Inverse and distinct modulation of tau ... - Wiley Online Library

Journal of Neurochemistry, 2007, 101, 1303–1315

doi:10.1111/j.1471-4159.2006.04435.x

Inverse and distinct modulation of tau-dependent neurodegeneration by presenilin 1 and amyloid-b in cultured cortical neurons: evidence that tau phosphorylation is the limiting factor in amyloid-b-induced cell death Julia Leschik,* Alfred Welzel,* Carina Weissmann,* Anne Eckert and Roland Brandt* *Department of Neurobiology, University of Osnabru¨ck, Osnabru¨ck, Germany Neurobiology Research Laboratory, Psychiatric University Clinic, Basel, Switzerland

Abstract Alzheimer’s disease (AD) is characterized by massive neuron loss in distinct brain regions, extracellular accumulations of the amyloid precursor protein-fragment amyloid-b (Ab) and intracellular tau fibrils containing hyperphosphorylated tau. Experimental evidence suggests a relation between presenilin (PS) mutations, Ab formation, and tau phosphorylation in triggering cell death; however, how Ab and PS affect taudependent degeneration is unknown. Using herpes simplex virus 1-mediated gene-transfer of fluorescent-tagged tau constructs in primary cortical neurons, we demonstrate that tau expression exerts a neurotoxic effect that is increased with a construct mimicking disease-like hyperphosphorylation [pseudohyperphosphorylated (PHP) tau]. Live imaging revealed that PHP tau expression is associated with increased perikarya suggesting the development of a ‘ballooned’ phe-

notype as a specific feature of tau-mediated cell death. Transgenic expression of PS1 suppressed tau-induced neurodegeneration. In contrast, Ab amplified degeneration in the presence of wt tau but not of PHP tau. The data indicate that PS1 and Ab inversely modulate tau-dependent neurodegeneration at distinct steps. They indicate that the mode by which PHP tau causes neurotoxicity is downstream of Ab and that tau phosphorylation is the limiting factor in Ab-induced cell death. Suppression of tau expression or inhibition of tau phosphorylation at disease-relevant sites may provide an effective therapeutic strategy to prevent neurodegeneration in Alzheimer’s disease. Keywords: Alzheimer’s disease, amyloid precursor protein/ Ab, herpes simplex virus, presenilin, tau, transgenic cortical cultures. J. Neurochem. (2007) 101, 1303–1315.

Histopathologically, brains of Alzheimer’s disease (AD) patients are characterized by the presence of extracellular amyloid plaques containing the aggregated amyloid precursor protein (APP) peptide fragment amyloid-b (Ab) and intracellular neurofibrillary tangles composed of hyperphosphorylated tau protein. Familial AD (FAD) cases occurring below the age of 60 years are caused by mutations in both the very homologous presenilin (PS) genes 1 and 2 and the APP. All of the so far characterized FAD-mutations lead to an increase in the production of Ab42 (for reviews see Mattson 2004; Tanzi and Bertram 2005). In AD, tau displays an abnormally high phosphorylation stoichiometry of six to eight phosphate/mol compared with 1.9 phosphate/mol tau isolated from healthy brain (Kennessey and Yen 1993) and exists in straight or paired helical filaments (PHFs), the major components of intracellular

Received June 28, 2006; revised manuscript received November 25, 2006; accepted December 7, 2006. Address correspondence and reprint requests to Prof. Dr Roland Brandt, Department of Neurobiology, University of Osnabru¨ck, Barbarastraße 11, Osnabru¨ck D-49076, Germany. E-mail: [email protected] Abbreviations used: AD, Alzheimer’s disease; APP, amyloid precursor protein; Ab, amyloid-b; DAPI, 4,6-diamidino-2-phenylindole; DAPT, N-[N-(3,5-difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester; DMEM, Dulbecco’s modified Eagle’s medium; eGFP, enhanced green fluorescent protein; FAD, familial Alzheimer’s disease; HSV-1, herpes simplex virus type 1; MAP, microtubule-associated protein; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; NGF, nerve growth factor; PBS, phosphate buffered saline; PDGF, plateletderived growth factor; PHFs, paired helical filaments; PHP, pseudohyperphosphorylated; PS, presenilin; SDS, sodium dodecyl sulfate; wt, wild type.

2007 The Authors Journal Compilation 2007 International Society for Neurochemistry, J. Neurochem. (2007) 101, 1303–1315

1303

1304 J. Leschik et al.

neurofibrillary tangles (Kidd 1963). PHF-tau is hyperphosphorylated at several sites, which may result from a pathologically changed equilibrium of the concerted action of various kinases and phosphatases in the diseased brain (Gong et al. 1993; Jicha et al. 1999; Patrick et al. 1999). Several in vitro studies indicate that Ab can induce an increased phosphorylation of tau at disease-relevant sites (Busciglio et al. 1995; Greenberg et al. 1995; Takashima et al. 1998; Rapoport and Ferreira 2000; Zheng et al. 2002) and can cause tau aggregation into PHF-like filaments (Rank et al. 2002). Furthermore, evidence has been provided that tau is essential for Ab-induced neurotoxicity (Rapoport et al. 2002) and studies using diverse FAD-transgenic mouse models indicate a contribution of mutated tau, the PS, and Ab in cognitive defects and neuronal degeneration (Sturchler-Pierrat et al. 1997; Calhoun et al. 1998; Tomidokoro et al. 2001; Otth et al. 2002; Echeverria et al. 2004). However, the role of AD-like tau hyperphosphorylation in the disease process and the functional interactions between tau hyperphosphorylation, Ab and PS are unsolved. Previously, we have shown that the expression of tau mutants mimicking key structural and functional aspects of hyperphosphorylated tau [pseudohyperphosphorylated tau (PHP tau)] exerts a cytotoxic effect in neural cell lines providing a unique model to determine the role of tau modification in the disease process (Eidenmuller et al. 2000, 2001; Fath et al. 2002). Recently, we could show that PHP tau also induces neuronal degeneration associated with the development of a ‘ballooned’ phenotype as distinct feature of cell death in organotypic hippocampal slices (Shahani et al. 2006). In this study, we addressed the question whether and how other pathologically relevant proteins interfere with taudependent neurodegeneration. Primary cortical cultures from non-transgenic mice, mice transgenic for PS1, and mice transgenic for FAD-mutated APP (APPSDL) that secreted Ab were infected with human wild type (wt) and PHP tau using herpes simplex virus amplicon vectors (HSV-1), which drive a highly efficient, transient expression of exogenous proteins with minimal toxicity in neurons (Neve et al. 2005). Infected neurons were analyzed by live imaging, immunocytochemistry, and immunoblot. The data provide evidence that PS1 and Ab inversely modulate tau-dependent neurodegeneration and that tau hyperphosphorylation is central in this process.

