The endocytotic pathway is required for increased Aβ42 secretion during apoptosis

Molecular Brain Research 128 (2004) 201 – 211 www.elsevier.com/locate/molbrainres

Research report

The endocytotic pathway is required for increased Ah42 secretion during apoptosis Chhinder P. Sodhi a, Srinivas Rampalli a, Ruth G. Perez b, Edward H. Koo c, Bruce Quinn a, Numa R. Gottardi-Littell a,* a

Department of Pathology, Northwestern University Feinberg School of Medicine, 303 East Chicago Avenue, Chicago, IL 60611, USA b Department of Neurology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA c Department of Neurosciences, University of California at San Diego, La Jolla, CA, USA Accepted 17 June 2004 Available online 20 July 2004

Abstract Secretion and progressive cerebral accumulation of h-amyloid peptides (Ah), which derive by endoproteolytic (‘amyloidogenic’) processing of h-amyloid precursor protein (APP), are felt to represent collectively an early and necessary event in the pathogenesis of Alzheimer’s disease. APP amyloidogenic processing can occur via secretory or endocytotic pathways, but the relative contribution of these pathways to Ah secretion remains to be established. The effect of apoptosis on amyloidogenic processing and Ah secretion similarly is incompletely understood. We tested the hypothesis that APP processing by the endocytotic pathway represents a stress-related neural cell response, by comparing Ah secretion after induction of apoptosis in PC12 cells transfected either for endocytosis-competent or -deficient APP. Newly prepared adenoviral vectors encompassing targeted mutagenesis of the cytoplasmic tail YENP tetrapeptide sequence, which serves as the principal APP internalization signal, were used to express endocytosis-deficient holoprotein. We report that the endocytotic pathway is required for the generation and secretion of Ah42, and that secretion of this neurotoxic peptide increases significantly during apoptosis. We demonstrate additionally that more Ah40 apparently is generated in secretory compartments during apoptosis when APP processing by the endocytotic pathway is impaired. D 2004 Elsevier B.V. All rights reserved. Theme: Disorders of the nervous system Topic: Degenerative disease: Alzheimer’s-beta amyloid Keywords: Adenoviral vector; Alzheimer’s disease; Apoptosis; APP; h-amyloid; Endocytosis

1. Introduction Alzheimer’s disease (AD), a neurodegenerative disorder of unknown etiology, represents the most common cause of dementia in the elderly, with prevalence rate approaching 10% in the population over 65 years of age. Considerable evidence exists that secretion and progressive cerebral accumulation of h-amyloid peptide(s) (Ah), which derive by endoproteolytic processing from a type I transmembrane glycoprotein of poorly (incompletely) characterized physiologic function termed h-amyloid precursor protein or APP, represent collectively an early and necessary event in the molecular pathogenesis of AD. Secreted Ah, in aggregated * Corresponding author. Tel.: +1-312/926-9487; fax: +1-312/926-9830. E-mail address: [email protected] (N.R. Gottardi-Littell). 0169-328X/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.molbrainres.2004.06.012

form (e.g., soluble oligomers), is toxic to neurons, as clearly has been established [8,30]. Moreover, of the two principal species of secreted Ah, namely, Ah1-40 (Ah40) and Ah1-42 (Ah42), the longer and more hydrophobic Ah42 represents the principal neurotoxic species (see Ref. [20]). APP amyloidogenic processing (i.e., endoproteolytic processing that results in the generation of Ah) is complex and incompletely understood, particularly in neural cells. Amyloidogenic processing can occur in secretory or endocytotic compartments, but the relative contribution of the secretory and endocytotic pathways to Ah (including Ah42) secretion by neurons and non-neuronal cells is not known. The secretory pathway may be the principal route for Ah42 generation, based on the finding of some studies [24] that expression of APP with targeted endoplasmic reticulum retention signal results in markedly increased Ah42 secre-

202

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

tion relative to wild type holoprotein. Other related studies [3], however, fail to confirm this finding, and, additionally, demonstrate that the endocytotic pathway contributes significantly to Ah42 generation, based on markedly decreased Ah42 secretion with expression of APP minus the cytoplasmic tail (i.e., lacking all potential signals for internalization from the cell surface). Similarly and more specifically, targeted mutagenesis of the cytoplasmic tail YENP tetrapeptide sequence that functions as the principal APP internalization signal, as elucidated by Perez et al. [17], results in markedly decreased Ah42 secretion, providing further evidence that the endocytotic pathway participates significantly in the generation and secretion of this neurotoxic species. In the present study, we test the hypothesis that neural cells (e.g., vulnerable subpopulations of cerebral cortical neurons) respond to cytotoxic stimuli with increased Ah (including Ah42) secretion originating from APP processing by the endocytotic pathway. Two previous, independent reports that cultured central nervous system neurons undergoing apoptosis have 3- to 4-fold increased Ah secretion [6,13] did not address specifically Ah42 secretion or contribution by the endocytotic pathway. We have performed comparative analysis of Ah (including Ah42) secretion after induction of apoptosis in PC12 cells transfected in parallel for endocytosis-competent or -deficient APP using adenoviral vectors. We report that the endocytotic pathway is required for, taken collectively, the generation and secretion of Ah42, and that secretion of this neurotoxic peptide increases significantly during apoptosis.

2. Materials and methods 2.1. Antibodies Polyclonal antibody or pAb CT15 [21] recognizes the Cterminal 15 amino acid residues of APP, and was used to immunoprecipitate holoprotein or C-terminal fragments from cell lysates; CT15 is specific for APP and does not cross-react with its homologue APLP-2 [23]. Monoclonal antibodies or mAb 4G8 and 6E10 (Signet Pathology Systems) recognize, respectively, amino acid residues 17 –24 and 1 –17 of the APP Ah sequence, and were used to immunoprecipitate, respectively, Ah (total species) and APPs-a from conditioned medium. Monoclonal antibody 22C11 (Chemicon) recognizes an extreme N-terminal epitope in the APP extracellular domain, and was used to detect holoprotein from cell lysates by Western blot analysis. Monoclonal antibody to h-actin (Sigma) was used for Western blot analysis. 2.2. Cell culture Naive (undifferentiated) PC12 cells were grown on collagen-coated dishes in RPMI 1640 medium supplemented with 10% horse serum (v/v; JRH Biosciences),

