Caspase-9 Activation and Caspase Cleavage of tau in the Alzheimer's Disease Brain

Neurobiology of Disease 11, 341–354 (2002)doi:10.1006/nbdi.2002.0549

Caspase-9 Activation and Caspase Cleavageof tau in the Alzheimer’s Disease Brain

Troy T. Rohn,* ,1 Robert A. Rissman,† Michael C. Davis,*Young Eun Kim,* Carl W. Cotman,† and Elizabeth Head†

*Department of Biology, Science/Nursing Building, Room 228, Boise State University,Boise, Idaho 83725 and †Institute for Brain Aging and Dementia,University of California at Irvine, Irvine, California 92697

Received March 21, 2002; revised July 15, 2002; accepted for publication August 29, 2002

Accumulating evidence supports a role for the activation of proteolytic enzymes, caspases, in theAlzheimer’s disease (AD) brain. Neurons committed to apoptosis may do so through a mitochondrialpathway employing caspase-9 or through an alternative, receptor-mediated pathway involvingcaspase-8. Considering the role of mitochondrial dysfunction in AD, we examined the possibleactivation of caspase-9 in the AD brain using an antibody that recognizes the active fragments ofcaspase-9, but not the full-length proform of the enzyme. In vivo immunohistochemical analysisdemonstrated little caspase-9 activation in the majority of hippocampal brain sections from controlbrains. However, labeling of neurons as well as dystrophic neurites within plaque regions wasobserved in all AD hippocampal brain sections examined. In addition, active caspase-9 was colocal-ized with active caspase-8 and the accumulation of caspase-3-cleavage products of fodrin. Theactivation of caspase-9 was also observed in neurons positive for oxidative damage to DNA/RNA. Aquantitative analysis indicates that as the number of neurons containing neurofibrillary tangles (NFTs)increases, the extent of caspase-9 activation decreases, supporting the idea that caspase-9 activationmay precede NFT formation. In addition, a site-directed caspase-cleavage antibody was designed tothe amino-terminal caspase-3 consensus cleavage site located in tau, and shown to be an effectivemarker for caspase-cleaved fragments of tau in vitro. Analysis with this antibody using age-matchedcontrol or AD brain sections demonstrated no staining in control brains while widespread labeling ofNFTs, neuropil threads, and dystrophic neurites was observed in AD sections. Taken together, theseresults demonstrate the activation of caspases and cleavage of tau in the AD brain, events which mayprecede and lead to the formation of NFTs. © 2002 Elsevier Science (USA)

INTRODUCTION

One mechanism responsible for the extensive neu-ronal cell loss associated with Alzheimer’s disease(AD) may be the activation of apoptotic pathways(Marx, 2001). Apoptosis is a highly conserved form ofcell death that is characterized by chromatin conden-sation, nuclear fragmentation, cytoplasmic membraneblebbing, and cell shrinkage (Kerr et al., 1972). Criticalto the execution of apoptosis is a family of aspartateproteases, known as caspases, that are activated from

1 To whom correspondence should be addressed. Fax: 208-426-

0969-9961/02 $35.00© 2002 Elsevier Science (USA)All rights reserved. 341

procaspases to their active form by proteolysis (Thorn-berry & Lazebnik, 1998). Caspases can serve both asinitial transducers of apoptotic stimuli and also asfinal executioners of death. Two major pathways ofapoptosis have been characterized: the death-receptorpathway in which caspase-8 plays a pivotal role andthe mitochondrial pathway involving oxidative dam-age and activation of caspase-9 (Ashkenazi & Dixit,1999). In both pathways, caspase-8 and -9, respectively,act as initiator caspases leading to downstream activa-tion of executioner caspases, particularly caspase-3.

Several recent studies have employed caspase-cleavage site-directed antibodies as specific probes for

Key Words: Alzheimer’s disease; tau; caspase
spase-3; neurofibrillary tangles. -9; ca
apoptosis and demonstrated the activation of apopto-4267. E-mail: [email protected].

tic pathways in the AD brain (Yang et al., 1998; Gervaiset al., 1999; Rohn et al., 2001b). Using this approach, wehave recently developed an antibody to the active18-kDa fragment of caspase-8 and demonstrated thelocalization of this fragment within neurofibrillarytangles (NFTs) and in neurites in plaque-rich regionsin the AD brain (Rohn et al., 2001a). On the basis of thisstudy, it was suggested that cell–surface interaction of�-amyloid in neurons might trigger stimulation of thereceptor-mediated pathway of apoptosis and cleavageof critical cellular proteins (Rohn et al., 2001a). How-ever, evidence also suggests significant oxidativedamage in the AD brain possibly related to directoxidizing effects of �-amyloid (Mattson et al., 1999;Nunomura et al., 2001), leading to the hypothesis thatthe mitochondrial pathway of apoptosis may also beactivated in selected neurons within the AD brain.

