Aβ promotes Alzheimer’s disease-like cytoskeleton abnormalities with consequences to APP processing in neurons


Address correspondence and reprint requests to Odete A. B. da Cruz e Silva, Centro de Biologia Celular, Secção Autónoma de Ciências da Saúde, Universidade de Aveiro, 3810-193 Aveiro, Portugal. E-mail: odetecs@ua.pt


J. Neurochem. (2010) 113, 761–771.


Aβ is proteolytically produced from the Alzheimer’s amyloid precursor protein (APP). Major properties attributed to Aβ include neurotoxic effects that contribute to Alzheimer’s disease neurodegeneration. However, Aβ can also affect APP processing and trafficking that, in neurons, is anterogradelly transported via microtubules in a kinesin-associated manner. Herein we show that Aβ can induce accumulation of intracellular sAPP in primary neuronal cultures. Subcellular fractionation studies and immunofluorescence analysis revealed that upon Aβ exposure sAPP retention was localized to cytoskeleton associated vesicular structures along the neurite processes, positive for an APP N-terminal antibody and negative for an APP C-terminal antibody. These vesicular structures were also positive for kinesin light chain 1 (KLC). We confirm that Aβ alters both actin and microtubule networks. It increases F-actin polymerization and we report for the first time that Aβ decreases α-tubulin acetylation. The use of cytoskeleton associated drugs partially reversed the Aβ-induced effects on sAPP secretion. The data here presented show that Aβ causes intracellular sAPP retention by inducing alterations in the cytoskeleton network, thus contributing to impaired APP/sAPP vesicular transport. Moreover, the data strengthens the hypothesis that Aβ-induces neurodegeneration and provides a potential mechanism of action, as impaired vesicular and axonal transport have been linked to Alzheimer’s disease pathology.

Abbreviations used:

amyloid β peptide


Alzheimer’s disease


amyloid precursor protein


bovine serum albumin


C-terminal fragment


extracellular sAPP


holo APP


intracellular sAPP


kinesin light chain 1


polyacrylamide gel electrophoresis


phosphate-buffered saline


sodium dodecyl sulfate

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by the presence of amyloid plaques mainly composed of Aβ peptide (Glenner and Wong 1984), neurofibrillary tangles (Goedert et al. 1992), and synaptic and neuronal loss in distinct brain areas, including the neocortex and hippocampus (Hyman et al. 1984; Masliah et al. 1991; Terry et al. 1991). These changes are accompanied by severe disruption of both axonal and dendritic cytoskeleton leading to failure in neuritic transport, and consequently impaired traffic of neurotransmitters and neuropeptides essential for neuronal viability and function. Axonal abnormalities precede amyloid deposition in some AD mouse models, suggesting that axonal defects play a crucial role in the earliest stages of AD pathogenesis (Stokin et al. 2005). AD axonal pathology includes atypical axonal accumulations of the amyloid precursor protein (APP) (Cras et al. 1991) and its fragments (Sennvik et al. 2004; Takahashi et al. 2004), suggesting altered APP transport and processing.

Amyloid precursor protein anterograde axonal transport is KLC-driven and most probably mediated by JIP-1 (Inomata et al. 2003; Matsuda et al. 2003; Lazarov et al. 2005), which also interacts with the microtubule associated phospho-Tau (Ittner et al. 2009). Kinesin 1 reduction induces axonal abnormalities and enhances aberrant Aβ generation and amyloid deposition (Gunawardena and Goldstein 2001). Aβ derives from the amyloidogenic processing of APP (Kang et al. 1987; Gandy et al. 1994), as it is sequentially cleaved by β-secretase (BACE-1) (Vassar et al. 1999; Yan et al. 2001) and the γ-secretase complex (comprising presenilin, niscatrin, APH1 and PEN2) (Li et al. 2000; Esler et al. 2002; Lee et al. 2002; Steiner et al. 2002). Generation of Aβ is accompanied by release of extracellular sAPPβ and intracellular APP C-terminal domain. Alternatively, APP can be processed via a non-amyloidogenic pathway by α- and γ-secretases (Buxbaum et al. 1998; Lammich et al. 1999; Allinson et al. 2003), hence precluding Aβ formation and generating the p3 peptide and sAPPα. Contrary to Aβ, sAPPα has neurotrophic and neuroprotective properties when added to cultured cells (Turner et al. 2003; Thornton et al. 2006).