Experimental procedures Materials and antibodies used Chemicals were purchased from Sigma (Deisenhofen, Germany). Cell culture media and supplements were obtained from Sigma and Invitrogen (Gaithersburg, MD, USA), culture flasks, plates, and dishes were from Nunc (Roskilde, Denmark), and primers were from MWG-Biotech AG (Ebersburg, Germany) unless stated otherwise. c-Secretase inhibitors N-[N-(3,5-difluorophenacetyl)-L-

alanyl]-S-phenylglycine t-butyl ester (DAPT) and WPE-III-31C were purchased from Merck (Darmstadt, Germany). The following primary antibodies were used: Mouse monoclonal Tau-5 (PharMingen, San Diego, CA, USA), PHF1 (a generous gift of Peter Davies, Albert Einstein College of Medicine, Bronx, NY, USA), AT8 (Pierce Biotechnology, Rockford, IL, USA), anti-MAP2 (AP20; Chemicon, Temecula, CA, USA), anti-human PS1 (17C2; Assay Designs Inc., Ann Arbor, MI, USA), anti-tubulin (DM1A), and rabbit polyclonal anti-activated caspase 3 (9661; Cell Signaling, Beverly, MA, USA). As secondary antibodies, Cy3-coupled donkey anti-rabbit and donkey anti-mouse antibody (Dianova, Hamburg, Germany) and peroxidase-conjugated anti-mouse and anti-rabbit antibodies (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) were used. Construction of HSV expression vectors Eukaryotic expression plasmids for fetal (352 aa) human tau with Nterminally fused enhanced green fluorescent protein (eGFP) were constructed in pHSV expression vectors. PHP tau was constructed by changing the codons for S198, S199, S202, T231, S235, S396, S404, S409, S413, and S422 to glutamate as described previously (Eidenmuller et al. 2000). As a control, an expression vector was constructed with eGFP alone. Amplification and purification of HSV-1 constructs was carried out as described previously (Fath et al. 2002). Cell culture and HSV infection Primary cortical cultures were prepared from cerebral cortices of mouse embryos (day 15–17 of gestation) transgenic for human PS1 under the control of a modified 3-hydroxy-3-methylglutaryl-coenzyme A reductase promoter (Terro et al. 2002) or mice transgenic for human APP695 with three FAD-mutations [Swedish, S: KM595/ 596NL; Dutch, D: E618Q; London, L: V642I (SDL)] under the control of the platelet-derived growth factor (PDGF)b promoter (Blanchard et al. 2003) according to Terro et al. (2002) with modifications. The cultures were obtained by breeding of heterozygous transgenic males with non-transgenic females C57BL/6 NCrl (Charles River Laboratories, Inc., Wilmington, MA, USA) and C57BL/6 Jico (Harlan Winkelmann GmbH, Borchen, Germany) and were performed for each embryo individually with genotype determined separately from DNA extracted from liver tissue (DNeasy tissue kit; Qiagen GmbH, Hilden, Germany) by PCR using the following primers: human ps1 gene, 5¢-TAATTGGTCCATAAAA GGC-3¢ (forward) and 5¢-GCACAGAAAGGGAGTCACAAG-3¢ (reverse) and human app gene, 5¢-GTAGCAGAGGAGGAAGAAGTG-3¢ (forward) and 5¢-CATGACCTGGGACATTCTC-3¢ (reverse). It cannot be completely excluded that effects not related to the expressed construct influence the behavior of the individual mouse clones. However, we did not obtain any evidence for an unusual phenotype of the mice or of the dissociated neurons indicating their suitability for the described experiments. Cortices from each embryo were dissected in HBSS medium [Ca–Mg-free Hanks balanced salts solution (PAA Laboratories GmbH, Pasching, Austria)] and mechanically dissociated with a Pasteur pipette in Earle’s minimum essential medium with Na-bicarbonate, without L-Gln (PAA Laboratories). The cell suspension was centrifuged for 15 min at 100 g, resuspended in NB/B27 [neurobasal medium containing 2% (v/v) B27 supplement, 2 mmol/L Gln, 1% (v/v) heat-inactivated fetal


Modulation of tau-dependent neurodegeneration 1305

calf serum, 1% (v/v) horse serum, 25 lmol/L b-mercaptoethanol, and 100 lg/mL primocin (InvivoGen, San Diego, CA, USA)] and plated at 5 · 104 cells/cm2 on laminin-coated (0.6 lg/cm2) and poly-Llysine pre-coated coverslips. Human model neurons (NT2-N cells) were produced by in vitro differentiation of NT2 cells for 5 weeks, enriched by serial replatement, plated at 104 cells/cm2 on collagen-coated coverslips, treated with cytostatica for 2 weeks as described previously (Piontek et al. 1999), and further cultured in serum-Dulbecco’s modified Eagle’s medium (DMEM) [DMEM containing 10% (v/v) heatinactivated fetal calf serum, 5% (v/v) heat-inactivated horse serum, 2 mmol/L Gln, 50 U/mL penicillin, and 50 lg/mL streptomycin]. After 5 days in vitro (cortical cultures) or 4 days in serumDMEM (hNT neurons), the culture medium was replaced with fresh medium. For cytochemistry, cells were infected with 2500 (cortical cultures) and 5000 (hNT neurons) amplicons per cm2. For immunoblot analysis, cells were infected with more virus (1– 5 · 104 amplicons per cm2). After 4 h, the medium was exchanged against fresh medium and incubation continued for 2–4 days. PC12 cells stably transfected with FLAG-epitope (DYKDDDDK) tagged wt tau or PHP tau encoding the 352 residues human tau isoform or with vector only were cultured in serum-DMEM supplemented with 250 lg/mL geneticin as described previously (Fath et al. 2002). Cytochemistry Cells were washed with phosphate buffered saline (PBS), fixed with 4% p-formaldehyde, and treated with glycine followed by permeabilization with Triton X-100 as described previously (Brandt et al. 1995) or incubated with PEM-buffer (100 mmol/L PIPES/KOH/pH 6.8, 5 mmol/L EGTA, and 2 mmol/L MgCl2) for 10 min, washed twice with PBS, and incubated for 5–10 min at )20C with precooled ethanol. After washing with blocking buffer [PBS containing 1% (w/v) BSA, and 0.1%(v/v) Tween 20], cells were incubated for 1 h with primary antibodies diluted in blocking buffer, washed with blocking buffer, and incubated with the respective secondary antibody for 30 min. To stain nuclei, 4,6-diamidino-2-phenylindole (DAPI, 2 lg/mL) was added. The coverslips were then washed in blocking buffer and mounted in anti-fade medium [0.1% (w/v) p-phenylendiamine in 90% (v/v) glycerol, and 10% (v/v) PBS buffered to pH 9 with carbonate/bicarbonate]. Cells were imaged with a dry 20·/0.5 Plan Fluor, a dry 40·/0.75 Plan Fluor, and an oilimmersion 40·/1.0 Plan Apo objective on a Nikon Eclipse TE2000U fluorescence microscope (Nikon GmbH, Du¨sseldorf, Germany) equipped with a digital camera (Vossku¨hler COOL-1300). Live imaging All images were acquired on a Nikon laser scanning microscope (Nikon Eclipse TE2000-U inverted), equipped with a C1 confocal laser scanning unit and EZ-C1 software. Neurons expressing eGFP or eGFP-tau were visualized using the 488 nm argon laser line and 510–540 nm band-pass emission filter. Microscope objectives (Nikon) used were a dry 20·/0.45 Plan Fluor, a dry 40·/0.95 Plan Fluor Phase, and an oil-immersion 40·/1.3 Plan Fluor objective. The microscope was enclosed in an incubation chamber maintained at 37C and 5% CO2 (Solent scientific Limited, Sagensworth, UK). Image stacks (1024 · 1024 pixels) were collected at 0.6 lm z-axis steps (typically between 18–22 optical sections per cell). Collapsed images were generated from the resulting Z stacks. In parallel,