In the present study, the role of activated caspase-9was investigated in the AD brain using a caspase-cleavage site-directed antibody directed toward theN-terminus of the auto activation cleavage site ofcaspase-9. This commercial antibody recognizes theactive fragments of caspase-9, but not the inactive,precursor form of the enzyme. A recent study, usingthis antibody, has demonstrated the presence of acti-vated caspase-9 in the AD brain by western blot anal-ysis (Lu et al., 2000). By immunohistochemical analy-sis, we now show the presence of activated caspase-9in the AD brain and within tangle-bearing neurons.Furthermore, active caspase-9 colocalized with activecaspase-8 and caspase-3 cleavage products of fodrinwithin neurons exhibiting NFT morphology. Theseresults support a role for the activation of both thereceptor-mediated and mitochondrial-mediated path-ways of apoptosis in the AD brain. In addition, evi-dence is presented demonstrating caspase cleavage oftau in the AD brain. Based on this study and alongwith previous studies (Rohn et al., 2001a,b; Su et al.,2001), a putative model is discussed whereby the roleof caspases in the cleavage of important cellular pro-teins, per se, and not the full execution of apoptosismay actually be more important for driving neurondysfunction and pathology observed in AD.

MATERIALS AND METHODS

Materials

All chemicals used were of the highest grade avail-able. The rabbit (polyclonal) anti-caspase-9 cleavagesite (315/316) specific antibody (caspase-9 CSSA) was

purchased from Biosource International (Camarillo,CA). The monoclonal antibody, AT8 (neurofibrillarytangles) was purchased from Innogenetics (Ghent,Belgium) and 6E10 (anti-A�-16) was purchased fromSignet Laboratories, Inc. (Dedham, MA). Apopain/YAMA/caspase-3 (nickel-NTA-agarose associated)expressed in E. coli was from Upstate Biotechnology(Lake Placid, NY). The monoclonal antibody, 8oxodG,was from QED Bioscience, (San Diego, CA). This anti-body recognizes 8-hydroxy-2�-deoxyguanosine (oh8dG),8-hydroxyguanine (oh8G), as well as 8-hydroxy-guanosine (oh8G), and serves as a marker of oxidativedamage to DNA and RNA. A multiple antigen pep-tide (MAP) containing the QGGYTMHQDQ sequenceto the amino terminal caspase-3 cleavage site of tauwas used for the generation of polyclonal antibodiesand was synthesized by Research Genetics, Inc.(Huntsville, AL). The peptide was injected into twodifferent rabbits and collected sera were monitored forantibody titer by ELISA. All injections and collectionof antisera were contracted out to Research Genetics,Inc. (Huntsville, AL). A Sulfolink kit used to affinitypurify antibody was purchased from Pierce (Rock-ford, IL).

Human Subjects

Autopsy brain tissue from the hippocampus andentorhinal cortex of seven neuropathologically con-firmed AD cases and seven nondemented cases withno evidence of senile plaques or NFTs was studied.Preliminary studies indicated that paraformaldehyde-fixed tissue samples were necessary to visualizecaspase-9 CCSA with either no or weak labeling ob-served in formalin-fixed tissue. In addition, cases withshort post mortem intervals (PMI) were selected tominimize possible post mortem effects leading tocaspase activation. Case demographics are presentedin Table 1. Age at death was not significantly differentbetween AD (mean, 68 � 8.29) and controls (mean,74 � 11.1). For quantification experiments, autopsybrain tissue from the hippocampus, parahippocampal,and frontal cortex of 13 cases with a range of MiniMental Status Examination Scores (MMSE) between 0and 30 (mean, 10.46 � 2.21) was examined. Agesranged from 69 to 92 years (mean, 77.31 � 1.94) andpost mortem intervals ranged from 2.75 to 7.5 h (mean,4.61 � 0.41). Frontal cortex was available from 9/13cases and the hippocampus was available on 8/13cases with 6/13 cases having both regions availablefor study. Case demographics are presented in Table2. Human brain tissues used in this study were pro-

342 Rohn et al.

© 2002 Elsevier Science (USA)All rights reserved.

vided by the Institute for Brain Aging and DementiaTissue Repositories at the University of California,Irvine.