5% fetal bovine serum (U.S. Biotechnologies), HEPES buffer (10 mM), L-glutamine (2 mM), and penicillin-streptomycin (100 units/ml each); medium was changed on alternate days and cells were passaged weekly. Human embryonic kidney (HEK) 293 cells used to propagate recombinant adenovirus (see below) were grown in Dulbecco’s modified eagle medium (high glucose, Invitrogen) supplemented with 10% fetal bovine serum and penicillinstreptomycin (100 units/ml each); medium was changed every third day and cells were passaged weekly. 2.3. Preparation of adenoviral vectors Replication-deficient, recombinant adenoviruses that express wild type (designated APPwt) or mutant human APP751 (see below), as well as control adenovirus that expresses E. coli h-galactosidase (designated lacZ), were prepared by direct ligation using the Adeno-X Expression System kit (Clontech). Adenoviruses Y738A, N740A, and 738/741 (denotes Y738A/P741A double mutation) express, respectively, APP-751 with corresponding mutation(s) in the cytoplasmic tail YENP tetrapeptide sequence that functions as the principal APP internalization signal [17]. Adenovirus V717F expresses APP-751 with corresponding mutation linked to familial Alzheimer’s disease or FAD [16]. Briefly, the respective h-galactosidase or APP-751 expression cassettes, which include a 5V CMV immediate – early promoter and a 3V polyadenylation sequence, were inserted into the E1a genomic region of human type 5 adenovirus (Ad5), and the resulting constructs were propagated in permissive (HEK 293) cells. Prior to propagation in HEK 293 cells, Western blot and cDNA partial sequence analysis were carried out for all APP-751 constructs to confirm, respectively, expression of full-length protein and incorporation of desired mutation(s) (absence thereof for APPwt; see above). Serial dilution, plaque formation assay after final propagation revealed titer of 109 PFU/ml for all APP-751 adenoviruses (1010 PFU/ml for lacZ). The APP-751 adenoviruses (including APPwt) and lacZ were prepared in the same genomic background as used for adenovirus Ad5/CMV-APP that expresses wild type human APP-695, prepared previously as described [35], with single exception that all new adenoviruses incorporate a larger deletion in the E3 genomic region, which is protective against the emergence of viral wild type recombinants during propagation (see Ref. [22]). 2.4. Transfection of PC12 cells using adenoviral vectors Naive PC12 cells (80% confluent) were infected with APP-expressing (or control) adenoviral vectors at 50:1 multiplicity-of-infection (50 MOI) for 90 min at 37 jC using minimal volume of medium (RPMI 1640 containing 5% serum, 2:1 horse/fetal bovine), then returned to usual volume of medium (at 37 jC) and analyzed after 48 h. Infection parameters as indicated were derived from studies with control adenoviral vector (lacZ) as follows. Cells were

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

infected with lacZ at concentration ranging from 10 to 100 MOI for 90 min (37 jC), then fixed (4% formaldehyde prepared freshly from paraformaldehyde powder) and incubated with x-gal reagent after 24, 48, and 60 h to determine respective infection fractions (i.e., percent x-gal positive or ‘blue’ cells); cell viability after 24, 48, and 60 h was estimated by trypan blue (0.4%) exclusion. 2.5. Induction and quantification of apoptosis To induce apoptosis, APP-transfected PC12 cells (see above) were treated with the broad-spectrum protein kinase inhibitor staurosporine [12], which was added to the culture medium 36 h after infection with recombinant adenovirus (1 AM final concentration, 12 h total exposure at 37 jC). Induction of apoptosis was confirmed by DNA oligo-nucleosomal ‘laddering’ using the Apoptotic DNA Ladder kit (Roche). Briefly, harvested DNA (3 –5 Ag) was treated with RNase A (Qiagen) and separated by electrophoresis (2% agarose gel in TBE buffer containing 0.5 Ag/ ml ethidium bromide). Additionally, quantitative assessment of apoptotic fraction was carried out by annexin V labeling [27] using the Annexin V-FITC kit (Immunotech) as follows. Cells were harvested by aspiration with PBS containing 2 mM EDTA, washed twice with PBS/EDTA at 4 jC, and incubated for 10 min at 4 jC (in dark) with Binding Buffer (100 Al; Immunotech) containing FITCconjugated annexin V (1 Ag/ml) and propidium iodide/PI (12.5 Ag/ml). After dilution with Binding Buffer (0.4 ml), analysis was carried out using a FACScan flow cytometer (Becton Dickinson Pharmigen). 2.6. Metabolic labeling (steady-state and pulse) For steady-state metabolic labeling, APP-transfected PC12 cells (see above) were incubated with Trans 35S-Label reagent (ICN) in serum- and methionine/cysteine/cystinefree medium 36 h after infection with recombinant adenovirus (250 ACi/ml final concentration, 12 h total exposure at 37 jC), preceded by washing and incubation for 20 min in serum- and met/cys-free medium. For pulse metabolic labeling to determine the half-life of APP holoprotein, cells were incubated for 20 min with 200 ACi/ml Trans 35S-Label reagent 44 h after infection with recombinant adenovirus, then chased for 1, 2, or 4 h (0 h, no chase); for each experimental condition, half-life calculations were normalized for respective phosphorimage signals at 0 h.

203

of solubilization buffer (10 mM Tris, pH 7.4; 5 mM EDTA; 2 mM EGTA; 0.1 mM DTT) containing 1% NP-40 and 1 mM PMSF with continuous mixing for 60 min at 4 jC; lysates were cleared by centrifugation (16 000  g for 20 min) and protein concentration was determined using BCA reagent (Pierce). For detection of C-terminal fragments in steady-state labeled cells, 25 Ag of lysate protein were used; for detection of holoprotein in pulse-labeled cells, 50 Ag were used. After extensive pre-clearing with protein Asepharose (Amersham Biosciences), lysates were incubated for 2 h at 4 jC with CT15 conjugated to protein Gsepharose (rabbit IgG-conjugated beads were used as negative control). At completion of the incubation, immunobeads were washed extensively then heated (95 jC for 10 min) to release bound proteins, which were separated by SDS-PAGE and detected by exposure to X-ray film (see below). For C-terminal fragments, 16.5% Tris/tricine gel [19] was used; for holoprotein, 7.5% Laemmli gel was used. Holoprotein additionally was detected from lysates (50 Ag protein/lane of 7.5% Laemmli gel) by Western blot analysis with mAb 22C11, using Hybond membrane and ECL kit (Amersham Biosciences). Ah (total species) and APPs-a were detected from conditioned medium of APP-transfected, metabolically labeled cells by immunoprecipitation, respectively, with mAb 4G8 and 6E10 conjugated to protein G-sepharose, as described above for holoprotein or C-terminal fragments (mouse IgG conjugated beads were used as negative control). Conditioned medium was centrifuged (5000  g for 20 min at 4 jC) to remove cell debris prior to pre-clearing with protein A-sepharose, but otherwise was used at original (protein) concentration. For Ah, 300 Al of conditioned medium were used, and for APPs-a, 20 Al were used; the amounts of conditioned medium used in all comparative studies were normalized for respective protein concentration of corresponding cell lysates. To resolve Ah40 and Ah42 species after immunoprecipitation, 10% bicine/urea gel [33] was used; 16.5% Tris/tricine gel was used for total Ah. Laemmli gel (7.5%) was used for APPs-a. At completion of SDS-PAGE, gels were fixed, impregnated with fluorographic amplification reagent (Amersham Biosciences), dried, and exposed to BioMax MR film (Kodak; TranScreen LE was used) at 80 jC. Eagle Eye II (Stratagene) imaging system was used for densitometric (comparative) analysis of Ah phosphorimage signal (system also used for analysis of APP holoprotein and APPs-a signals in pulse-chase studies). 2.8. Detection of APP and BACE expression