phase-contrast images were obtained at phase-contrast settings using an attached digital camera (Nikon Coolpix). Live imaging of infected neurons expressing eGFP alone or eGFP tau constructs was performed over a period of 48 h after infection at three time intervals. For these experiments, cells were plated on 35-mm gridded glass-bottom culture dishes (MatTek Corporation, Ashland, MA, USA) at 4 · 104 cells/cm2 and infected with 2000 amplicons per cm2. Images were taken with lowest practical laser intensity and shortest practical illumination time to limit photodynamic damage. Between imaging, cells were put back into the incubator. For morphological analysis, perikaryal areas were measured from the collapsed fluorescence images using ScionImage analysis software (beta 4.0.2; Scion Corporation, Frederick, MD, USA). Immunoblot analysis Primary cortical cells were harvested in radioimmunoprecipitation assay buffer (50 mmol/L Tris/HCl, pH 7.5, 150 mmol/L NaCl, 1 mmol/L EDTA, 1% NP-40, 0.5% deoxycholate, and 0.1% sodium dodecyl sulfate (SDS), pH 8.0) containing protease inhibitors (1 mmol/L phenylmethylsulphonylfluoride, 10 lg/mL each of leupeptin and pepstatin, and 1 mmol/L EGTA) and phosphatase inhibitors (1 mmol/L sodium o-vanadate, 20 mmol/L sodium fluoride, and 1 mmol/L sodium pyrophosphate) incubated for 30 min at 4C while shaking and centrifuged for 10 min at 10 000 g at 4C. The supernatant (lysate) was subjected to SDS– polyacrylamide gel elcetrophoresis and transferred to Immobilon-P (Millipore, Bedford, MA, USA) followed by immunoblotting. Detection used enhanced chemiluminescence using SuperSignal West Dura extended duration substrate (Pierce Biotechnology) and was performed according to the manufacturer’s protocol. Quantification of the blots was performed with Gel-Pro Analyzer 4.0 (Media Cybernetics L.P, Silver Spring, MD, USA). Other methods Cell toxicity was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) conversion assays. For MTT conversion assays, PC12 cells were cultured in triplicate in 96-well plates. MTT conversion was determined as described previously (Piontek et al. 1999). When indicated, pre-aggregated Ab(1–42) (Bachem, Bubendorf, Switzerland) that had been incubated at 400 lmol/L for 3 days at 37C, or, as an inactive control, the reverse peptide Ab(42–1) that had been treated identically was added. Concentrations of Ab40 and Ab42 from medium samples of transgenic cultures were determined using the ‘hAmyloid b40’ and ‘hAmyloid b42’ ELISA kits from The Genetics Company (Schlieren, Switzerland) according to the manufacturer’s protocol. Statistical analysis among experimental groups was performed using paired Student’s t-test. p values are *p < 0.05; **p < 0.01; ***p < 0.001.

Results

Disease-like tau modifications induce neurodegeneration in mouse primary cortical cultures To analyze a potential tau-dependent neurodegeneration in dissociated mouse neurons, fetal tau constructs were



expressed in primary cortical cultures using an ampliconbased HSV1 system, which leads to an efficient, transient expression of exogenous genes in post-mitotic neurons with minimal toxicity. Tau sequence was aminoterminally fused with eGFP as an epitope tag. In addition to wt tau sequence, PHP tau that mimics key structural and functional aspects of hyperphosphorylated tau isolated from AD brain was used (Fig. 1a). As a control, eGFP alone was expressed. The infection resulted in many cells expressing the respective constructs as evaluated by counter-staining with the nuclear dye DAPI (Fig. 1b). Staining against the neuronal marker protein microtubule-associated protein (MAP)2 revealed that a high number of neurons was infected. Fusion proteins from infected cultures as separated on SDS– polyacrylamide gel elcetrophoresis showed an apparent molecular weight of approximately 85 kDa in agreement with a calculated total mass for fetal tau (48 kDa) and eGFP (27 kDa). Expression levels were similar for all constructs as compared with tubulin as a reference (Fig. 1c). For all constructs, the expression peaked at day 2 post-infection (data not shown). Quantitative immunofluorescence analysis of infected cultures confirmed similar expression of the eGFP-fusion proteins in neurons (8.2 ± 0.7 and 8.8 ± 1.3 rel. units for wt tau and PHP tau, respectively). Changes in nuclear morphology as a result of the expressed protein were analyzed by fluorescence microscopy of DAPIstained infected neurons. Many PHP tau expressing neurons showed fragmented nuclei as a sign for neurodegeneration (Fig. 1d, middle). In contrast, most nuclei of neurons infected with HSV1-eGFP-wt tau or -eGFP alone were intact (Fig. 1d, top and bottom). Quantitation revealed that the amount of infected neurons with fragmented nuclei was about twice as high for PHP tau compared with wt tau expressing cells and threefold compared with eGFP (Fig. 1e, left). The results were very similar for infected human model neurons (Fig. 1e, right), indicating that PHP tau-induced a neurodegenerative effect independent of the neuronal cell type and species. Moreover, in both cell types, the expression of eGFP-wt tau resulted in a small but significant increase in the number of neurons with fragmented nuclei compared with eGFP alone. In order to determine whether the tau-induced neurodegeneration is associated with apoptotic mechanisms, the level of activated caspase 3 as a key enzyme for apoptosis was detected by immunoblot of lysates from infected cultures. Activated caspase 3 was about 2.5-fold higher for PHP tau compared with wt tau or eGFP (Fig. 1f). We did not observe a difference between wt tau- and eGFP-infected cultures. The data suggest that PHP tau-induced neurodegeneration is associated with an induction of apoptotic mechanisms. It is a matter of debate whether neurodegeneration in AD occurs via apoptotic or non-apoptotic mechanisms. For patients with AD and other tauopathies, it has been reported that degenerating neurons establish a ‘ballooned’ phenotype which appears to represent a non-apoptotic type of cell death

(Gleckman et al. 1999). eGFP as an epitope tag enables the observation of the fate of individual infected neurons by live imaging. In order to classify the neurons in different categories, MAP2 stainings of control cultures were performed and individual cells were imaged at high resolution. Based on size and morphological features, four anatomical classes could be differentiated: small and large pyramidal neurons, stellate (granule) cells, and bipolar or bitufted cells (Fig. 2a). The morphological criteria could be used to separate the neuronal subpopulations during live analysis. Low-density cultures of primary cortical cells were infected with HSV1 constructs and individual neurons were identified by their characteristic morphology. The same individual neurons could be located at subsequent time intervals of 24, 32, and 48 h post-infection by fluorescence and phase-contrast microscopy (Figs 2b and c). We did not observe obvious morphological alterations such as fragmented nuclei or beaded processes by this kind of analysis. However, we observed that PHP tau expressing neurons exhibited a slight but significant increase in the mean perikaryal area that was already evident after 24 and 32 h post-infection compared with wt tau or eGFP-expressing cells (Fig. 2d). When only pyramidal cells (small and large pyramidal neurons; see Fig. 2a) were evaluated, the increased perikaryal area was more pronounced than in non-pyramidal neurons (data not shown). This may suggest the development of a ‘ballooned’ phenotype in PHP tau expressing neurons. However, it should be noted that our approach does not allow to distinguish between a flattening of the cell body and an actual increase in cell volume and that ‘ballooned’ neurons could not be conserved by standardfixation protocols, probably due to a breakdown of cytoskeletal structure. Interestingly, PHP tau expression did not result in a selective loss of neurites. With all constructs, we saw a decrease in the mean number of primary neurites which was similar between wt tau expressing neurons (from 3.34 ± 0.27 at 24 h to 1.42 ± 0.32 at 48 h; n = 38) and PHP tau expressing neurons (from 3.75 ± 0.29 at 24 h to 1.78 ± 0.35 at 48 h; n = 36), corresponding to a loss of 57% and 53% of primary neurites, respectively. This is in agreement with our previous observation that tau-dependent neurodegeneration in organotypic hippocampal slices develops in the absence of changes in spine density and morphology (Shahani et al. 2006). The data suggest that the development of a ‘ballooned’ phenotype represents an early event during tau-dependent degeneration and that tau-induced cell death may involve apoptotic and non-apoptotic mechanisms. Transgenic expression of PS1 suppresses neurodegeneration that is induced by disease-like tau To determine the effect of PS on tau-induced neurodegeneration, primary cortical cultures from mice transgenic for human PS1 were prepared. The expression of the transgene was under the control of the human 3-hydroxy-3-methyl-



(a)

(b)

(c)

(d) (f) (e)