Cell Culture

Rat PC12 cells (ATCC, Manassas, VA) were grownon rat tail collagen-coated dishes in F12K media (LifeTechnologies, Inc.) containing 15% horse serum, 2.5%fetal bovine serum, 100 units/ml penicillin, and 100�g/ml streptomycin. To induce differentiation to a

neuronal phenotype, PC12 cells were maintained inmedium containing 100 ng/ml NGF for 10–12 days.

Western Blot Analysis

Western blot analysis was performed as previouslydescribed (Rohn et al., 2000). PC12 extracts were pre-pared by adding ice-cold lysis buffer (50 mM Tris-HCL, 150 mM NaCl, 1% NP-40, 0.25% deoxycholate, 1mM EGTA, pH 7.4), followed by centrifugation. Theresulting supernatant was analyzed for protein con-

TABLE 1

Case Demographics

Case GroupAge(yrs) Sex PMI MMSE

Neuropathologicaldiagnosis

Braak & Braakstage Cause of death

1 AD 66 M 7 7 AD N/A Respiratory failure2 AD 71 F 7 7 AD VI Unknown3 AD 51 F 2.5 3 AD N/A Respiratory failure4 AD 77 M 3.6 22 AD V Respiratory failure5 AD 68 F 5 5 AD VI Cardiac Arrest6 AD 70 M 3.8 9 AD VI End stage AD7 AD 73 M 2.7 7 AD VI End stage AD8 Control 69 M 6.6 N/A Normal 0 N/A9 Control 68 F 6 N/A Normal II N/A

10 Control 90 F 6.5 N/A Normal N/A Heart failure11 Control 70 M 5.2 N/A Normal 0 N/A12 Control 64 F 8 N/A Normal N/A Myocardial infarction13 Control 90 F 8 N/A Normal I N/A14 Control 67 M 2.7 N/A Normal N/A Cardiomyopathy

Note: PMI, post mortem interval; MMSE, Mini Mental State Examination; N/A, not available.

TABLE 2

Case Demographics for Quantification Study

CaseAge(yrs) Sex PMI MMSE

Neuropathologicaldiagnosis

Braak & Braakstage Cause of death

1 74 M 2.75 0 AD V Dehydration2 88 F 4.33 0 AD V N/A3 85 M 5.3 4 AD V AD4 73 M 4.3 7 AD V Respiratory failure5 69 M 7 8 AD VI N/A6 73 M 4 8 AD N/A N/A7 70 M 3.75 9 AD VI AD8 92 F 3.8 10 AD V Respiratory failure9 74 F 7.5 13 AD N/A N/A

10 77 M 6.17 14 AD IV Pneumonia11 77 M 3.6 16 AD V Respiratory failure12 79 F 4.75 17 AD V N/A13 74 F 2.8 30 Normal II Cancer

Note: PMI, post mortem interval; MMSE, Mini Mental State Examination; N/A, not available.

343Caspase-9 Activation in the AD Brain


tent using the BCA assay (Pierce) to ensure equalprotein loading. Blots were developed using an en-hanced chemiluminescence system (Amersham).

Immunohistochemistry

Free-floating 50-�m-thick serial sections were usedfor immunohistochemical studies as previously de-scribed (Rohn et al., 2001b). For single-labeling exper-iments, sections were incubated overnight in eithercaspase-9 CSSA (5–10 �g/ml) or anti-A�1-16 (6E10,1:1000). For immunostaining with �-amyloid, sectionswere pretreated in 90% formic acid for 4 min prior toincubation in 6E10. For double-labeling experiments,sections were incubated with caspase-9 CSSA (2.5 �g/ml) and either AT8 (1:30,000) or an antibody to theactive 18 kDa fragment of caspase-8 (CASP-8p18,1:200). To determine if cross-reactivity to reagents wasa factor in double-labeling experiments, experimentswere replicated with the antibodies in reverse. Anti-gen visualization was determined using ABC complex(ABC Elite immunoperoxidase kit, Vector labs, Burlin-game, CA) followed by reaction with diaminobenzi-dine (DAB, Vector Labs, Burlingame, CA).