Amyloid precursor protein proteolytic processing occurs in specific subcellular compartments and trafficking routes. Cleavage by α-secretase takes place preferentially in the secretory anterograde pathway, while proteolysis by β-secretase is favored in the retrograde endocytic pathway. Thus, alterations in the intracellular transport of APP likely affect it fate and the balance between alternative processing pathways, provoking differences in the proteolytic fragments produced. As APP anterograde transport has been associated with kinesin-1, and thus dependent on microtubule tracks, APP vesicular trafficking is susceptible to cytoskeleton network alterations. Interestingly, Aβ was reported to induce cytoskeleton reorganization and morphological alterations in astrocytes (Salinero et al. 1997). In addition, previous studies from our laboratory in a non-neuronal cell line demonstrated that Aβ affects the release of the neuroprotective sAPPα fragment leading to intracellular sAPP (isAPP) retention in cytoskeleton-associated vesicular-like densities (Henriques et al. 2009b). Consequently, high density clusters are clearly visible in the cytoplasm. We also observed that isAPP retention occurs in non-neuronal, neuronal-like cells and primary cultures, although sAPP secretion was affected to different degrees. Hence, and as APP vesicular trafficking has been associated with kinesin-driven transport along the microtubule network, we characterized Aβ-induced effects on neuronal APP/sAPP vesicular trafficking and relate these to altered cytoskeleton network dynamics.

Materials and methods

Cell culture

Primary rat cortical and hippocampal neuronal cultures were established from 18 days rat embryos as previously described (Henriques et al. 2007). Following dissociation with trypsin and deoxyribonuclease I (0.15 mg/mL) in Hank’s balanced salt solution (0.45 mg/mL for cortical cultures or 0.75 mg/mL for hippocampal cultures during 5–10 min at 37°C), cells were plated onto poly-d-lysine coated dishes at a density of 1.0 × 105 cells/cm2 in B27-supplemented Neurobasal medium (Gibco, Invitrogen, Portugal), a serum-free medium combination (Brewer et al. 1993). The medium was further supplemented with glutamine (0.5 mM), gentamicin (60 μg/mL), and glutamate (25 μM, for hippocampal cultures only). Cultures were maintained in an atmosphere of 5% CO2 at 37°C for 9 days, before being used for experimental purposes.

Neuronal-like PC12 cells (a rat pheochromocytoma cell line, ATCC) were grown in RPMI 1640 medium (Gibco, BRL) supplemented with 0.85 g/L sodium bicarbonate, 10% horse serum and 5% FBS, and maintained at 37°C and 5% CO2. For experimental procedures, cells were plated at a density of 5.0 × 105 cells/cm2 in poly-l-ornithine coated dishes.

Exposure to Aβ

Both primary neuronal cultures and PC12 cells were incubated with different Aβ species (Sigma, Portugal) at the specified concentrations during 24 h in the appropriate cell line medium, which was replaced in the last 3 h by serum free medium with or without Aβ peptide. This was the conditioned medium collected to monitor extracellular sAPP (esAPP) production. For primary neuronal cultures, experiments were carried out in B27-free Neurobasal medium. The Aβ species used was Aβ25–35 but results were confirmed with Aβ1–42, and this is indicated throughout.

Sample collection and immunodetection

Following Aβ treatment, conditioned media and cells were collected as previously described (Amador et al. 2004). Cells were harvested into boiling 1% sodium dodecyl sulfate (SDS), sonicated and boiled for 10 min. Protein determination of the cellular lysates was carried out using the BCA kit (Pierce, Dagma, Portugal). Samples normalized for protein content were separated on 7.5% SDS–polyacrylamide gel electrophoresis (PAGE) and then electrophoretically transferred onto a nitrocellulose membrane. Immunoblotting detection of APP and sAPP was carried out using antibodies that permit distinguishing between holo APP (hAPP)/APP C-terminal fragments (CTFs) and sAPP. The antibodies used were an APP N-terminal antibody (22C11, Boehringer) and an APP C-terminal antibody (rabbit anti-β-APP, Zymed Laboratories Inc., Alfredo Cavalheiro, Portugal).