Fig. 1 Disease-like tau modifications induce neurodegeneration in primary cortical cultures. (a) Schematic representation of enhanced green fluorescent protein (eGFP) -tagged human tau constructs. The constructs are based on the fetal (352 aa) wild type (wt) tau isoform with an aminoterminally fused eGFP sequence (light gray box). Serine (S)/threonine (T) residues, which have been substituted with glutamate (E) to create a pseudohyperphosphorylation of these residues (PHP tau) are indicated. Numbers refer to the longest isoform containing 441 residues. The microtubule-binding repeats of tau are indicated by the dark grey box. (b) Fluorescence micrograph of herpes simplex virus type 1-eGFP-wt tau infected primary cortical cultures. 4,6-Diamidino-2-phenylindole (DAPI) -stained nuclei and eGFP (top), MAP2-staining (middle), and overlay (bottom) are shown. Note that many neurons express the fusion construct. (c) Immunoblot of lysates of cultures infected with the indicated constructs. Detection of GFP (top), tau (middle), and tubulin as a control for equal loading (bottom) is shown. Tau fusion proteins and eGFP separate at apparent molecular masses of 85 and 27 kDa, respectively. Endogenous mouse tau is detected in all cultures at 48 kDa. (d) Fluorescence micrograph of cultures infected with herpes simplex virus type 1eGFP-wt tau (top), -eGFP-PHP tau (middle), and -eGFP (bottom). Overlay (left panel) of eGFP (green), DAPI- (blue) and MAP2-staining (red), and DAPI -staining alone (right panel) is shown. Infected neurons are indicated by arrows and a fragmented nucleus in a PHP tau expressing neuron by an arrowhead. (e) Quantification of the percentage of fragmented nuclei of infected primary cortical cultures (left) and infected human model neurons (right). Cultures infected with PHP tau show an increased fraction of degenerated neurons compared with

wt tau or eGFP alone. (f) Immunoblot of lysates (top) and quantification of signal intensities of activated caspase 3 relative to tubulin (bottom) of cultures infected with the indicated constructs. Expression of PHP tau leads to an increased activation of caspase 3 compared with wt tau and eGFP. Cell culture, infections, cytochemistry, and immunoblot analysis were performed as described in ‘Experimental Procedures.’ For immunoblot analysis, 6 · 105 cells (c) or 2 · 105 cells (f) were plated on laminin coated coverslips, infected after 5 days, lysates prepared 2 days later, and separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 10% (c) or 7.5% (f) acrylamide. Detection was with polyclonal anti-GFP, polyclonal antiactivated caspase 3, monoclonal Tau-5, and monoclonal anti-tubulin (DM1A) antibody. For cytochemistry of primary cortical cultures, 5 · 104 cells/cm2 were plated on laminin coated coverslips, infected after 5 days, fixed with 4% p-formaldehyde 2 days later, and stained with monoclonal anti-MAP2 antibody and DAPI. Human model neurons were plated at 104 cells/cm2 on collagen-coated coverslips, treated for 2 weeks with cytostatica, further cultured for 4 days in DMEM-serum, infected, and processed for fluorescence analysis 4 days later. Neurodegeneration was determined by scoring infected cells for the presence of fragmented nuclei. Between 228 and 564 neurons from six sets of experiments (cortical neurons) and between 21 and 130 neurons from three sets of experiments (human model neurons) were evaluated per construct. For (f), immunoblots from three-independent sets of experiments were quantitated and the ratio of activated caspase 3 to tubulin calculated. Mean and SEM are shown. Scale bars: 50 lm (b) and 20 lm (d).

glutaryl-coenzyme A reductase promoter, a housekeeping promoter which drives a strong and ubiquitous expression with a high activity in neurons (Czech et al. 1997, 1998;

Leutner et al. 2000; Terro et al. 2002). About 50% of the cells from transgenic cultures exhibited a clear staining with a monoclonal PS1 antibody directed against the N-terminal



(a)

(b)

(c)

(d)

Fig. 2 Live imaging of the effect of tau expression on the degeneration of primary cortical cultures. (a) Anatomical classes of neuronal subpopulations based on MAP2 stainings. (b) Live enhanced green fluorescent protein (eGFP) -fluorescence (top) and phase-contrast image (bottom) of two adjacent neurons infected with herpes simplex virus type 1-eGFP (arrowheads and arrows) after 24, 32, and 48 h post-infection. Imaging of infected and morphologically intact neurons in low-density cultures was possible up to 48 h post-infection. (c) Time-lapse imaging of neurons infected with the constructs as indicated. Minor shifting of cell bodies as seen in the pseudohyperphosphorylated (PHP) tau panel was common. Many neurons that expressed PHP tau exhibited an increased perikaryal area. (d)

Quantitation of the perikaryal area of living neurons at 24 and 32 h post-infection with the constructs as indicated. Note the increased perikaryal area in neurons infected with PHP tau compared with wild type (wt) tau or eGFP-infected neurons. Cell culture, stainings, infections, and morphometric analysis were performed as described in ‘Experimental Procedures.’ For cytochemistry, cells were fixed 2 days after infection with 4% formaldehyde and stained with anti-MAP2 antibody. For live imaging, 4 · 104 cells/cm2 were plated on laminincoated gridded coverslips, infected after 5 days, and imaged at times indicated. Perikaryal areas were measured from the fluorescence images from 37–38 infected neurons per construct from 9–10 experiments. Mean and SEM are shown. Scale bars: 20 lm.

fragment of human PS1, which was consistent with an endoplasmic reticulum (ER) localization (Fig. 3a). Expression of the PS1 transgene completely abolished PHP tau-induced neurodegeneration and no difference in the percentage of neurons with fragmented nuclei was observed between wt tau and PHP tau expressing neurons (Fig. 3b). To test whether the absence of PHP tau-induced neurodegeneration on the PS1 background was due to a suppression of caspase 3 activation (see Fig. 1f), the level of activated caspase 3 was determined from immunoblots of lysates from PHP tau-infected PS1-transgenic and non-transgenic cultures (Fig. 3c). Compared with non-infected control cultures, PHP tau induced an about 70% increased level of activated caspase 3 in non-transgenic, but not in PS1-transgenic cultures (Fig. 3d, left). Also in wt tau infected cultures, the PS1 transgene appeared to reduce caspase 3 activation (Fig. 3d, right). The protective activity of PS1 with respect to tauinduced neurodegeneration was not affected by treatment with known c-secretase inhibitors (DAPT and WPE-III-31C; data not shown) indicating that c-secretase activity is not involved.

Ab amplifies tau-induced neurodegeneration Experimental evidence indicates that Ab and tau functionally interact in mediating neurodegeneration (Rapoport et al. 2002; Park and Ferreira 2005). To test whether Ab increases neuronal degeneration in tau-infected neurons, primary cultures from trangenic mice expressing the 695 aa long isoform of APP containing three FAD-mutations (SDL) under the control of the PDGF-promoter were prepared. The PDGF-promoter drives a high expression of the transgene in the mouse brain, which leads to the formation of amyloid plaques at an age of 18 month (Blanchard et al. 2003). To determine the amount and identity of Ab that is secreted by the cultured neurons, ELISAs of medium samples were performed with human-specific antibodies directed against Ab40 and Ab42 at time points equivalent to the experiments. As expected, we did not detect Ab in the medium of nontransgenic control cultures (detection limit: 100 pg/mL). In the medium of transgenic cells between 140–160 pg/mL (corresponding to 30–40 pmol/L) of Ab40 and Ab42 were detected (Fig. 4a). The equimolar concentration of Ab40 and



(a)

(c)

(b)

(d)