For quantification studies involving cell counts, sec-tions were incubated overnight in caspase-9 CCSA(1:100) and visualized with biotinylated anti-rabbitIgG, avidin–biotin complex and finally with DAB (1 heach incubation at RT). After a series of washes, sec-tions were subsequently incubated in anti-PHF-1 (1:1000; provided by Dr. P. Davies, Albert Einstein Col-lege of Medicine, Bronx, NY) with a blue SG substrate.Both brain regions and all cases were immunostainedin one study to reduce experimental variability. Forquantification involving load values, two sectionswere incubated in either anti-caspase-9 CCSA or in6E10 and similar areas were outlined and captured forimage analysis. A load analysis was used for deter-mining the relationship between active caspase-9CSSA and �-amyloid because several cases showedextensive diffuse plaque formation leading to lowplaque counts that did not accurately represent theextent of �-amyloid accumulation. Thus loads wereobtained for both caspase-9 activation and �-amyloidto provide numbers that were comparable for the twodifferent markers.

Immunofluorescence and Confocal Microscopy

Immunofluorescence studies were performed by in-cubating sections with caspase-9 CSSA (20 �g/ml)overnight at room temperature, followed by ABC

complex containing a secondary rabbit-HRP antibody(ABC Elite immunoperoxidase). Antigen visualizationwas accomplished using a tyramide signal amplifica-tion kit (Molecular Probes, Eugene, OR) consisting ofAlexa fluor 488-labeled tyramide (Ex/Em � 495/519).Caspase-9 CSSA labeled sections were further incu-bated with anti-fodrin caspase-cleavage product(CCP) antibody overnight at room temperature (0.4 to0.5 �g/ml). This antibody specifically recognizescaspase-3 cleavage products of fodrin but not full-length fodrin (Rohn et al., 2001b). A secondary biotin-ylated antibody was incubated for 1 h followed bystreptavidin Cy-3 (1:200) for an additional hour andobserved under fluorescent optics or with a confocalmicroscope. Confocal images were collected on anOlympus IX70 inverted microscope using both a 20�and 40� objective for image analysis and barrier filtersat 510 and 605 nm. A z-series at 1-�m intervals wascaptured to determine the spatial colocalization char-acteristics of staining within individual neurons. Todetermine whether cross-reactivity to reagents was afactor in double-labeling experiments, experimentswere replicated in reverse. The order of primary anti-body application did not affect the results.

Quantification

Cell counts using double-labeled sections includedfive brain regions: CA3, hilus, the average of two areasof CA1, superficial/deep layers of the parahippocam-pal cortex, and superficial/deep layers of frontal cor-tex. Photographs of double-labeled regions for eachcase were taken at 20� and two independent countsby EH and RR was obtained while blind with respectto disease severity. The number of caspase-9 CCSA-only, PHF-1-only, and colocalized cells was counted.Reliability correlations between the two independentcounts were significant and ranged between 0.805 and0.997. The average of the two counts was used as rawdata and analyzed using SPSS for Windows with analpha of 0.05. Load values for single labels of eitheractive-caspase 9 CSSA or A�1-16 were obtained usingpreviously described methods (Cummings et al., 1996).

RESULTS

Caspase-9 Activation in the AD Brain

To examine a possible role for the activation ofcaspase-9 in the AD brain, hippocampal sections fromAD or aged-matched control brains were analyzed by

344 Rohn et al.


immunohistochemistry. Little immunoreactivity wasseen following incubation of caspase-9 CSSA in sec-tions from age-matched control brains (Fig. 1A). Of theseven control cases examined, one case (64 years, PMI� 8 h, myocardial infarction) showed extensive neu-ronal staining within the hippocampus (data notshown). In contrast, extensive labeling of neurons inthe hippocampus (CA1, CA2, and subiculum) andentorhinal cortex was observed in all tissue sectionsexamined from severe AD brains (Fig. 1B, C, and D).In addition, swollen dystrophic neurites withinplaque-rich regions was also observed following incu-bation with caspase-9 CSSA (Fig. 1D arrow and Fig.2D). In a previous study, we demonstrated the activa-tion of caspase-8 in NFTs, as well as in reactive astro-

cytes of the AD brain (Rohn et al., 2001a). In thepresent study, there was no evidence for the activationof caspase-9 within reactive astrocytes in any of theAD cases examined (data not shown).