For the subcellular fraction studies, specific organelle markers used included HSP70/72 (cytosolic marker, Stressgen, Citomed, Portugal) and syntaxin 6 (Golgi marker, BD Biosciences Enzifarma, Portugal). Actin (Stressgen) and β-tubulin (Zymed) confirmed subcellular enrichment in cytosolic- and cytoskeleton-associated fractions. Primary antibody detection made use of horseradish peroxidase-conjugated secondary antibodies (Amersham Pharmacia Biotech, VWR, Portugal) for enhanced chemiluminescence detection (ECL, Amersham Pharmacia). Enhanced chemiluminescence plus was used to detect extracellular sAPP, syntaxin, carboxipeptidase E and intracellular hAPP.

Conditioned media were also subjected to silver stain, to detect proteins secreted. Briefly, the SDS–PAGE gel, processed as described above, was fixed in a solution containing 50% methanol and 5% acetic acid for 30 min. The fixing solution was replaced with 50% methanol for 15 min, after which sensitizing solution [0.02% (w/v) sodium thiosulphate] was added for 1 min. Subsequently, a staining solution of 0.2% (w/v) of silver nitrate chilled to 4°C was added to the gel and allowed to incubate for 25 min. Finally, the gel was incubated with developing solution of 3% sodium carbonate and 0.025% (v/v) of 37% formaldehyde and developed for a maximum of 10 min, using a 1.4% (w/v) sodium EDTA solution to stop color development.

Quantification and statistical analysis

Quantity One densitometry software (Bio-Rad Laboratories, Portugal) was used to quantify band intensity and correlate it to protein levels. Data are expressed as mean ± SEM, from at least three independent experiments. Statistical analysis was carried out using one way analysis of variance (anova). When the F values were significant, the Dunnett test was applied to compare all groups versus control. The level of significance accepted was p < 0.05.

Subcellular fractionation

Primary neuronal cultures and PC12 cells were exposed to Aβ as previously described and subcellular fractions were prepared using the ProteoExtract Subcellular Proteome Extraction Kit (Calbiochem, VWR, Portugal). The differential solubility of the subcellular compartments in specific reagent mixtures enables the differential extraction of proteins according to their subcellular localization. Sequential extraction steps yielded fractions containing cytosolic proteins (‘Cytosol’), plasma membrane and organelle proteins (‘Memb + Org’), and cytoskeleton and cytoskeleton-associated proteins (‘Cytosk’). Fractions obtained were separated on a 5–20% gradient SDS–PAGE gel and immunoblotted for specific proteins, as indicated.

APP and sAPP subcellular localization

For immunofluorescence analysis cells were plated onto coverslips at a confluency of approximately 50%. Following exposure to Aβ for 24 h, cells were fixed in 4% paraformaldehyde, permeabilized with methanol and blocked with 3% bovine serum albumin (BSA). Subsequently, cells were immunolabeled with specific antibodies. To distinguish between APP and sAPP, or other APP cleavage fragments, primary cultures were incubated with an antibody against APP N-terminal and an anti-APP C-terminus antibody as indicated above. The kinesin light chain [KLC 1 (V-17), Santa Cruz Biotechnology, Frilabo, Portugal] antibody was used as a marker for the APP vesicular anterograde transport.

Primary antibody complexes were visualized using Texas Red- (Molecular Probes, Enzifarma, Portugal) and Fluorescein-conjugated (Calbiochem) secondary antibodies. Coverslips were mounted on microscope glass slides using antifading reagents containing, as indicated, DAPI for nucleic acids staining (Vectashield, Baptista Marques, Portugal). Epifluorescence images were acquired using a Zeiss LSM 510-Meta confocal microscope, and a 63×/1.4 oil immersion lens. Argon laser lines of 405 nm and 488 nm were used to excite DAPI and Fluorescein, respectively, and a 561 nm DPSS laser was used to excite Texas Red. Microphotographs were acquired in a sole section in the z-axis (xy-mode), and represent a mean of 16 scans.