Fig. 3 Effect of presenilin (PS) 1 on tau-induced neurodegeneration. (a) Fluorescence micrograph of primary cortical cultures from mice transgenic for human PS1 (top) and non-transgenic mice (bottom). 4,6-Diamidino-2-phenylindole-stained nuclei (right) and PS stainings (all panels) are shown. (b) Quantification of the percentage of fragmented nuclei of infected primary cortical cultures from PS1-transgenic and non-transgenic (non tg) animals. The presence of PS1 abolishes pseudohyperphosphorylated (PHP) tau-induced neurodegeneration. (c) Immunoblot of activated caspase 3, tubulin, and tau of lysates from PS1- and non-tansgenic (non tg) cultures with or without (con) infection with herpes simplex virus type 1-PHP tau. (d) Quantification of signal intensities of activated caspase 3 relative to tubulin from lysates from PS1- and non-transgenic (non tg) cultures with or without (con) infection with herpes simplex virus type 1-PHP tau (left) and wild type (wt) tau (right). In contrast to non-transgenic cultures, infection with PHP tau does not lead to elevated activated caspase 3 in the presence of PS1. The PS1 transgene also appears to reduce wt

tau-induced caspase 3 activation. Cell culture, infections, cytochemistry, and immunoblot analysis were performed as described in ‘Experimental Procedures.’ For cytochemistry, 5 · 104 cells/cm2 were plated on laminin coated coverslips, fixed after 7 days with ethanol/ PEM, and stained with a monoclonal antibody against the N-terminal fragment of human PS1 and 4,6-diamidino-2-phenylindole. For immunoblot analysis, 2 · 105 cells were plated on laminin coated coverslips, infected after 5 days, lysates prepared 2 days later, and separated by sodium dodecyl sulfate–polyacrylamide gel elcetrophoresis. Detection was with polyclonal anti-activated caspase 3, monoclonal anti-tubulin (DM1A), and monoclonal Tau-5 antibody. Immunoblots from four to six sets of experiments were quantitated and the ratio of activated caspase 3 to tubulin was calculated. Mean and SEM are shown. The percentage of infected neurons with fragmented nuclei was determined by fluorescence microscopy. Between 326 and 798 neurons from four to five sets of experiments were evaluated per construct. Mean and SEM are shown. Scale bar: 20 lm.

Ab42 is in agreement with previous data showing that FAD mutations lead to an increased ratio of Ab42 to Ab40, whereas normally Ab42 represents only a minor fraction of the Ab produced (Citron et al. 1997; Eckmann et al. 1997; Araki et al. 2001). Infection with HSV1-wt tau of APPSDL-transgenic cultures resulted in an approximate 50% increase in the number of fragmented nuclei compared with non-transgenic controls (Fig. 4b). In contrast, neurotoxicity was not increased after expression of eGFP which is in agreement with previous results showing that tau is essential for Ab-induced toxicity

(Rapoport et al. 2002). Interestingly, PHP tau-induced toxicity and the degenerative effect of wt tau on an APPSDL-background were very similar to the toxicity of PHP tau in non-transgenic cultures. To test whether the effects of mutant APP-expression was through the production of Ab, the effect of two different c-secretase inhibitors known to prevent Ab peptide production was tested. Both the non-transition state inhibitor DAPT and the transition state analog WPE-III-31C reduced tau-induced neurodegeneration in APPSDL expressing cultures at concentrations that have previously been shown to inhibit c-secretase-mediated



(a)

(b)

(c)

(d)

Fig. 4 Effect of familial Alzheimer’s disease-mutated amyloid precursor protein (APP) on tau-induced neurodegeneration. (a) Concentration of amyloid b (Ab) 40 and 42 in the medium of APPSDLtransgenic cultures as measured by ELISA at times indicated following medium exchange. Ab40/42 of non-transgenic control cultures was below the detection limit. (b) Percentage of fragmented nuclei of infected neurons from APPSDL-transgenic and non-transgenic (non tg) cultures. The toxicity of wild type (wt) tau is amplified in the presence of the APP transgene. No increase was observed with pseudohyperphosphorylated (PHP) tau. (c) Effect of c-secretase inhibitors DAPT and WPE-III-31C on wt tau-induced neurodegeneration in APPSDLtransgenic cultures. Both inhibitors reduce the percentage of fragmented nuclei in infected neurons. (d) Quantification of signal intensities of PHF1-phosphorylated tau relative to total tau of lysates from APPSDL- and non-transgenic (non tg) cultures after infection with herpes simplex virus type 1-wt tau. The presence of APPSDL leads to an elevated phosphorylation of tau at PHF1 and AT8 sites. Cell culture, ELISA, infections, and immunoblot analysis were performed as described in ‘Experimental Procedures.’ For (a), 1.5 · 105 cells/cm2

were plated on laminin coated coverslips. After 5 days, the medium was exchanged against 500 lL fresh NB/B27 and the assay was performed with 50 lL sample each in duplicates. For (b), between 358 and 598 neurons from four to five sets of experiments were evaluated per construct for the presence of fragmented nuclei. For (c), cultures were kept in the presence of carrier (DMSO) or c-secretase inhibitors as indicated. The percentage of infected neurons with fragmented nuclei was determined from six randomly chosen microscopic fields per condition. No effect of the inhibitors in non-transgenic cultures was observed (12.5–25% neurons with fragmented nuclei for all conditions). For immunoblot analysis, 105 cells were plated on laminin coated coverslips, infected after 5 days, lysates prepared 3 days later, and separated by sodium dodecyl sulfate–polyacrylamide gel elcetrophoresis on 10% acrylamide. Detection was with phosphorylation-sensitive monoclonal PHF1 or AT8 and monoclonal Tau-5 antibody against all tau isoforms. Immunoblots from four- to fiveindependent sets of experiments were quantitated and the ratio of PHF1 or AT8 to Tau-5 signal calculated. Mean and SEM are shown for all experiments.

cleavage in cultured neuronal cells (Hass and Yankner 2005; Kim et al. 2005) (Fig. 4c). No effect of the inhibitors was observed in wt-tau infected non-transgenic control cultures (data not shown). Taken together the data suggest that (i) hyperphosphorylated tau is the limiting factor in Ab-induced

cell death, (ii) the mode by which PHP tau causes neurotoxicity is downstream of Ab, and (iii) tau mediates Ab-induced toxicity only when it can be phosphorylated at the disease-relevant sites that have been mutated in PHP tau. To confirm that Ab-induced degeneration is associated with



phosphorylation at sites that have been mutated in PHP tau, quantitative immunoblotting was performed with the antibodies PHF1 and AT8 that detect phosphorylated S396/S404 and S199/S202/T205, respectively. In fact, immunoreactivity for both antibodies in APPSDL-transgenic cultures was more than double compared with non-transgenic controls (Fig. 4d), indicating that Ab promotes neurodegeneration by increasing tau phosphorylation at AD relevant sites. To test whether the effect of APPSDL expression in infected primary neurons can be reproduced also by adding exogenous Ab in a cell line model, previously established PC12 clones stably transfected with different tau constructs were used (Fath et al. 2002). Also in these cells, PHP tau expression resulted in a reduced survival, which was most evident when the cells were differentiated with nerve growth factor (NGF) (Fig. 5a). Addition of pre-aggregated Ab(1–42) to the medium reduced the survival of tau expressing cells compared with a control culture in a Ab-concentrationdependent manner (Fig. 5b). The relative survival as calculated from MTT-conversion in the presence of Ab(1–42) compared with an inactive control peptide (a peptide with the reverse sequence 42–1) was significantly reduced in wt tau expressing cells compared with control cultures (Fig. 5c), indicating that the expression of human tau sensitizes the cells to Ab-induced neurotoxicity. Also in this system, Ab reduced the survival of wt tau expressing cells much more than the survival of PHP tau expressing or control cells (Fig. 5d). The relative survival was reduced in wt tau but not in PHP tau expressing cells with respect to a vector control line (Fig. 5e) or a tau fragment that lacked most of the phosphorylatable sites (data not shown). Taken together, the data also indicate that extracellular, aggregated Ab amplifies tau-dependent degeneration by a mechanism that requires tau-modification and confirm that the mode by which PHP tau causes neurotoxicity is downstream of Ab. It should however be noted that the concentration of pre-aggregated Ab that was required to achieve a significant effect was much higher than the Ab concentration in the medium of the primary cultures and that we did not see a detectable effect of aggregated Ab at 1 ng/mL or below. It is possible that cytosolic Ab is produced in the neurons of APPSDLtransgenic mice that is much more effective than exogenously added Ab.