Caspase-9 Activation Colocalized with AT8 andAdditional Markers of Caspase Activation

Neuronal staining by caspase-9 CSSA in the ADbrain appeared to be within neurons containing NFTs.To examine a possible relationship between caspase-9activation and NFTs, double-labeling experimentswere performed using the NFT marker, AT8. Colocal-ization of these two antibodies was evident in neuronswith clear NFT morphology (Fig. 2A).

FIG. 1. Caspase-9 CSSA immunoreactivity in Alzheimer’s disease. (A) Representative immunohistochemical staining with caspase-9 CSSAin a control case showing absence of staining. (B) Low magnification (4x) of a brain section from a representative AD case depicting labelingof neurons in the hippocampus. (C) High magnification (60x) of a brain section from a representative AD case showing labeling of two neuronswith apparent tangle morphology. (D) Immunolabeling of caspase-9 CSSA in a representative AD case demonstrating labeling of apparentNFTs and neurites within a plaque-rich region (arrow).



To examine a possible relationship between the ac-tivation of caspase-8 and caspase-9, experiments wereperformed using caspase-9 CSSA and CASP-8p18, anantibody that specifically recognizes the activated 18kDa fragment of caspase-8 (Rohn et al., 2001a). Colo-calization of these two antibodies was seen in sectionsfrom the AD brain (Fig. 2B, arrows and inset), sup-porting the hypothesis that activation of bothcaspase-8 and -9 occurs within the same neurons ofthe AD brain. In addition, the colocalization of thesetwo markers appeared to be delegated to neuronsexhibiting a “flame-like” morphology consistent withNFTs (Fig. 2B).

We have previously demonstrated that fodrin CCPantibody is a specific marker for caspase-3 cleavagefragments for fodrin, a neuronal cytoskeleton protein(Rohn et al., 2001b). Initial attempts to perform double-labeling immunohistochemical analysis using brownDAB and blue SG staining proved difficult due to alack of color separation in neurons that were double-labeled (data not shown). Therefore, immunofluores-cence double-labeling experiments were undertakenfollowed by confocal analysis. In this case, clear dem-onstration of colocalization of activated caspase-9 withcaspase-3 products of fodrin within neurons of the ADbrain was observed (Fig. 2C). In addition, colocaliza-tion of these two markers was observed within dys-trophic neurites within plaque-rich regions in the ADbrain (Fig. 2D). Thus, not only is activated caspase-9localized together with activated caspase-8, but is alsofound in close association with stable CCPs of fodrinformed following cleavage by caspase-3. It should benoted that no pretreatment to abolish autofluores-cence was necessary to observe fodrin CCP or acti-vated caspase-9 immunofluorescence as the stainingpatterns of both these antibodies could be clearly dif-ferentiated from the characteristic morphological fea-tures associated with autofluorescence.

Caspase-9 Activation, DNA/RNA OxidativeDamage and Mitochondrial Dysfunction

Because caspase-9 is a key member of the mitochon-drial pathway of apoptosis, we examined whether it is

FIG. 2. Colocalization of the caspase-9 CSSA with a NFT markerand antibodies to the active fragments of caspase-8 as well ascaspase-3 cleavage products of fodrin. (A) Double-labeling immu-nohistochemical analysis for caspase-9 CSSA and AT8 demonstratecolocalization of active caspase-9 (brown) and AT8 (blue) in hip-pocampal neurons from the AD brain. (B) Double-labeling immu-nohistochemical analysis for caspase-9 CSSA and CASP-8p18 anti-body demonstrate colocalization of active caspase-9 (brown) andactive caspase-8 (blue) in hippocampal neurons. Inset representscolocalization of active caspase-8 and -9 in a hippocampal brainsection from an additional AD case. (C) Confocal fluorescent mi-croscopy with the fodrin CCP antibody (red) and caspase-9 CSSA(green) in a severe AD case illustrating the high degree of colocal-ization between these two markers. (D) Identical to panel C, withfodrin CCP antibody (red) and caspase-9 CSSA (green) colocalizing

within dystrophic neurites within a cored plaque of a severe ADcase. (E and F) Double-labeling immunohistochemical analysisfor caspase-9 CSSA (blue) and 8oxodG (brown) in two separateAD cases demonstrates colocalization of active caspase-9 withmarkers for oxidative DNA/RNA damage. Scale bars are equivalentto 50 �m.