2.7. Detection of APP holoprotein and processing derivatives APP holoprotein or C-terminal fragments were detected from lysates of transfected, metabolically labeled PC12 cells (see above) by immunoprecipitation with pAb CT15 as follows. Cells were scraped from dishes, washed twice with PBS containing 1 mM PMSF at 4 jC, and lysed by addition

To detect expression of transgene-derived (human) APP and endogenous BACE-1 in PC12 cells, RT-PCR analysis was carried out as follows. Total RNA isolated with Trizol reagent (Gibco BRL) was processed for cDNA synthesis using reverse primer (see below) and reagents provided with GeneAmp RNA PCR Core kit (Perkin-Elmer), including MuLV reverse transcriptase. PCR product subsequently was

204

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

generated using forward primer (see below) and AmpliTaq DNA polymerase (provided with GeneAmp kit). Respective PCR products were separated by electrophoresis (1.0 – 1.5% agarose gel), identified with ethidium bromide, and visualized with UV trans-illuminator. For APP, the following primer set was used, based on the cDNA sequence for human APP-751, as reported [18]: forward primer, 5V-TGA GAA GAG TAC CAA CTT GCA TGA CTA CGG-3V; reverse primer, 5V-ATG TTC ATT CTC ATC CCC AGG TGT CTC GAG-3V (601 – 1231; spans sequence corresponding to KPI-domain; expected product size, 630 bp). For BACE-1, the following primer set was used, based on the cDNA sequence from rat, as reported [29]: forward primer, 5V-ATT CCC TAT ACA CTG GCA GTC TC-3V; reverse primer, 5V-TTG TAG CAA CAG TCT TCC ATG TC-3V (1152 –1764; expected product size, 614 bp).

and non-infected controls (latter not shown). Baseline (i.e., minus staurosporine) apoptotic fraction of 20– 25% in APPtransfected cells, as well as lacZ-transfected and noninfected controls, reflects increased background inherent with the annexin V assay in PC12 cells. Clear oligonucleosomal ‘laddering’ was seen after staurosporine treatment in APP-transfected cells (Fig. 1c), as well as lacZ-

3. Results 3.1. Staurosporine treatment induces apoptosis uniformly in PC12 cells transfected for endocytosis-competent or -deficient APP Adenovirus-derived vectors were used to modulate APP expression in naive (undifferentiated) PC12 cells. Adenoviral vectors were selected because they readily infect neural cells with minimal toxicity and do not disrupt or otherwise alter usual endocytotic trafficking of endogenous APP [35]. Five replication-deficient, recombinant adenoviruses were prepared in APP-751 background (see Materials and methods), comprising vectors N740A, Y738A, and 738/741 that express endocytosis-deficient APP, and vectors APPwt and V717F that express endocytosis-competent APP. Pilot studies (data not shown) with control adenoviral vector lacZ at titer 50 MOI (multiplicity-of-infection) revealed peak infection fraction of 70 –80% occurring approximately 48 h after infection, which correlated tightly with detection of secreted APP processing derivatives (see below) as determined by time-course study with vector APPwt. Of note, identical time-course results for secretion of APPs-a and Ah (total species) were observed after transfection with vector APPwt or previously reported vector Ad5/CMV-APP, which expresses wild type APP-695 [35], consistent with the notion that, at least within experimental parameters for this communication, principal non-neuronal isoform (i.e., APP751; as carried out for this communication) and neuronal isoform (i.e., APP-695) are processed comparably. To induce apoptosis, APP-transfected PC12 cells were treated for 12 h at 1 AM concentration with the broadspectrum protein kinase inhibitor staurosporine (see Fig. 1). Quantitative analysis using annexin V binding revealed uniform apoptotic fraction of 70– 75% after staurosporine treatment in cells transfected, respectively, with all five newly prepared APP vectors (Fig. 1a, b), which was identical to that seen in control vector (lacZ) transfected

Fig. 1. Staurosporine treatment generates uniform apoptotic fraction in APP-transfected cells. Naive PC12 cells (60 – 70% confluent) were infected with respective APP-expressing adenoviral vectors or control (lacZ) vector at 50 MOI, treated after 36 h with staurosporine (1 AM for 12 h), then analyzed quantitatively for apoptosis induction by annexin V binding (a, b) and, qualitatively, by DNA ‘laddering’ (c). (a) Representative biparametric flow cytograms with or without staurosporine treatment (shown for APPwt transfectants). The lower right and upper right quadrants represent early and late apoptotic fractions, respectively; lower left quadrants represent baseline fraction (i.e., viable cells; negative for annexin V and PI/propidium iodide), whereas upper left quadrants represent necrotic fraction (i.e., cells damaged during the assay; positive for PI). (b) Bar graph depicting respective apoptotic fractions (early plus late) with or without staurosporine treatment for control or APP-vector transfectants (results shown represent mean F SD from four independent observations). Note that staurosporine treatment results in uniform apoptotic fraction of 70 – 75% for all transfectants (identical for non-infected control, not shown). (c) Conventional agarose gel electrophoresis of DNA extracted from control or APP-vector transfectants with (+) or without ( ) staurosporine treatment. Note that clear oligo-nucleosomal ‘laddering’ is present after staurosporine treatment.