Discussion

tagged PHP tau exerts a neurotoxic effect in transfected differentiated PC12 cells and virus-infected human model neurons (Fath et al. 2002). Here, we report that eGFP-tagged tau constructs have a very similar effect when expressed in mouse primary cortical neurons indicating that disease-like modified tau exerts a robust neurotoxic effect independent of the type of neuron and the epitope-tag used. Apoptotic DNA-fragmentation (Su et al. 1994; Lassmann et al. 1995; Smale et al. 1995) and increased level of activated caspase 3 have been reported in post-mortem brain tissue of AD patients (Gervais et al. 1999; Su et al. 2001). However, the relative contribution of apoptosis to neurodegeneration occurring in AD is unclear and other mechanisms may play an important role. Using live-imaging of degenerating neurons, we found that tau-mediated neurodegeneration is accompanied by an increase in the perikaryal area. This may indicate that the development of a ‘ballooning phenotype’ is involved in tau-mediated neurodegeneration. ‘Ballooning’ of neurons as a typical event in AD and other tauopathies has been reported (Dickson et al. 1986; Fujino et al. 2004) and apoptosis and the development of a ballooned morphology appear as two distinct death mechanisms as no clear connection could be made between apoptotic DNA-fragmentation and the ballooned phenotype in Pick’s disease (Gleckman et al. 1999). Interestingly, ballooned neurons could not be conserved by and came off the coverslip during the fixation procedure but were only evident by live-imaging, a constraint probably caused by a breakdown of the cytoskeletal structure. This also implies that methods which are based on a quantitative assessment of fixed cells or tissue may overestimate apoptotic versus nonapoptotic mechanisms of cell death due to the loss of necrotic cells. It should however be noted that in our experiments, ballooning occurred as an early event (24–36 h postinfection) and it is possible that neurons that balloon eventually undergo apoptosis at a later time point. The molecular mechanisms by which PHP tau exerts a neurotoxic effect still remain to be elucidated. It seems likely that the function of tau as a microtubule-associated protein is not involved as PHP tau does not actively destabilize microtubules and taxol-induced microtubule stabilization does not reverse the neurodegenerative effect (Fath et al. 2002). Tau interacts with a variety of proteins involved in signaling cascades and subtle changes in cellular signaling caused by pathologically modified tau might induce neurodegeneration (for review see Brandt and Leschik 2004).

Hyperphosphorylated tau is neurotoxic in primary cortical cultures Simulation of tau hyperphosphorylation by introducing negatively charged residues at sites which are modified in AD provides a unique tool to determine the effect of sitespecific and permanent phosphorylation characteristic for pathological tau. Previously, we have shown that FLAG-

PS1 and Ab inversely modulate tau-induced neurodegeneration To analyze for a potential functional interference of PS1 with tau-mediated toxicity, tau constructs were expressed in cortical neurons from transgenic animals. Transgenically expressed PS1 completely abolished neurodegeneration induced by PHP tau expression and reduced activated



caspase 3 to background levels. This is in agreement with data describing PS1 to be anti-apoptotic (Bursztajn et al. 1998; Terro et al. 2002; Baki et al. 2004) and to protect against FAD mutation-induced amyloid pathology (Wang et al. 2006). It remains to be elucidated what is the relevant mechanism how PS1 counteracts PHP tau-induced neurotoxicity as PS1 also appeared to reduce wt tau-induced caspase 3 activation (Fig. 3d). The use of neurons prepared from mice that are transgenic for FAD-mutated APP allows the analysis of the functional interference between amyloid pathology and tau. We

observed that APPSDL expression elevated wt tau-induced neurodegeneration while PHP tau-induced neurodegeneration was not affected indicating that hyperphosphorylated tau is the limiting factor in Ab-induced neurodegeneration and that tau mediates Ab-induced toxicity only when it can be phosphorylated at disease-relevant sites. The same results were obtained with a cell line model treated with aggregated Ab(1–42), which may provide a useful system to screen for drugs that interfere with the generation of AD-relevant tau modifications downstream of Ab. Several studies have shown that Ab can induce or amplify apoptosis in cultured

(a)

(b)

(d)

(c)

(e)



(a)

(b)

(c)

Fig. 6 Schematic representation of the effects of amyloid b (Ab) and presenilin (PS) 1 on tau-mediated degeneration in a scenario without Ab (a) and with Ab in the absence (b) or presence (c) of PS1. Tau is the limiting factor in Ab-induced neurodegeneration and the mode by which hyperphosphorylated tau (HP-tau) causes neurotoxicty is downstream of Ab. Ab and PS1 inversely modulate neurodegeneration at distinct steps with hyperphosphorylated tau (HP-tau) being central in the degenerative cascade.

cell lines or neurons (Estus et al. 1997; Troy et al. 2000; McPhie et al. 2001) and that tau is essential for Ab-toxicity (Rapoport et al. 2002). However, it is less clear which pathway mediates Ab-toxicity and how tau is involved. It is known that Ab can increase tau-phosphorylation at AD-relevant sites (Busciglio et al. 1995; Greenberg et al. 1995; Takashima et al. 1998; Rapoport and Ferreira 2000) and we also observed increased phosphorylation at the PHF-1 and AT8 epitopes, i.e. at sites that – with the exception of T205 – have been mutated in our PHP tau Fig. 5 Effect of pre-aggregated amyloid b (Ab) on tau-induced neurodegeneration in a neural cell line model. (a) 3-(4,5-Dimethylthiazol2-yl)-2,5-diphenyltetrazolium bromide (MTT) conversion assay of PC12 cultures stably transfected with FLAG-epitope tagged human wild type (wt) tau, pseudohyperphosphorylated (PHP) tau, or a vector control in the absence or presence of NGF. PHP tau expression resulted in a decreased MTT conversion compared with wt tau or vector control cells, which was most evident after differentiation with NGF. (b) MTT conversion of PC12 cultures stably transfected with FLAG-epitope tagged human wt tau (tau) or a vector control treated with different concentrations of pre-aggregated Ab42(1–42) or a control peptide with reverse sequence Ab(42–1). Note that tau expression sensitizes cells compared with control cells when treated with Ab42(1– 42). (c) Relative survival was calculated from MTT conversion with pre-aggregated Ab(1–42) relative to MTT conversion with a control peptide with reverse sequence Ab(42–1) of PC12 cultures stably transfected with FLAG-epitope tagged human wt tau (wt tau) and a

construct. It should however be noted that the correlation between Ab-induced neurodegeneration and increased tau phosphorylation does not necessarily imply a causal relation. For example, it has been shown that glycogen synthase kinase 3b induces both phosphorylation of tau and of a mitochondrial enzyme with the latter causing a toxic effect (Hoshi et al. 1996). However, our finding that Ab does not affect PHP tau-induced neurodegeneration provides strong evidence that Ab induces a disease-like tau modification, which is both required and sufficient for neurodegeneration (Fig. 5). The results are in agreement with the amyloid hypothesis, which predicts Ab to be the primary cause of the disease leading to neurodegeneration as a downstream effect, and they confirm a mandatory role for tau modification in the downstream process. Interestingly, ELISA-based determination of Ab concentration in the medium indicated that very low concentrations were able to potentiate tau-mediated toxicity to a level that was observed for PHP tau. Although, our data show that aggregated exogenous Ab has also a very similar effect albeit at a much higher concentration, it cannot be excluded that intracellular Ab is the species that triggers taudependent degeneration. In fact, several transgenic mice constructed to express Ab only intracellularly develop neuronal degeneration (La Ferla et al. 1995; Chui et al. 1999) and tau hyperphosphorylation (Echeverria et al. 2004). Taken together, the data indicate that Ab induces diseaserelevant tau modifications that are both required and sufficient for neurodegeneration. It should be noted that tau-dependent toxicity does not require the presence of mutations that have been identified in frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) cases and which have never been observed in AD but provide the basis for most transgenic mouse models to date (for review see Brandt et al. 2005). Furthermore, our data indicate that PS1 is able to protect neurons against the vector control. Human wt tau expression results in an Ab-dependent reduction of survival. (d) MTT conversion of PC12 cultures stably transfected with FLAG-epitope tagged human wt tau (wt-tau), PHP tau, and a vector control treated with 5 lmol/L Ab(1–42) or a control peptide (Ab(42–1)). Note that Ab reduces the survival of wt tau expressing cells much more than the survival of PHP tau expressing or control cells. (e) Relative survival as calculated from the data shown in (d). Wt tau expression but not PHP tau expression sensitizes the cells for Ab-dependent reduction of survival. MTT conversion data were from six experiments using two independent clonal lines per construct. Cells were cultured on ELISA plates for 8 days in the presence of 100 ng/mL NGF (b–e) or as indicated (a). Pre-aggregated Ab(1–42) or an inactive control peptide Ab(42–1) were added as indicated. For (d), MTT conversion at day 8 was calculated as percentage from MTT conversion at day 1 as determined from an ELISA-plate that had been prepared in parallel. Mean and SEM are shown.