Labeling F-actin and acetylated α-tubulin

Cells fixed with paraformaldehyde were permeabilized with a solution of acetone at ≤−20°C for 3 min. Subsequently, cells were washed with phosphate-buffered saline (PBS) and then incubated with PBS containing 1% BSA for 1 h. This blocking solution was removed and phallotoxin staining solution (labels F-actin) was added to cells (1.5 U/100 μL in PBS containing 1% BSA) for 30 min at 25°C. After PBS washing, cells were incubated with an anti-acetylated α-tubulin antibody (Zymed) for 2 h and primary complexes labeled using Fluorescein conjugated secondary antibody (Calbiochem). Coverslips were mounted and visualized as previously described.

Biochemical assays on cytoskeleton dynamics

Cells were incubated with Aβ for 21 h, followed by a 3-h incubation period with fresh medium with or without Aβ in the presence or absence of 10–20 μM cytochalasin D (Sigma-Aldrich), or 20–40 μM taxol (Sigma-Aldrich). Cytochalasin D (cytD) is a compound that drives actin depolymerization and taxol drives microtubule stabilization.


Aβ-induced intracellular sAPP accumulation is associated with the cytoskeleton in primary neuronal cultures

Primary neuronal cultures and PC12 cells were incubated with Aβ peptides for 24 h, and in the last 3 h of incubation the medium was replaced by fresh medium with (Aβ) or without Aβ (Aβ-Aβ). To distinguish between the cleaved APP (sAPP) fragment and the full-length hAPP protein, two antibodies were used. An anti-APP N-terminus 22C11 antibody, which recognizes both hAPP and sAPP, and an anti-APP C-terminal antibody, that recognizes hAPP but not sAPP. We observed that in both primary cortical and hippocampal neuronal cultures, Aβ induced an increase in APP intracellular levels (Fig. 1a) of 2.0 ± 0.18 and 1.6 ± 0.13-fold, respectively. Conversely, Aβ decreased hAPP levels (C-terminal antibody) to 0.6 ± 0.01 for cortical cultures and 0.7 ± 0.05 for hippocampal cultures. In agreement with these observations, a decrease in APP mRNA was observed for neuronal cultures (Henriques et al. 2009a). The data indicate that the intracellular increase observed is because of isAPP accumulation. Accompanying the isAPP retention, esAPP secretion markedly decreased in both neuronal cultures, to 0.2 ± 0.04 (cortical) and 0.1 ± 0.04 (hippocampal) when compared to basal levels. Further, if Aβ was removed in the last 3 h the secretory block was released (Fig. 1a), leading to an increase in extracellular sAPP, reaching control levels in both hippocampal (1.1 ± 0.09) and cortical (1.3 ± 0.14) neurons. This confirms that isAPP retention/sAPP secretory block in primary neuronal cultures are indeed Aβ-specific inhibitory effects. Similar results were obtained when using the Aβ1–42 peptide (Fig. 1b). Two concentrations, 5 μM and 20 μM, were tested and showed augmenting effects in terms of isAPP retention. Furthermore, the control peptide (Scramble peptide, Scrb) did not have any detectable effect on sAPP intracellular retention/secretion. As the Aβ25–35 peptide shows similar effects to the naturally occurring Aβ1–42 peptide and does not require to be previously aggregated, the former was used in the subsequent experiments. Similar data were obtained for PC12 cells, both in terms of intracellular APP accumulation and sAPP secretion (Fig. 1c). The Aβ inhibitory effect was not generalized given that the levels of many secreted proteins remained unchanged after Aβ treatment, as revealed by silver staining proteins present in the conditioned medium (Fig. 1d).

Figure 1.

 Aβ induces intracellular sAPP retention in primary neuronal cultures and PC12 cells. Immunoblot analysis of cellular lysates and conditioned medium was performed with the 22C11 and the C-terminal antibodies to monitor hAPP and sAPP (a–c). Primary neuronal cultures and PC12 cells were incubated with Aβ has described in Material and Methods. Aβ and Scramble peptides (Scrb) were aggregated during 48 h at 37°C and added to primary cultures for 24 h. The Aβ species used in (b) was 1–42; in the remaining panels the Aβ species used was 25–35. (d) The total protein content in primary neuronal cultures conditioned medium was analyzed by silver staining. Full arrows: bands with increased intensity upon Aβ treatment. Dashed arrows: bands with decreased intensity upon Aβ treatment. Unfilled arrowheads: band intensity unchanged. ◂ isAPP, intracellular sAPP; esAPP, extracellular sAPP; C, control cells; Aβ, cells exposed to Aβ for 24 h; Aβ-Aβ, cells exposed to Aβ for 21 h, and further incubated in Aβ-free medium for 3 h.