346 Rohn et al.


active in neurons exhibiting oxidative damage. Todetermine this, we performed colocalization experi-ments with caspase-9 CSSA and anti-8oxodG, an an-tibody that detects oxidative damage to either DNA(nuclear or mitochondrial) or RNA. In previous stud-ies, 8oxodG was observed in association with dysfunc-tional mitochondria, which may serve as a stimulusfor the activation of caspase-9 (Aliev et al., 2002). Asshown in Fig. 2E and F, the activation of caspase-9occurred in neurons positive for oxidized DNA/RNAin the AD brain. Whereas 8oxodG staining appearedto be punctate within the cell body, the majority ofcaspase-9 CSSA labeling appeared to be more concen-trated in the apical dendrites.

Caspase-9 Activation, NFT and Senile PlaqueFormation in the AD Brain

A consistent feature of caspase-9 CSSA labeling inthe AD brain was immunolabeling of neurons with“flame-like” inclusions, consistent with tangle forma-tion. Therefore, to examine in further detail a possiblerelationship between caspase-9 activation and NFTs,quantitative analysis of 13 cases with a range of MMSEscores between 0 and 30 was performed usingcaspase-9 CSSA and PHF-1, a marker for matureNFTs. The number of cells containing both caspase-9CCSA and PHF-1 was relatively small and rangedfrom 0 to a maximum of 18.6% across all the brainregions sampled. In a control case with a MMSE of 30used for the quantification study, no tangles wereobserved but a significant number of caspase-9 CCSApositive neurons were identified, suggesting thatcaspase-9 activation can occur in the absence of overttangle formation. Further, a significant negative corre-lation was observed for the number of PHF-1 positiveneurons and those positive for caspase-9 CCSA in theparahippocampal cortex (r � �0.75, P � .053) and thefrontal cortex (r � �0.688, P � .013), and for area CA1(r � �0.94, P � .002) (Fig. 3, top) but not in the hilus(r � 0.42, P � n.s.) or in the CA3 subregion (r � 0.40,P � n.s.) (data not shown). Thus, as tangle formationincreased, the extent of caspase-9 activation decreased.Figure 3 (lower panel) shows representative stainingof caspase-9 CSSA and PHF-1 in area CA1 and thefrontal cortex in cases with increasing numbers ofNFTs (left to right) illustrating the inverse relationshipbetween caspase-9 activation and NFT formation inthe AD brain. No significant correlation was observedbetween caspase-9 CSSA and MMSE, age at death, orpost mortem interval.

To determine the relationship between activecaspase-9 CCSA and �-amyloid deposition, loadswere obtained from single label experiments for eachantibody. No significant correlations were found be-tween these two variables (data not shown). Thus,more extensive �-amyloid was not associated withhigher levels of caspase-9 activation.

Synthesis and Characterization of a Caspase-Cleavage Site-Directed Antibody to theMicrotubule-Associated Protein, tau

Based upon the previous results indicating a rela-tionship between the activation of the enzymecaspase-9 and NFT formation, experiments were un-dertaken to examine if tau, whose modification leadsto PHFs and NFTs in AD, is cleaved by caspases andwhether fragments accumulate in the AD brain. Todetermine if tau is cleaved by caspases, in vivo, wedesigned a caspase-cleavage site-directed antibody totau based upon a known consensus caspase-cleavagesite within the tau sequence located toward the amino-terminal end, DRKD-QGGYTMHQDQ (Canu et al.,1998). Following purification of this antibody (hereintermed tauCCP antibody) by affinity chromatography,its validity as a specific probe for tau CCPs was ex-amined using a cell-free system. Human brain or pri-mary cultured rat cortical neuron extracts were incu-bated with or without caspase-3. Following digestionwith caspase-3, samples were immunoblotted with thetau-CCP antibody to determine if recognition of CCPsoccurred. Two prominent bands were immunolabeledwith the tau-CCP antibody in digested extracts corre-sponding to 67 and 28 kDa, that were not presentunder control conditions (Fig. 4A, left panel). Underthese conditions, the tau CCP antibody did not labelfull-length tau in control lanes, which normally has amolecular weight of 70 kDa. To verify that full-lengthtau was indeed present in these extracts the same blotwas stripped and reprobed with an anti-tau antibodythat recognizes both full-length tau and its associatedcaspase-cleaved fragments. Numerous bands corre-sponding to full-length tau and CCPs were detectedfollowing western blot analysis with this antibody(Fig. 4A, right panel). These initial results suggest thatthe tau-CCP antibody recognizes CCPs of tau, butdoes not recognize full-length tau.