toxic effect of disease-relevant tau modifications. Thus, as schematically depicted in Fig. 6, Ab and PS1 inversely modulate neurodegeneration at distinct steps with tau hyperphosphorylation being central in the neurodegenerative cascade. Suppression of tau expression or inhibition of tau modification (e.g. by overexpression of competing tau fragments) may provide a useful strategy to prevent neurodegeneration that has been triggered by different AD-relevant upstream mechanisms. Acknowledgements We thank Dr Laurent Pradier (Sanofi-Aventis, France) for supplying mice transgenic for PS1 and APP695 with the SDL mutations, Dr Peter Davies (Albert Einstein College of Medicine, Bronx, NY, USA) for PHF1 antibody, Dr Eberhard Buse (Covance, Mu¨nster) for supplying amyloid ELISA kits and Anglika Hilderink for expert technical assistance. This work was supported by the Ministry for Science and Culture of Lower Saxony and a fellowship of the graduate college 612 of the Deutsche Forschungsgemeinschaft (CW).

References Araki W., Yuasa K., Takeda S., Takeda K., Shirotani K., Takahashi K. and Tabira T. (2001) Pro-apoptotic effects of presenilin 2 (PS2) overexpression is associated with down-regulation of Bcl-2 in cultured neurons. J. Neurochem. 79, 1161–1168. Baki L., Shioi J., Wen P., Shao Z., Schwarzman A., Gama-Sosa M., Neve R. and Robakis N. K. (2004) PS1 activates PI3 K thus inhibiting GSK-3 activity and tau-overphosphorylation: effects of FAD mutations. EMBO J. 23, 2586–2596. Blanchard V., Moussaoui S., Czech C. et al. (2003) Time sequence of maturation of dystrophic neurites associated with Abeta deposits in APP/PS1 transgenic mice. Exp. Neurol. 184, 247–263. Brandt R. and Leschik J. (2004) Functional interactions of tau and their relevance for Alzheimer’s disease. Curr. Alzheimer Res. 1, 255–269. Brandt R., Leger J. and Lee G. (1995) Interaction of tau with the neural plasma membrane mediated by tau’s amino-terminal projection domain. J. Cell Biol. 131, 1327–1340. Brandt R., Hundelt M. and Shahani N. (2005) Tau alteration and neuronal degeneration in tauopathies: mechanisms and models. Biochim. Biophys. Acta 1739, 331–354. Bursztajn S., De Souza R., Mc Phie D. L., Berman S. A., Shioi J., Robakis N. K. and Neve R. L. (1998) Overexpression in neurons of human presenilin 1 or presenilin 1 familial Alzheimer disease mutant does not enhance apoptosis. J. Neurosci. 18, 9790–9799. Busciglio J., Lorenzo A., Yeh J. and Yankner B. A. (1995) Beta-amyloid fibrils induce tau phosphorylation and loss of microtubule binding. Neuron 14, 879–888. Calhoun M. E., Wiederhold K. H., Abramowski D., Phinney A. L., Probst A., Sturchler-Pierrat C., Staufenbiel M., Sommer B. and Jucker M. (1998) Neuron loss in APP transgenic mice. Nature 395, 755–756. Chui D. H., Tanahashi H., Ozawa K. et al. (1999) Transgenic mice with Alzheimer presenilin 1 mutations show accelerated neurodegeneration without amyloid plaque formation. Nat. Med. 5, 560–564. Citron M., Westaway D., Xia W., Carlson G., Diehl T., Levesque G. and Selkoe D. E. A. (1997) Mutant presenilins of Alzheimer’s disease increase the production of 42-residue amyloid beta-protein in both transfected cells and transgenic mice. Nat. Med. 3, 67–72.

Czech C., Delaere P., Macq A. F. et al. (1997) Proteolytic processing of mutated human amyloid precursor protein in transgenic mice. Brain Res. Mol. Brain Res. 47, 108–116. Czech C., Lesort M., Tremp G. et al. (1998) Characterization of human presenilin 1 transgenic rats: increased sensitivity to apoptosis in primary neuronal cultures. J. Neurosci. 87, 325–336. Dickson D. W., Yen S. H., Suzuki K. I., Davies P., Garcia J. H. and Hirano A. (1986) Ballooned neurons in select neurodegenerative diseases contain phosphorylated neurofilament epitopes. Acta Neuropathol. (Berl.) 71, 216–223. Echeverria V., Ducatenzeiler A., Dowd E. et al. (2004) Altered mitogenactivated protein kinase signaling, tau hyperphosphorylation and mild spatial learning dysfunction in transgenic rats expressing the beta-amyloid peptide intracellularly in hippocampal and cortical neurons. J. Neurosci. 129, 583–592. Eckmann C. B., Mehta N. D., Crook R. et al. (1997) A new pathogenic mutation in the APP gene (I716 V) increass the relative proportion of Ab42(43). Hum. Mol. Gen. 5, 2087–2089. Eidenmuller J., Fath T., Hellwig A., Reed J., Sontag E. and Brandt R. (2000) Structural and functional implications of tau hyperphosphorylation: information from phosphorylation-mimicking mutated tau proteins. Biochemistry 39, 13 166–13 175. Eidenmuller J., Fath T., Maas T., Pool M., Sontag E. and Brandt R. (2001) Phosphorylation-mimicking glutamate clusters in the proline-rich region are sufficient to simulate the functional deficiencies of hyperphosphorylated tau protein. Biochem. J. 357, 759–767. Estus S., Tucker H. M., Van Rooyen C., Wright S., Brigham E. F., Wogulis M. and Rydel R. E. (1997) Aggregated amyloid-beta protein induces cortical neuronal apoptosis and concomitant ‘apoptotic’ pattern of gene induction. J. Neurosci. 17, 7736–7745. Fath T., Eidenmuller J. and Brandt R. (2002) Tau-mediated cytotoxicity in a pseudohyperphosphorylation model of Alzheimer’s disease. J. Neurosci. 22, 9733–9741. Fujino Y., Delucia M. W., Davies P. and Dickson D. W. (2004) Ballooned neurones in the limbic lobe are associated with Alzheimer type pathology and lack diagnostic specificity. Neuropathol. Appl. Neurobiol. 30, 676–682. Gervais F. G., Xu D., Robertson G. S. et al. (1999) Involvement of caspases in proteolytic cleavage of Alzheimer’s amyloid-beta precursor protein and amyloidogenic A beta peptide formation. Cell 97, 395–406. Gleckman A. M., Jiang Z., Liu Y. and Smith T. W. (1999) Neuronal and glial DNA fragmentation in Pick’s disease. Acta Neuropathol. (Berl.) 98, 55–61. Gong C. X., Singh T. J., Grundke-Iqbal I. and Grundke-Iqbal K. (1993) Phosphoprotein phosphatase activities in Alzheimer disease brain. J. Neurochem. 61, 921–927. Greenberg S. M., Qiu W. Q., Selkoe D. J., Ben-Itzhak A. and Kosik K. S. (1995) Aminoterminal region of beta-amyloid precursor protein activates mitogen-activated protein kinase. Neurosci. Lett. 198, 52–56. Hass M. R. and Yankner B. (2005) A c-secretase-independent mechanism of signal transduction by the amyloid precursor protein. J. Biol. Chem. 280, 36 895–36 904. Hoshi M., Takashima A., Noguchi K., Murayama M., Sato K., Kondo S., Saitoh Y., Ishiguro K., Hoshino T. and Imahori K. (1996) Regulation of mitochondrial pyruvate dehydrogenase activity by tau protein kinase I/glycogen synthase kinase 3beta in brain. Proc. Natl Acad. Sci. U S A 93, 2719–2723. Jicha G. A., Weaver C., Lane E., Vianna C., Kress Y., Rockwood J. and Davies P. (1999) CAMP-dependent protein kinase phosphorylations on tau in Alzheimer’s disease. J. Neurosci. 19, 7486–7494.