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

205

transfected and non-infected controls. Cell viability by trypan blue exclusion (data not shown) was approximately 70% after staurosporine treatment for all transfectants and non-infected control, with caveat that significant cell clumping, which occurred independently from adenoviral infection, made assessment somewhat difficult; identical baseline viability in excess of 80% was observed for all transfectants and non-infected control. 3.2. Induction of apoptosis results in increased Ab42 secretion after transfection for endocytosis-competent APP To determine whether induction of apoptosis results in increased APP amyloidogenic processing by the endocytotic pathway, comparative analysis of Ah secretion was carried out in (naive) PC12 cells transfected in parallel for endocytosis-competent or -deficient APP. Secreted Ah (total species) was immunoprecipitated with mAb 4G8 from conditioned medium of APP-transfected, metabolically labeled cells in the presence or absence of staurosporine treatment, as above (see Fig. 2). Baseline (i.e., minus staurosporine) Ah secretion was decreased in cells transfected for endocytosis-deficient APP, relative to endocytosis-competent transfectants, with maximal effect for vector N740A and intermediate effect for vectors Y738A and 738/ 741, strictly concordant with the results of Perez et al. [17] in stably transfected CHO cells. Baseline p3 secretion, as anticipated (see [17]), was increased commensurately in cells transfected for endocytosis-deficient APP (data not shown). With respect to endocytosis-competent isoforms, baseline Ah secretion was decreased somewhat relative to wild type in cells transfected with vector V717F, which expresses the corresponding FAD-linked mutant APP. Ah secretion after staurosporine treatment, however, was increased significantly in cells transfected for endocytosiscompetent as well as endocytosis-deficient APP, with greater differential effect for the latter (7-fold increase after staurosporine for N740A transfectants, versus 3- and 4-fold increase for APPwt and V717F transfectants, respectively; 5-fold increase for Y738A and 738/741 transfectants; Fig. 2a). To determine whether Ah40 and Ah42 secretion are modulated (up-regulated) differentially during apoptosis, total Ah immunoprecipitated from conditioned medium was separated with 10% bicine/urea gel [33], which resolves these two species (see Fig. 3). Baseline Ah42 secretion, expressed as fraction of total Ah secreted, was decreased significantly in cells transfected for endocytosis-deficient APP, with maximal effect for vector N740A and intermediate effect for vectors Y738A and 738/741, strictly concordant, again, with the results of Perez et al. [17]. Most interestingly, and in contrast to results obtained for total Ah secreted, Ah42 secretion after staurosporine treatment was increased significantly only in cells transfected for endocytosis-competent APP. Moreover, cells transfected with vector V717F had increased baseline Ah42 secretion

Fig. 2. Total Ah secretion increases during apoptosis in cells transfected for endocytosis-competent or -deficient APP. Naive PC12 cells were transfected in parallel for endocytosis-competent (adenoviral vectors APPwt, V717F) or -deficient (vectors N740A, Y738A, 738/741) APP (lacZ, control vector), labeled metabolically, treated with staurosporine to induce apoptosis, then analyzed for Ah secretion (total species) by immunoprecipitation from conditioned medium using mAb 4G8 (a, autoradiogram of immunoprecipitated Ah with or without staurosporine treatment after separation by SDS-PAGE, 16.5% Tris-tricine; b, bar graph depicting comparative Ah secretion by densitometry, normalized for baseline secretion in APPwt transfectants). Note that baseline Ah secretion (total species) is reduced significantly for endocytosis-deficient transfectants (maximal effect for N740A), concordant with findings by Perez et al. [17], but increases significantly after staurosporine treatment both for endocytosis-competent and -deficient transfectants, with increased differential effect for the latter (7-fold increase after staurosporine for N740A versus 3to 4-fold increase for APPwt, V717F). Amounts of conditioned medium used were normalized for respective protein concentration of corresponding cell lysates. Results shown for b represent meanFSD from three independent observations (*, significant at p<0.05 or better relative to wild type; **, significant at p<0.001 relative to baseline).

relative to wild type transfectants, consistent with wellestablished effect for FAD-linked mutant APP (see Ref. [20]), but, interestingly, Ah42 secretion during apoptosis was increased comparably for both (3.6- and 4-fold increase after staurosporine for V717F and APPwt transfectants, respectively). 3.3. Increased Ah42 secretion during apoptosis is directly dependent on endocytosis-competent APP isoform To confirm that the results for comparative analysis of Ah (including Ah42) secretion, as above, are directly dependent on the specific APP transgene expressed, additional studies were carried out as follows. By RT-PCR analysis, APP expression was increased significantly and comparably in cells transfected respectively with all five newly prepared adenoviral vectors, and remained otherwise unchanged after staurosporine treatment (see Fig. 4a). APP

206

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

Fig. 3. Ah42 secretion increases during apoptosis only in cells transfected for endocytosis-competent APP. Total Ah species immunoprecipitated from conditioned media of naive PC12 cells transfected in parallel for endocytosis-competent or deficient APP with or without staurosporine treatment (see above, Fig. 2) were separated by SDS-PAGE using 10% bicine/urea system [33], which resolves Ah40 and Ah42 (a, autoradiogram of resolved Ah40, Ah42 species; b, bar graph depicting Ah42 as ratio of total Ah secreted by densitometry). Note that baseline Ah42 secretion is reduced significantly for endocytosis-deficient transfectants (maximal effect for N740A), again concordant with findings by Perez et al. [17], and especially, that Ah42 secretion increases significantly during apoptosis only for endocytosis-competent transfectants. Note also that baseline Ah42 secretion for V717F transfectants is increased relative to wild type, consistent with well-established effect for FAD-linked mutant APP, but increases comparably for both after staurosporine treatment (3.6- and 4-fold increase for V717F and APPwt, respectively). Amounts of conditioned medium used were normalized for respective protein concentration of corresponding cell lysates. Results shown for b represent mean F SD from three independent observations (*, significant at p < 0.05 or better relative to wild type; **, significant at p < 0.001 relative to baseline).