In order to determine the intracellular sites of isAPP retention induced by Abeta, we performed both subcellular fractionation analysis and immunofluorescence studies. An association of isAPP retention with cytoskeleton structures was observed, for both PC12 cells and primary cortical neurons (Fig. 2a and b) when treated with Abeta. The cytoskeleton/microtubule containing fraction was actin and β-tubulin-positive and free of other organelle markers (Fig. 2), while the cytosolic fractions were HSP70 positive. Immunoblot analysis of the subcellular fractions with 22C11 or C-terminal antibodies allowed us to distinguish hAPP (unfilled arrowheads, ◃) from sAPP (solid arrowheads, ◂). For PC12 cells, Aβ exposure led to isAPP (solid arrowheads, ◂) accumulation in the cytosolic fraction, but more markedly so in the cytoskeleton enriched fraction. This was accompanied by an increase in mature hAPP751/770 C-terminal positive isoforms in the cytoskeleton fraction (arrowhead ◃, Fig. 2a). For primary neuronal cortical cultures similar data were obtained. isAPP accumulated mainly in the cytoskeleton fraction (solid arrowheads, ◂), and hAPP isoforms (unfilled arrowheads, ◃) slightly increased in this fraction, but decreased in the membrane and organelle enriched fraction (Memb + Org). Further, carboxypeptidase E, a protein involved in regulated vesicular secretion, was also found to increase in the cytoskeletal fraction upon Aβ treatment, although its total intracellular level decreased. The isAPP accumulation in cytoskeleton and cytosolic fractions, which are fractions known to be associated with cytoplasmic vesicles, together with the decrease in hAPP in Memb+Org fraction, suggested that hAPP was being cleaved intracellularly before reaching the plasma membrane. Accordingly, previous data from our group also showed intracellular CTFs accumulated in response to Aβ for primary neurons, reinforcing that hAPP is intracellularly cleaved (Henriques et al. 2009a). These results strengthen the hypothesis that Aβ exerts an inhibitory effect at the vesicular secretory level, leading to intravesicular isAPP production and retention.

Figure 2.

 Intracellular cytoskeletal sAPP retention. Following incubation with Aβ25–35, PC12 cells (a) and primary cortical cultures (b) were fractionated as described, and the resulting fractions analyzed by immunoblotting. Total, total cell lysate; Cytosol, cytosolic protein fraction; Memb + Org, membrane and organelles enriched fraction; and Cytosk, cytoskeleton protein fraction. ◂, isAPP forms; ◃, mature hAPP; esAPP, extracellular sAPP. Intracellular sAPP and APP levels (N-terminal 22C11 antibody), HSP70 (cytosolic marker), actin and β-tubulin (cytosolic and cytoskeleton marker) were detected using ECL; intracellular hAPP (C-terminal antibody), CPE (carboxypeptidase E, marker of regulated vesicular secretion), and syntaxin-6 (Golgi marker) were detected using the highly sensitive ECL plus reagent. C, control; Aβ, Aβ exposure.

isAPP retention occurs in secretory vesicles

To characterize the location of intracellular sAPP retention in primary hippocampal neurons we have performed immunofluorescence microscopy analyses. As already mentioned, the vesicular axonal transport of APP involves the microtubule-associated motor protein kinesin 1. The latter is known to be responsible for a wide range of protein vesicular secretion and is composed of two subunits of KHC and two subunits of KLC (heavy and light chains, respectively). Co-localization of KLC with APP/sAPP was therefore addressed under basal and Aβ exposure conditions. These studies were carried out with the 22C11 (N-terminal APP) and C-terminal antibodies to distinguish between hAPP and sAPP. Under basal conditions 64 ± 3% of the 22C11 positive population co-localizes with the C-terminal immunoreactive population (hAPP), while with Aβ treatment only 44 ± 5% of the 22C11 positive population co-localizes with the C-terminal antibody, indicating that the remaining 22C11 positive population is sAPP (36% under basal and 56% in Aβ conditions). This is in agreement with increased isAPP retention upon Aβ addition. We also observed that under basal conditions both 22C11 and C-terminal antibodies co-localize with vesicle-associated KLC in the majority of the neurites observed (Fig. 3a). Thus we can deduce that hAPP and KLC co-localize intracellularly (Fig. 3a, triple co-localization). However, Aβ exposure led to decreased C-terminal immunoreactivity and diminished co-localization with vesicle-associated KLC at the neurites (Fig. 3b, dashed arrows). In contrast, the 22C11 and KLC co-localization signal was maintained or increased in most neurites (Fig. 3b, solid arrows). This is consistent with isAPP accumulating in neuritic KLC-positive vesicles (Fig. 3b, mainly double co-localization).

