands was calculated as a percentage of the intensity of the protein band of cell lysates at the beginning of the chase (time 0). The results from a single representative experiment are presented. Less than 10% variation was observed between experiments. Symbols represent transfection with wild type ␤APP with vector (E) or with X11 (q) and transfection with ␤APP N684A with vector (Ⅺ) or with X11 (f).
FIG. 8. Coexpression of X11 with cystatin C does not affect its expression and secretion. 293 cells stably overexpressing the human cystatin C gene that were transiently transfected with X11 or vector cDNAs were labeled with [ 35 S]methionine/cysteine for 20 min and chased for different time periods. Cell lysates (A) and medium proteins (B) were immunoprecipitated with anti-cystatin C antibody, separated on 10% Tris-Tricine gel, and exposed to x-ray film. The protein bands were scanned and quantitated. Relative intensity of the bands was calculated as percentage of the intensity of the protein band of cell lysates at the beginning of the chase (time 0). The results from a single representative experiment are presented. Less than 10% variation was observed between experiments. Symbols represent transfection with cystatin C with vector (E) or with X11 (q). binding of sequences found on Fas and N-methyl-D-aspartate receptors (45,46). These domains are also implicated in PDZ dimerization as exemplified in the interaction of PDZ domains on nitric oxide synthase and PSD95 (47). Thus, PDZ dimerization may induce stabilization of the complex ␤APP-X11-X11 dimerization associate within cellular membranes. Stabilization of the complex may block the accessibility of the secretase cleavage sites of ␤APP, resulting in decreased secretion of both s␤APP and A␤. However, X11 may directly affect ␤APP processing, preventing secretion of its soluble fragments and thereby leading to an accumulation of ␤APP in the cell.
Although numerous functions have been attributed to ␤APP, the precise physiological role of the full-length protein, as well as that of the proteolytic peptides, remains to be defined. Cellassociated ␤APP and secreted ␤APP products have been implicated in cell growth, neuronal survival and viability, neurite outgrowth, and neuroprotection (48 -51). s␤APP␣ more effectively protects hippocampal neurons against excitoxicity, A␤ toxicity, and glucose deprivation than does s␤APP␤. s␤APP␣ has trophic activity on various cell lines that was localized to amino acids 591-612 at the carboxyl terminus (48,52,53). Therefore, down-regulation of s␤APP␣ as a result of ␤APP-X11 coupling could result in reduced trophic activity. However, stabilization of the full-length protein within the cell surface may be neuroprotective.
The pathological processes that lead to the deposition of insoluble A␤ in the brain are unknown. Several observations indicate that aberrant expression or processing of the human ␤APP gene products may lead to the development of AD. Patients with trisomy 21 (Down's syndrome patients) have overexpression of the protein and tend to develop AD-type pathology early in life (54,55). Several familial forms of Alzheimer's disease have been described that are linked to mutations in the ␤APP gene that result in amino acid substitutions within A␤ or near its amino-or carboxyl-terminal ends. A Swedish kindred develops an AD phenotype that cosegregates with a double mutation, K595N/M596L, amino-terminal to the A␤ region (56). This double mutation leads to overproduction of A␤ and s␤APP␤ (57,58). Moreover, transgenic mice overexpressing human ␤APP exhibit age-dependent and brain region-specific extracellular deposits of fibrillar A␤ (59 -61). Thus, perhaps even slightly increased amounts of A␤ might be sufficient to cause AD, and inhibition of A␤ production could aid in prevention of AD. It was hypothesized that A␤ and s␤APP are physiological ligands with reciprocal effects on neurons and that alterations in the A␤ to s␤APP ratio may have beneficial or deleterious effects on neuronal survival and growth (62). Thus, stabilization of cellular ␤APP and inhibition of the amyloidogenic processing pathway by stimulating the interaction of ␤APP with an adaptor protein, such as X11 or another family member, may provide a novel approach for the pharmacological modulation of ␤APP processing in Alzheimer's disease. FIG. 9. Reduced secretion of A␤ in cells overexpressing both X11 and wild type ␤APP. N2a cells stably transfected with wild type ␤APP and transiently transfected with vector or X11 cDNAs were labeled with [ 35 S]methionine/cysteine for 4 h. Conditioned medium proteins were immunoprecipitated with anti-A␤ 1-17 and anti-A␤ [17][18][19][20][21][22][23][24][25][26][27] antibodies (6E10 and 4G8), separated in 16.5% Tris-Tricine gel, and exposed to x-ray film. The results from a single metabolic labeling representative experiment are presented. Less than 10% variation was observed between experiments (A). N2a cells stably transfected with wild type ␤APP were transiently transfected with vector (Ⅺ) or X11 (o) cDNAs. Three anti-A␤ antibodies were used for ELISA analysis of A␤ medium proteins (B): anti-A␤ 1-10 antibody (BAN-50), anti-A␤ x-40 antibody (BA-27), and anti-A␤ x-42 antibody (BC-05). The concentrations of A␤ 1-40 and A␤ 1-42 are presented in fmol/ml/h as the mean Ϯ S.D. of seven different experiments.