holoprotein, detected by Western blot analysis from cell lysates, was increased commensurately in all five transfectants, and, interestingly, was increased significantly (uniformly) over baseline level after staurosporine treatment (Fig. 4b; note that a portion of this apparent holoprotein increase after staurosporine treatment likely reflects increased lysate-associated APPs-a). Similarly, secreted APPs-a and corresponding membrane-retained, C-terminal fragment (a-CTF), detected by immunoprecipitation from conditioned medium or cell lysates with mAb 6E10 or pAb CT15, respectively, were increased significantly over baseline after staurosporine treatment (Fig. 4c, d; note somewhat greater differential effect for APPs-a in N740A transfectants relative to APPwt transfectants). These data suggest that the turnover of APP holoprotein is modulated (decreased) during apoptosis. Pulse-chase analysis was used to determine the half-life of APP holoprotein with or without staurosporine treatment (see Fig. 5), which was calculated on the basis of respective signals at time 0 (i.e., data were normalized for each

experimental condition). Baseline holoprotein half-life in cells transfected for endocytosis-deficient APP was decreased significantly (approximately 50%; Fig. 5a –c, upper panels, left aspect and Fig. 5d – f; compare especially respective upper panels, left aspect of Fig. 5a and b) relative to endocytosis-competent transfectants (60 F 7.2 min for N740A transfectants, versus 93 F 11 min and 85 F 6.9 min for APPwt and V717F transfectants, respectively; ANOVA, p < 0.01), with commensurately increased APPs-a secretion (Fig. 5a –c, lower panels, left aspect and Fig. 5j – l; compare especially respective lower panels, left aspect of Fig. 5a and b). As anticipated, holoprotein halflife was increased significantly over baseline level after staurosporine treatment in cells transfected for endocytosiscompetent APP (Fig. 5a and c, upper panels, right aspect and Fig. 5d and f), with somewhat greater differential effect for APPwt transfectants (compare especially respective upper panels, right aspect of Fig. 5a and c) but, interestingly, comparatively less apoptosis-related increase in holoprotein half-life was evident for endocytosis-deficient (N740A; Fig. 5b, upper panel, right aspect and Fig. 5e) transfectants (180 F 30 min and 150 F 9.8 min for APPwt and V717F transfectants, respectively, versus

Fig. 4. APP transgene expression is up-regulated comparably for all adenoviral vectors and remains unchanged during apoptosis, but levels of APP holoprotein and processing derivatives are increased during apoptosis. Naive PC12 cells transfected in parallel for endocytosiscompetent or -deficient APP were analyzed for APP mRNA expression (a), APP holoprotein in cell lysates (b), APPs-a secretion (c), and APP Cterminal fragments (a- and h-CTF) in cell lysates (d) with or without staurosporine treatment. APP mRNA levels, detected by RT-PCR analysis using 1 Ag total cellular RNA and primer set flanking cDNA sequence corresponding to the KPI-domain (expected product size, 630 bp), increase uniformly for all APP-vector transfectants and remain otherwise unchanged after staurosporine treatment (a; h-actin, positive control, primer set designed from rat sequence, NCBI accession number NM_031144, expected product size, 800 bp). In contrast, APP holoprotein levels, detected by Western blot analysis using mAb 22C11, increase significantly after staurosporine treatment for all transfectants (b; 7.5% Laemmli gel). Secreted APPs-a, detected by immunoprecipitation from conditioned media using mAb 6E10, also increases significantly after staurosporine treatment (c; note somewhat greater differential effect for N740A transfectants relative to wild type; 7.5% Laemmli gel), with commensurate increase in corresponding Cterminal fragment (a-CTF), detected by immunoprecipitation with pAb CT15 (d; 16.5% Tris-tricine gel).

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

207

Fig. 5. Half-life of APP holoprotein is increased during apoptosis, but less so for endocytosis-deficient isoform, with commensurately increased APPs-a secretion. Naive PC12 cells transfected in parallel for endocytosis-competent (vectors APPwt, V717F) or -deficient (vector N740A) APP were pulse-labeled metabolically (200 ACi/ml [35S]Label for 20 min) 44 h after infection with (+) or without ( ) staurosporine treatment, chased with serum-free medium for 1, 2, or 4 h (0 h, no chase), then analyzed for APP holoprotein in cell lysates (a – c, upper panel; 7.5% Laemmli gel) and APPs-a secretion (a – c, lower panel; 7.5% Laemmli gel). d – f, Bar graphic depiction of APP holoprotein (mature plus immature) levels during chase by densitometry. g – i, Line graphic depiction of APP holoprotein mature/immature ratio during chase by densitometry. j – l, Line graphic depiction of secreted APPs-a (as fraction of holoprotein signal at time 0) during chase by densitometry. Note that the half-life of APP holoprotein increases significantly after staurosporine treatment, but considerably less so for endocytosis-deficient (N740A) isoform, which otherwise reveals commensurately increased APPs-a secretion. Results shown for d – l represent mean F SD from four independent observations (data are normalized for signal at time 0 for each experimental condition).

80 F 5.6 min for N740A transfectants; ANOVA, p < 0.01; data represent mean of two observations each from two independent experiments). Moreover, the ratio of mature/ immature APP holoprotein 2 h after pulse labeling during baseline and after staurosporine treatment was decreased significantly for N740A transfectants, relative to APPwt and V717F transfectants (see especially Fig. 5g –i). These data indicate that the turnover of endocytosis-deficient APP is increased relative to endocytosis-competent holoprotein, with comparatively increased a-secretase processing, and additionally, that decreased holoprotein turnover as occurs during apoptosis is comparatively less for endocytosisdeficient APP.

Relatedly, to determine if increased amyloidogenic processing and Ah secretion during apoptosis are directly dependent on increased h-secretase expression (i.e., increased synthesis of BACE-1 mRNA), RT-PCR analysis for endogenous BACE-1 [29] gene expression was carried out with or without staurosporine treatment (see Fig. 6). Baseline BACE-1 expression was comparable in all APP vector-transfectants, as well as lacZ-transfected and noninfected controls (latter not shown), and was not increased after staurosporine treatment (in fact, was decreased uniformly after staurosporine treatment). Expression of a ‘housekeeping’ gene (h-actin) remained unchanged after staurosporine treatment. No product was detected when

208

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

Fig. 6. BACE-1 expression (endogenous) is not up-regulated during apoptosis. Naive PC12 cells transfected in parallel for endocytosis-competent or -deficient APP were analyzed for (endogenous) BACE-1 mRNA expression with or without staurosporine treatment by RT-PCR analysis, using 1 Ag total cellular RNA and primer set designed from the rat cDNA sequence (expected product size, 614 bp). Note that BACE-1 mRNA levels do not increase (in fact, decrease uniformly) after staurosporine treatment (h-actin, positive control).

total RNA was processed for PCR without reverse transcription, excluding contamination with genomic DNA (data not shown). These data indicate that increased Ah secretion during apoptosis is not directly related to up-regulation of BACE-1 gene expression.