Figure 3.

 Aβ induces intracellular sAPP retention in secretory vesicles. Following Aβ treatment, primary hippocampal cultures were prepared for immunofluorescence analysis. Both 22C11 (Fluoresceine-labeled, green) and C-terminal (labeled with Alexa 350, blue) antibodies allowed to discriminate between hAPP and the sAPP. KLC antibody was used as a marker of secretory vesicles (Texas Red-labeled, red). (a) Represents hAPP localization under basal conditions. (b) Represents sAPP localization under Aβ treated conditions. Solid arrows indicate KLC +ve/22C11 +ve structures, and dashed arrows indicate KLC +ve/C-terminal −ve structures. Images were acquired using a Zeiss confocal microscope. C, control cells; Aβ, cells exposed to Aβ for 24 h; ROI, region of interest. Bar, 10 μm.

Aβ interferes with cytoskeleton network

The results described above indicate that Aβ affects KLC-driven APP vesicular transport which is known to be associated with the cytoskeleton network and dependent on microtubule tracks. Therefore, alterations in the vesicular secretion may be associated with altered cytoskeletal dynamics. Thus we addressed Aβ effects on the major proteins constituting the cytoskeleton network polymers, actin and tubulin. As the polymerization/depolymerization dynamics of these proteins are key processes in transport, F-actin polymerization (filamentous actin deriving from globular actin polymerization) and α-tubulin acetylation (an indirect measurement of microtubule stability) were evaluated in both primary neuronal cultures and PC12 cells (Fig. 4). For primary neuronal cultures, exposure to Aβ leads to a dramatic decrease in α-tubulin acetylation, suggesting decreased microtubule stability. For PC12 cells a redistribution of α-tubulin acetylation to the periphery could be observed. With respect to F-actin polymerization, Aβ exposure leads to a significant increase, as is evident for PC12 cells (Fig. 4b and S1). Increased polymerization is clearly visible around the plasma membrane and in the cytoplasm (solid arrows). This response is less evident in primary neuronal cultures but can still be observed (Fig. 4a).

Figure 4.

 Aβ leads to alterations in the cytoskeletal networks of neuronal cultures and PC12 cells. Upon Aβ treatment, primary hippocampal cultures (a) and PC12 cells (b) were stained with Alexa Fluor 568 conjugated phallotoxin solution (labeled filamentous F-actin) and with acetylated α-tubulin antibody (labeled with Fluoresceine, green staining). Confocal images were acquired using a Zeiss confocal microscope. C, control cells; Aβ, Aβ for 24 h. Bar, 10 μm.

Aβ block on APP/sAPP trafficking can be reversed by altering cytoskeleton dynamics

The Aβ-induced effect on isAPP retention and its correlation with altered cytoskeleton dynamics was further established biochemically, by using an actin depolymerizing agent (cytochalasin D, cytD) and a microtubule stabilizing drug (taxol). Aβ was added to PC12 cells as previously described, but each of the cytoskeleton altering drugs was also added in the last 3 h. Under these conditions, isAPP levels remained mainly unchanged (data not shown) but the effect of Aβ on esAPP levels was considerably reversed (Fig. 5). In PC12 cells, co-incubation of Aβ and cytD in the last 3 h significantly reversed the effect of Aβ on esAPP (Fig. 5a). Although to a lesser extent than cytD, taxol could also counteract the Aβ effect in this cell line. At 40 μM, taxol considerably attenuated the Aβ inhibition of esAPP release (Fig. 5b). Hence, Aβ inhibitory effect on sAPP secretion could be partially reversed by drugs which affect both actin and microtubule networks. In primary cultures, taxol could not revert the Aβ effects at the conditions tested (data not shown), possibly because it renders microtubules stable but less dynamic thus less functional in transport, what may particularly impair neuritic transport. On the other hand, cytD when added to primary cultures partially reversed the Aβ induced effect on esAPP secretion (Fig. 5c), presumably via cytD-induced depolymerization of the actin network.