Kennessey A. A. and Yen S. H. (1993) The extent of phosphorylation of fetal tau is comparable to that of PHF-tau from Alzheimer paired helical filaments. Brain Res. 629, 40–46. Kidd M. (1963) Paired helical filaments in electron microscopy of Alzheimer’s disease. Nature 197, 192–193. Kim D. Y., MacKenzie L. A., Care B. W., Pettingell W. H. and Kovacs D. M. (2005) Presenilin/c-secretase-mediated cleavage of the voltage-gated sodium channel b2-subunit regulates cell adhesion and migration. J. Biol. Chem. 280, 23 251–23 261. La Ferla F. M., Tinkle B. T., Bieberich C. J., Haudenschild C. C. and Jay G. (1995) The Alzheimer’s Ab peptide induces neurodegeneration and apoptotic cell death in transgenic mice. Nat. Genet. 9, 21–29. Lassmann H., Bancher C., Breitschopf H., Wegiel J., Bobinski M., Jellinger K. and Wisniewski H. M. (1995) Cell death in Alzheimer’s disease evaluated by DNA fragmentation in situ. Acta Neuropathol. (Berl.) 89, 35–41. Leutner S., Czech C., Schindowski K., Touchet N., Eckert A. and Mu¨ller A. W. (2000) Reduced antioxidant enzyme activity in brains of mice transgenic for human presenilin-1 with single or multiple mutations. Neurosci. Lett. 292, 87–90. Mattson M. P. (2004) Pathways towards and away from Alzheimer’s disease. Nature 430, 631–639. McPhie D. L., Golde T., Eckmann C. B., Yager D., Brant J. B. and Neve R. L. (2001) Beta-secretase cleavage of the amyloid precursor protein mediates neuronal apoptosis caused by familial Alzheimer’s disease mutations. Brain Res. Mol. Brain Res. 97, 103–113. Neve R. L., Neve K. A., Nestler E. J. and Carlezon W. A. Jr (2005) Use of herpes virus amplicon vectors to study brain disorders. Biotechniques 39, 381–391. Otth C., Concha I. I., Arendt T., Stieler J., Schliebs R., Gonzalez-Billault C. and Maccioni R. B. (2002) A/betaPP induces cdk5-dependent tau hyperphosphorylation in transgenic mice Tg2576. J. Alzheimers Dis. 4, 417–430. Park S. Y. and Ferreira A. (2005) The generation of a 17 kDa neurotoxic fragment: an alternative mechanism by which tau mediates beraamyloid-induced neurodegeneration. J. Neurosci. 25, 5363–5375. Patrick G. N., Zukerberg L., Nicolic M., De la Monte S., Dikkes P. and Tsai L. H. (1999) Conversion of p35 to p25 deregulates Cdk5 activity and promotes neurodegeneration. Nature 402, 614–622. Piontek J., Chen C. C., Kempf M. and Brandt R. (1999) Neurotrophins differentially regulate the survival and morphological complexity of human CNS model neurons. J. Neurochem. 73, 139–146. Rank K. B., Pauley A. M., Bhattacharya K., Wang Z., Evans D. B., Fleck T., Johnston J. A. and Sharma S. K. (2002) Direct interaction of soluble human recombinant Abeta 1-42 results in tau aggregation und hyperphosphorylation by tau protein kinase II. FEBS Lett. 514, 263–268. Rapoport M. and Ferreira A. (2000) PD98059 prevents neurite degeneration induced by fibrillar beta-amyloid in mature hippocampal neurons. J. Neurochem. 74, 125–133.

Rapoport M., Dawson H. N., Binder L. I., Vitek M. P. and Ferreira A. (2002) Tau is essentiell for beta-amyloid-induced neurotoxicity. Proc. Natl Acad. Sci. U S A 99, 6364–6369. Shahani N., Subramaniam S., Wolf T., Tackenberg C. and Brandt R. (2006) Tau aggregation and progressive neuronal degeneration in the absence of changes in spine density and morphology after targeted expression of Alzheimer’s disease-relevant tau constructs in organotypic hippocampal slices. J. Neurosci. 26, 6103–6114. Smale G., Nichols N. R., Brady D. R., Finch C. E. and Horton W. E. Jr (1995) Evidence for apoptotic cell death in Alzheimer’s disease. Exp. Neurol. 133, 225–230. Sturchler-Pierrat P. C., Abramowski D., Duke M. et al. (1997) Two amyloid precursor protein transgenic mouse models with Alzheimer disease-like pathology. Proc. Natl Acad. Sci. U S A 94, 13 287–13 292. Su J. H., Anderson A. J., Cummings B. J. and Cotman C. W. (1994) Immunohistochemical evidence for apoptosis in Alzheimer’s disease. Neuroreport 5, 2529–2533. Su J. H., Zhao M., Anderson A. J., Srinivasan A. and Cotman C. W. (2001) Activated caspase-3 expression in Alzheimer’s and aged control brain: correlation with Alzheimer pathology. Brain Res. 898, 350–357. Takashima A., Honda T., Yasutake K., Michel G., Murayama M., Murayama O., Ishiguro K. and Yamaguchi H. (1998) Activation of tau protein kinase I/glycogen synthase kinase-3beta by amyloid beta peptide (25-35) enhances phosphorylation of tau in hippocampal neurons. Neurosci. Res. 31, 317–323. Tanzi R. E. and Bertram L. (2005) Twenty years of the Alzheimer’s disease amyloid hypothesis: a genetic perspective. Cell 120, 545–555. Terro F., Czech C., Esclaire F., Elyaman W., Yardin C., Baclet M.-C., Touchet N., Tremp G., Pradier L. and Hugon J. (2002) Neurons overexpressing mutant presenilin-1 are more sensitive to apoptosis induced by endoplasmic reticulum-golgi stress. J. Neurosci. Res. 69, 530–539. Tomidokoro Y., Ishiguro K., Harigaya Y., Matsubara E., Ikeda M., Park J. M., Yasutake K., Kawarabayashi T., Okamoto K. and Shoji M. (2001) Abeta amyloidosis induces the initial stage of tau accumulation in APP(Sw) mice. Neurosci. Lett. 299, 169–172. Troy C. M., Rabacchi S. A., Friedman W. J., Frappier T. F., Brown K. and Shelanski M. L. (2000) Caspase-2 mediated neuronal cell death induced by b-amyloid. J. Neurosci. 20, 1386–1392. Wang R., Wang B., He W. and Zheng H. (2006) Wild-type presenilin 1 protects against Alzheimer disease mutation-induced amyloid pathology. J. Biol. Chem. 281, 15 330–15 336. Zheng W. H., Bastianetto S., Mennicken F., Ma W. and Kar S. (2002) Amyloid beta peptide induces tau phosphorylation and loss of cholinergic neurons in rat primary septal cultures. J. Neurosci. 115, 201–211.