4. Discussion We test, in the present study, the hypothesis that neural cells respond to cytotoxic stimuli with increased Ah secretion originating from APP processing by the endocytotic pathway. We have carried out comparative analysis of Ah (including Ah42) secretion after induction of apoptosis in naive PC12 cells transfected in parallel for endocytosiscompetent or -deficient APP using adenovirus-derived vectors. We report that the endocytotic pathway is required for, taken collectively, the generation and secretion of Ah42, and that secretion of this neurotoxic peptide increases significantly during apoptosis. Our results indicate additionally that more Ah40 apparently is generated in secretory compartments during apoptosis when APP processing by the endocytotic pathway is impaired. The relative contribution of the secretory and endocytotic pathways to APP amyloidogenic processing and Ah secretion has not been established unequivocally, but our results support strongly and expand on the findings of Perez et al. [17] that the endocytotic pathway participates significantly in the generation and secretion of Ah, including the principal neurotoxic species, Ah42. Of note, our results address directly secreted Ah, and not intracellular compartments where these peptides actually are generated. Relatedly, we do not address direct effect(s) of intracellular Ah40 and Ah42 levels, including potential neurotoxic effect of free (cytosolic) Ah42. Previous reports that Ah42 in large part is generated in secretory compartments, namely, the endoplasmic reticulum/intermediate compartment (ER/IC) [24,32], especially in neurons [2,9], based primarily on the finding that increased Ah42 secretion results from APP retention in these compartments (i.e., expression of APP with targeted ER-retention signal, treatment with brefeldin A), did not address the role of unique cytoplasmic tail sorting signals (e.g., YENP tetrapeptide sequence) in APP processing, as

was addressed previously by Perez et al. [17], and by our laboratory vis-a-vis the present report. Of note, Cupers et al. [3] found that expression of APP with ER-retention signal in neurons resulted in significantly decreased Ah secretion (total species; treatment with brefeldin A, interestingly, restored secretion of Ah40 and Ah42), and, additionally, that expression of APP minus the cytoplasmic tail (i.e., with deletion of all potential internalization signals) resulted in significantly decreased Ah secretion, including Ah42. The effect of apoptosis on APP amyloidogenic processing and Ah secretion similarly is incompletely understood. Two previous, independent reports [6,13] that central nervous system neurons undergoing apoptosis have 3- to 4-fold increased Ah secretion did not address specifically Ah42 secretion or contribution by the endocytotic pathway. We demonstrate that Ah42 secretion, expressed as fraction of total Ah secreted, increases 3- to 4-fold during staurosporine-induced apoptosis in naive PC12 cells, as a direct consequence of APP processing by the endocytotic pathway. Of note, Ah42 secretion increases comparably during apoptosis in cells transfected for wild type or FAD-linked (V717F) mutant APP, even though baseline Ah42 secretion is increased relative to wild type for V717F transfectants, consistent with well-established effect for FAD-linked mutant APP (see Ref. [20]). Total Ah secretion also increases significantly during apoptosis, but somewhat surprisingly, with greater differential effect in cells transfected for endocytosis-deficient APP (7-fold increase) compared to endocytosis-competent transfectants (3- to 4-fold increase). Moreover, the relative increase in holoprotein half-life during apoptosis is significantly less in cells transfected for endocytosis-deficient APP compared to endocytosiscompetent transfectants. Taken together, these results suggest that more Ah40 is generated in secretory compartments during apoptosis when APP processing by the endocytotic pathway is impaired. Our results do not address participatory role of caspases in the apoptosis-related (increased) Ah secretory response. Specifically, potential modulatory effect of caspase inhibition was not tested, and monitoring for APP-derived peptide C31 (see below) was not carried out. Caspase-mediated cleavage of APP at aspartyl residue 720 in the cytoplasmic domain (APP-751 numbering) has been demonstrated pre-

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

viously by several laboratories [7,14], but this cleavage (as well as corresponding cleavage at two caspase-consensus sites in APP extracellular domain) likely does not contribute directly to amyloidogenic processing and Ah secretion [25,26]; the resultant C-terminal peptide (termed C31), however, may contribute to neuronal death, as has been postulated [14]. Our finding that Ah (including Ah42) secretion increases during apoptosis connotes functional increase in h-secretase (i.e., BACE-1) processing, which is rate-limiting for the generation of Ah (see Ref. [28]). We demonstrate, however, that increased Ah secretion during apoptosis is not accompanied by up-regulation of BACE-1 expression at the transcriptional level (corresponding mRNA levels actually decrease uniformly, i.e., irrespective of APP isoform expressed, during apoptosis). One possibility, which our data do not address, is increased sorting of APP holoprotein, during apoptosis, to compartments (e.g., endocytotic) where BACE-1 resides, or, conversely, increased sorting of BACE1 to compartments where APP holoprotein resides. A second (non-excluding) possibility, which again, our data do not address, is increased half-life for BACE-1 during apoptosis, analogous to our finding for APP holoprotein. Regulated sorting of BACE-1 to endocytotic compartments has been described [10,31], and of note, co-segregation of cell-surface APP and BACE-1 in cholesterol-rich, lipid ‘rafts’ recently has been shown to contribute significantly to Ah secretion in an endocytosis-dependent manner [5]. In this view, our finding that significantly decreased baseline Ah (including Ah42) secretion occurs with expression of endocytosis-deficient APP may originate from lack of cosegregation of (internalized) APP and BACE-1 in endocytotic compartments. However, Ah40 secretion increases significantly during apoptosis with expression of endocytosis-deficient APP, to an extent greater than is evident for endocytosis-competent holoprotein. The simplest explanation for these observations is that g-secretase processing of BACE-1 processed APP occurs preferentially, during apoptosis, at residue 42 of the corresponding Ah sequence in endocytotic compartments, by unknown mechanism, and that endocytosis-deficient APP undergoes comparatively increased (relative to endocytosis-competent holoprotein) amyloidogenic processing in secretory compartments during apoptosis, with resultant increased Ah40 secretion. Alternatively, more Ah42 could be generated in secretory compartments (i.e., ER/IC) during apoptosis with expression of endocytosis-deficient APP, but otherwise retained in these compartments. Relatedly, induction of apoptosis with expression of endocytosis-competent APP could result in increased secretion of previously sequestered Ah42. Another possibility that cannot be excluded, but remains unlikely based on our findings (e.g., increased a-secretase processing of endocytosis-deficient APP, which generally is held to occur in very large part at the cell surface; see Ref. [20]) and the findings of Perez et al. [17], is missorting of endocytosis-deficient APP resulting from altered interaction with