Figure 5.

 Aβ effects reversed by drugs acting on cytoskeleton dynamics. Extracellular sAPP secretion levels were evaluated upon co-incubation of Aβ with drugs able to modulate cytoskeleton dynamics. PC12 cells were treated with Aβ and cytochalasin D (a) or taxol (b) during the last 3 h of the 24-h incubation period. (c) Primary cultures treatment with cytochalasin D. ◂; isAPP, intracellular sAPP; esAPP, extracellular sAPP; C, control; Aβ, Aβ treatment during 24 h; CytD, cytochalasin D.


Given its ability to trigger a set of biochemical and cellular alterations, Aβ has been described as a key player in the amyloid cascade hypothesis leading to progressive neurotoxicity and neuronal death observed in AD (Hardy and Higgins 1992; Hardy and Selkoe 2002). However, besides the well described neurotoxic/apoptotic effects, Aβ also provokes alterations in APP metabolism (Davis-Salinas et al. 1995; Schmitt et al. 1997; Carlson et al. 2000). Previous work from our laboratory demonstrated that Aβ exerts an effect on APP trafficking/processing leading to isAPPα accumulation in cytoskeleton-associated vesicular-like structures in a non-neuronal cell line. Noticeably, this isAPP cleavage and retention was likewise observed in primary hippocampal cultures and PC12 cells (Henriques et al. 2009b). We have also detected an increase in APP CTFs in primary neuronal cultures (Henriques et al. 2009a), again suggesting enhanced intracellular hAPP cleavage. Nonetheless, differences could be observed at the level of sAPP secretion between these cell types. An important question regarding the specificity and the origin of this potentially pathological intracellular sAPP retention, as well as the cellular structures mediating this Aβ response in neurons and in PC12 cells arises. In the work here described we show Aβ-induced inhibition of sAPP secretion, suggesting an Abeta blocking effect on vesicular secretion. This effect was considerable in both primary cortical and hippocampal cultures, and was not generalized, as other medium secreted proteins were unaffected. Furthermore, we also observed that isAPP retention was associated with cytoskeleton structures, as previously shown for non-neuronal cells (Henriques et al. 2009b).

It is well established that vesicular motility and exocytosis are intimately associated with the cytoskeleton network (Meyer and Burger 1979; Hamm-Alvarez and Sheetz 1998; Lanzetti 2007; Potokar et al. 2007), and in the light of our data, disruption of this system may explain the observed intracellular sAPP retention and decreased secretion. It appears then, that as a response to Aβ exposure, sAPP can be intracellularly produced and retained in cytoskeleton-associated secretory vesicles. This hypothesis is supported by co-immunocytochemistry studies locating isAPP retention to KLC-positive secretory vesicles in Aβ treated cells. Our data clearly shows that isAPP and KLC localize to the same subcellular structure, in accordance with either a direct or indirect interaction between KLC and APP in neuritic anterograde transport.

As vesicular transport is related to the cytoskeleton network, hindered APP/sAPP vesicular transport induced by Aβ could be a consequence of altered cytoskeleton dynamics. Indeed, we have demonstrated that Aβ affects the dynamics of both actin and tubulin networks, thus providing a mechanism for sAPP retention. Further, our data clearly show that Aβ alters vesicular trafficking by affecting cytoskeleton networks. As observed with Aβ withdrawal, drugs that alter cytoskeleton dynamics were able to partially reverse esAPP secretion, even in a short period (3 h). These data correlate sAPP retention with hindered cystoskeleton dynamics and indicates that Aβ induced-effects can be counteracted. Aβ impacts on actin polymerization, and by reversing the polymerization state with cytD, the Aβ inhibition of esAPP secretion could be reversed, although more efficiently in PC12 cells. In these cells, protein secretion is highly dependent on the actin cytoskeleton, whereas in primary cultures protein/neuritic transport is mainly dependent on the microtubule network. This may explain the different degree to which the cytoskeleton modulating drugs reverse the Aβ effect on esAPP secretion in the different cell lines.