209

binding proteins (e.g., Fe65, X11; see Ref. [4]) that recognize site in the cytoplasmic domain inclusive for the YENP sequence. The precise mechanism whereby amyloidogenic processing in endocytotic compartments results in Ah secretion is not known (in particular, events distal to h-secretase processing of APP holoprotein), but the following scenario is plausible. APP holoprotein is internalized from the cell surface (possibly from lipid ‘rafts’; see Ref. [5]) and undergoes h-secretase processing in early endocytotic compartments; to wit, BACE-1, which also may be internalized from lipid ‘rafts’ (see Ref. [5]), is sorted to APP-containing early endosomal compartments. Subsequent g-secretase processing of h-site processed APP occurs at or near the cell surface, and resultant Ah is released into the extracellular space. Additional sorting applies with respect to gsecretase processing and resultant Ah secretion as follows. Vesicle derived from early endosomal compartment containing h-site processed APP can sort (fuse) to plasmalemma and thus deliver substrate to (plasmalemmal) g-secretase complex. Alternatively (non-exclusively), functional g-secretase complex can internalize from the cell surface (? via lipid ‘rafts’) and sort to early endocytotic compartments containing h-site processed APP; subsequent sorting (vesicle) from early endosomal compartment would deliver resultant Ah to plasmalemma for release into the extracellular space. Evidence exists, as noted above, that partitioning of APP and BACE-1 to lipid ‘rafts’ contributes to amyloidogenic processing and Ah secretion in an endocytosis-dependent manner [5]. Evidence also exists that functional g-secretase complex (obligate components thereof, namely, presenilin-1 and nicastrin) is present at or near the cell surface [1,11]. Lastly, the following scenario may relate our findings to the pathogenesis of Alzheimer’s disease. Vulnerable subpopulations of neurons in the aging brain (e.g., layer 2 stellate neurons in entorhinal cortex, cholinergic neurons in basal forebrain) may respond to cytotoxic stimuli with increased internalization of cell-surface APP from presynaptic sites [15,34], which undergoes amyloidogenic processing in endocytotic compartments, resulting in increased Ah (especially Ah42) secretion. Increased Ah42 burden functions as a neurotoxic and possibly pro-apoptotic signal that contributes to local synaptic dysfunction, which in turn results in increased APP internalization, amyloidogenic processing in endocytotic compartments, and feedback amplification of Ah42 secretion [7]. Continued APP internalization (loss) from presynaptic sites also may contribute to synaptic dysfunction. In summary, we have carried out comparative analysis of Ah (including Ah42) secretion during staurosporine-induced apoptosis in naive (undifferentiated) PC12 cells transfected in parallel for endocytosis-competent or -deficient APP. We report that APP processing by the endocytotic pathway is required for, taken collectively, the generation and secretion of Ah42, and that secretion of this

210

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211

neurotoxic peptide increases significantly during apoptosis. Extension of these studies to include cell types other than naive PC12 cells (e.g., PC12 cells with neuritic differentiation, central nervous system neurons in primary culture) and means other than staurosporine treatment for induction of apoptosis (e.g., genotoxic agents, NGF-withdrawal in PC12 cells with neuritic differentiation) remains for future communication.

Acknowledgements This work was supported by National Institutes of Health grant AG15066 (NRG-L).

References [1] O. Berezovska, P. Ramdya, J. Skoch, M.S. Wolfe, B.J. Bacskai, B.T. Hyman, Amyloid precursor protein associates with a nicastrin-dependent docking site on the presenilin 1-g-secretase complex in cells demonstrated by fluorescence lifetime imaging, J. Neurosci. 23 (2003) 4560 – 4566. [2] D.G. Cook, M.S. Forman, J.C. Sung, S. Leight, D.L. Kolson, T. Iwatsubo, V.M.-Y. Lee, R.W. Doms, Alzheimer’s Ah(1-42) is generated in the endoplasmic reticulum/intermediate compartment of NT2N cells, Nat. Med. 3 (1997) 1021 – 1023. [3] P. Cupers, M. Bentahir, K. Craessaerts, I. Orlans, H. Vanderstichele, P. Saftig, B. De Strooper, W. Annaert, The discrepancy between presenilin subcellular localization and g-secretase processing of amyloid precursor protein, J. Cell Biol. 154 (2001) 731 – 740. [4] B. De Strooper, W. Annaert, Proteolytic processing and cell biological functions of the amyloid precursor protein, J. Cell Sci. 113 (2000) 1857 – 1870. [5] R. Ehehalt, P. Keller, C. Haass, C. Thiele, K. Simons, Amyloidogenic processing of the Alzheimer h-amyloid precursor protein depends on lipid rafts, J. Cell Biol. 160 2003, pp. 113 – 123. [6] C. Galli, A. Piccini, M.T. Ciotti, L. Castellani, P. Calissano, D. Zaccheo, M. Tabaton, Increased amyloidogenic secretion in cerebellar granule cells undergoing apoptosis, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 1247-1252. [7] F.G. Gervais, D. Xu, G.S. Robertson, J.P. Vaillancourt, Y. Zhu, J. Huang, A. LeBlanc, D. Smith, M. Rigby, M.S. Shearman, E.E. Clarke, H. Zheng, L.H. Van Der Ploeg, S.C. Ruffolo, N.A. Thornberry, S. Xanthoudakis, R.J. Zamboni, S. Roy, D.W. Nicholson, Involvement of caspases in proteolytic cleavage of Alzheimer’s amyloid-h precursor protein and amyloidogenic Ah peptide formation, Cell 97 (1999) 395 – 406. [8] C. Geula, C.K. Wu, D. Saroff, A. Lorenzo, M. Yuan, B.A. Yankner, Aging renders the brain vulnerable to amyloid h-protein neurotoxicity, Nat. Med. 4 (1998) 827 – 831. [9] T. Hartmann, S.C. Bieger, B. Bruhl, P.J. Tienari, N. Ida, D. Allsop, G.W. Roberts, C.L. Masters, C.G. Dotti, K. Unsicker, K. Beyreuther, Distinct sites of intracellular production for Alzheimer’s disease Ah40/42 amyloid peptides, Nat. Med. 3 (1997) 1016 – 1020. [10] J.T. Huse, D.S. Pijak, G.J. Leslie, V.M. Lee, R.W. Doms, Maturation and endosomal targeting of h-site amyloid precursor protein-cleaving enzyme, the Alzheimer’s disease h-secretase, J. Biol. Chem. 275 (2000) 33729 – 33737. [11] C. Kaether, S. Lammich, D. Edbauer, M. Ertl, J. Rietdorf, A. Capell, H. Steiner, C. Haass, Presenilin-1 affects trafficking and processing of hAPP and is targeted in a complex with nicastrin to the plasma membrane, J. Cell Biol. 158 (2002) 551 – 561.

[12] T.-W. Kim, W.H. Pettingell, Y.-K. Jung, D.M. Kovacs, R.E. Tanzi, Alternative cleavage of Alzheimer-associated presenilins during apoptosis by a caspase-3 family protease, Science 277 (1997) 373 – 376. [13] A. LeBlanc, Increased production of 4 kDa amyloid h peptide in serum deprived human primary neuron cultures: possible involvement of apoptosis, J. Neurosci. 15 (1995) 7837 – 7846. [14] D.C. Lu, S. Rabizadeh, S. Chandra, R.F. Shayya, L.M. Ellerby, X. Ye, G.S. Salvesen, E.H. Koo, D.E. Bredesen, A second cytotoxic proteolytic peptide derived from amyloid h-protein precursor, Nat. Med. 6 (2000) 397 – 404. [15] N.R. Marquez-Sterling, A.C.Y. Lo, S.S. Sisodia, E.H. Koo, Trafficking of cell-surface h-amyloid precursor protein: evidence that a sorting intermediate participates in synaptic vesicle recycling, J. Neurosci. 17 (1997) 140 – 151. [16] J. Murrell, M. Farlow, B. Ghetti, M.D. Benson, A mutation in the amyloid precursor protein associated with hereditary Alzheimer’s disease, Science 254 (1991) 97 – 99. [17] R.G. Perez, S. Soriano, J.D. Hayes, B. Ostaszewski, W. Xia, D.J. Selkoe, X. Chen, G.B. Stokin, E.H. Koo, Mutagenesis identifies new signals for h-amyloid precursor protein endocytosis, turnover, and the generation of secreted fragments, including Ah42, J. Biol. Chem. 274 (1999) 18851 – 18856. [18] P. Ponte, P. Gonzalez-DeWhitt, J. Schilling, J. Miller, D. Hsu, B. Greenberg, K. Davis, W. Wallace, I. Lieberburg, F. Fuller, B. Cordell, A new A4 amyloid mRNA contains a domain homologous to serine proteinase inhibitors, Nature 331 (1988) 525 – 527. [19] H. Schagger, G. von Jagow, Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa, Anal. Biochem. 166 (1987) 368 – 379. [20] D.J. Selkoe, Alzheimer’s disease: genes, proteins, and therapy, Physiol. Rev. 81 (2001) 741 – 766. [21] S.S. Sisodia, E.H. Koo, P.N. Hoffman, G. Perry, D.L. Price, Identification and transport of full-length amyloid precursor proteins in rat peripheral nervous system, J. Neurosci. 13 (1993) 3136 – 3142. [22] R.S. Slack, F.D. Miller, Viral vectors for modulating gene expression in neurons, Curr. Opin. Neurobiol. 6 (1996) 576 – 583. [23] H.H. Slunt, G. Thinakaran, C.V. Koch, A.C.Y. Lo, R.E. Tanzi, S.S. Sisodia, Expression of a ubiquitous, cross-reactive homologue of the mouse h-amyloid precursor protein (APP), J. Biol. Chem. 269 (1994) 2637 – 2644. [24] S. Soriano, A.S. Chyung, X. Chen, G.B. Stokin, V.M. Lee, E.H. Koo, Expression of h-amyloid precursor protein-CD3g chimeras to demonstrate the selective generation of amyloid h1-40 and amyloid h1-42 peptides within secretory and endocytic compartments, J. Biol. Chem. 274 (1999) 32295 – 32300. [25] S. Soriano, D.C. Lu, S. Chandra, C.U. Pietrzik, E.H. Koo, The amyloidogenic pathway of amyloid precursor protein (APP) is independent of its cleavage by caspases, J. Biol. Chem. 276 (2001) 29045 – 29050. [26] G. Tesco, Y.H. Koh, R.E. Tanzi, Caspase activation increases hamyloid generation independently of caspase cleavage of the hamyloid precursor protein (APP), J. Biol. Chem. 278 (2003) 46074 – 46080. [27] M. van Engeland, F.C. Ramaekers, B. Schutte, C.P. Reutelingsperger, A novel assay to measure loss of plasma membrane asymmetry during apoptosis of adherent cells in culture, Cytometry 24 (1996) 131 – 139. [28] R. Vassar, M. Citron, Ah-generating enzymes: recent advances in hand g-secretase research, Neuron 27 (2000) 419 – 422. [29] R. Vassar, B.D. Bennett, S. Babu-Khan, S. Khan, E.A. Mendiaz, P. Denis, D.B. Teplow, S. Ross, P. Amarante, R. Loeloff, Y. Luo, S. Fisher, J. Fuller, S. Edenson, J. Lile, M.A. Jaronsinski, A.L. Biere, E. Curran, T. Burgess, J.-C. Louis, F. Collins, J. Treanor, G. Rogers, M. Citron, h-Secretase cleavage of Alzheimer’s amyloid precursor protein by the transmembrane aspartic protease BACE, Science 286 (1999) 735 – 741.

C.P. Sodhi et al. / Molecular Brain Research 128 (2004) 201–211 [30] D.M. Walsh, I. Klyubin, J.V. Fadeeva, W.K. Cullen, R. Anwyl, M.S. Wolfe, M.J. Rowan, D.J. Selkoe, Naturally secreted oligomers of amyloid h protein potently inhibit hippocampal long-term potentiation in vivo, Nature 416 (2002) 535 – 539. [31] J. Walter, R. Fluhrer, B. Hartung, M. Willem, C. Kaether, A. Capell, S. Lammich, G. Multhaup, C. Haass, Phosphorylation regulates intracellular trafficking of h-secretase, J. Biol. Chem. 276 (2001) 14634 – 14641. [32] C. Wild-Bode, T. Yamazaki, A. Capell, U. Leimer, H. Steiner, Y. Ihara, C. Haass, Intracellular generation and accumulation of amyloid h-peptide terminating at amino acid 42, J. Biol. Chem. 272 (1997) 16085 – 16088.

211

[33] J. Wiltfang, A. Smirnov, B. Schnierstein, G. Kelemen, U. Matthies, H. Klafki, M. Staufenbiel, G. Huther, E. Ruther, J. Kornhuber, Improved electrophoretic separation and immunoblotting of beta-amyloid (Ah) peptides 1 – 40, 1 – 42, and 1 – 43, Electrophoresis 18 (1997) 527 – 532. [34] T. Yamazaki, D.J. Selkoe, E.H. Koo, Trafficking of cell surface hamyloid precursor protein: retrograde and transcytotic transport in cultured neurons, J. Cell Biol. 129 (1995) 431 – 442. [35] H. Yuan, P. Zhai, L.M. Anderson, J. Pan, B. Thimmapaya, E.H. Koo, N.R. Marquez-Sterling, Recombinant adenovirus is an appropriate vector for endocytotic protein trafficking studies in cultured neurons, J. Neurosci. Methods 88 (1999) 45 – 54.