25–35 was previously shown to affect axonal transport by inducing neuronal actin polymerization and aggregation (Hiruma et al. 2003), and Abeta induction of local actin polymerization may impair vesicular exocytosis. Induction of F-actin polymerization by fibrillar Aβ1–42 was reported by Mendoza-Naranjo et al. (2007) in hippocampal neurons and related to increased activity of Rac1/Cdc42 Rho GTPases induced by this peptide. In addition, we have previously shown that Aβ can inhibit PP1 (Vintem et al. 2009), a phosphatase involved in many signaling cascades, including regulation of the actin/microtubule dynamics. Cofilin is an actin-binding protein whose regulation is critical to actin polymerization, and is regulated by protein phosphorylation and Protein Phosphatase 1 (PP1) , which as mentioned can be influenced by Aβ.

Presently, we hypothesize that Aβ effects on the microtubule network also has consequences on protein trafficking. The microtubule stabilizing drug, taxol, was able to attenuate the Aβ inhibition of sAPP release in PC12 cells. Even so, the resulting response in terms of the levels of esAPP and taxol reversal, specific to different cell lines, deserves further investigation. More interestingly, tubulin acetylation was also found compromised upon Aβ exposure, which in turn may affect APP microtubule mediated transport (Reed et al. 2006; Gardiner et al. 2007). It is consensual that α-tubulin acetylation is an indirect measure of the amount of the tubulin polymer and of microtubule stability, not conferring stability in itself (Black et al. 1989; Bloom 2004). Of note, Gardiner et al. (2007) suggested that increased α-tubulin acetylation was associated with enhanced transport along microtubules and Reed et al. (2006) showed that α-tubulin acetylation can influence the binding and the motility of the microtubule motor protein kinesin 1. In addition, work by Dompierre et al. (2007) suggested that hyperacetylation of neuronal tubulin leads to the release of neurotrophic factors-containing vesicles, via recruitment of motor proteins to microtubules, highlighting the importance of tubulin acetylation in vesicular secretion. The degree of tubulin acetylation is cell type specific, increasing with cell specialization (Black and Keyser 1987), and results point to a major role for α-tubulin acetylation in APP/sAPP neuritic transport. In fact, when Aβ is withdrawn in the last 3 h, the levels of α-tubulin acetylation recovered to basal levels (Fig. S2), concomitantly with esAPP secretion. Possible underlying mechanisms include Abeta stimulation of histone deacetylase 6 activity, rendering decreased levels of acetylated tubulin and lower rate of vesicular trafficking. In addition, we cannot exclude that Abeta may also be affecting microtubule polymerization and/or stabilization via Tau hyperphosphorylation (Ekinci and Shea 2000; Town et al. 2002), or other microtubule-associated proteins, such as MAP1b, which cross-talk between actin and microtubule dynamics.

In conclusion, in neurons, Aβ impairs APP/sAPP vesicular anterograde transport and exocytosis, in a mechanism mediated by altered cytoskeleton dynamics of both microtubule and actin networks. Aβ-mediated mechanisms leading to cytoskeleton abnormalities and impaired protein vesicular secretion consequently contribute to AD neurodegeneration. Microtubule destabilization has been reported to be associated with neurotoxicity, and Aβ-induced neurodegeneration could be prevented by microtubule stabilization drugs (Michaelis et al. 1998, 2005; Seyb et al. 2006). An important additional consequence of the Aβ effects here reported is altered APP/sAPP neuritic transport and decreased sAPP secretion. As extracellular sAPP has potential neurotrophic and neuroprotective properties, its depletion is of extreme importance in a background of neuronal damage and loss as in AD.


Supported by IV the European Union VI Framework Program (Project cNEUPRO), Fundação para a Ciência e Tecnologia of the Portuguese Ministry of Science and Technology (REEQ/1023/BIO/2005, POCTI/NSE/33520, PTDC/QUI-BIQ/101317/2008, BD/16071/2004, BPD/44604/2008), and the Center for Cell Biology at University of Aveiro. The authors gratefully acknowledge the contributions of Prof. Sam Sisodia for his critical discussions and suggestions.