To further characterize the tau-CCP antibody, ex-periments were undertaken using differentiated PC12cells. To verify the presence of tau in PC12 cells, ex-tracts from both nondifferentiated and differentiated



PC12 cells were immunoblotted with the anti-tau an-tibody that recognizes full-length tau. Figure 4B (leftpanel) depicts the results of such an experiment show-ing the expression of tau following differentiation ofPC12 cells in the presence of NGF. The full-lengthantibody also recognized numerous CCPs of tau fol-lowing digestion with caspase-3 (Fig. 4B, left panel). Inseparate experiments, NGF-treated PC12 cells wereincubated in the presence or absence of caspase-3, andprobed with the tau-CCP antibody. Although the tau-CCP antibody recognized several caspase-cleavedfragments including a prominent band running at 28kDa, it did not recognize full-length tau in nondi-gested extracts (Fig. 4B, right panel).

Tau-CCP Antibody as a Marker for CaspaseActivation in the AD Brain

Following validation of the tau-CCP antibody as aspecific probe for the caspase-cleavage of tau in vitro,immunohistochemistry experiments were performedusing hippocampal sections from AD or age-matchedcontrol brains. In all control sections examined, littlestaining was observed following application of thetau-CCP antibody (Fig. 4C). In contrast, widespreadlabeling of NFTs, neuropil threads, and dystrophicneurites within plaque-rich regions was observed fol-lowing immunohistochemical analysis with the tau-CCP antibody in all severe AD brain sections exam-

FIG. 3. Negative correlation between caspase-9 activation and PHF-1 immunolabeling. Significant negative correlations were observed for thenumber of PHF-1 positive neurons and those positive for caspase-9 CCSA in the frontal cortex (r � �0.69, P � .013) (A) and for area CA1 (r ��0.94, P � .002) (B) Caspase-9 activation decreases with increasing neurofibrillary tangle accumulation in area CA1 of the hippocampus (C–F)and frontal cortex (G–J). Cases were selected with increasing tangle-bearing neuron counts (lowest to highest, left–right) to illustrate thatcaspase-9 activation (brown) also appears to precede tangle formation (blue) and decline as the extent of tangle formation increases. Note,however, that in a severe AD case (F and J), extensive cell loss may also account for loss of active-caspase-9 containing neurons.

348 Rohn et al.


FIG. 4. Caspase cleavage of tau in the AD brain. (A) Brain extracts from human or primary cultured rat cortical neurons were digested with1.88 �g of caspase-3 overnight at RT. Proteins were separated by a 12% SDS-PAGE gel, transferred to nitrocellulose, and probed with 0.5–1.0�g/ml of purified tau-CCP antibody (left panel) or with a commercial antibody (1:100) that recognizes both full-length tau (70 kDa), as wellas any associated cleaved fragments (right panel). Unlike the commercial antibody, which recognized full-length tau and caspase-cleavedfragments, the tau-CCP antibody labeled two prominent caspase-3-cleaved fragments without labeling full-length tau corresponding to 70 kDa.(B) NGF-treated PC12 cell extracts were prepared, digested with caspase-3, and analyzed with either a commercial antibody to tau (left panel)or with the tau-CCP antibody (right panel) as described above. The tau-CCP antibody recognized caspase-3 digested fragments but did notrecognize full-length tau. (C) Representative immunohistochemical staining with anti-tau CCP purified antibody in a control case showing lackof immunoreactivity. (D) High-field magnification (40X) of a hippocampal section from a severe AD case depicting tau-CCP immunoreactivitywithin NFTs, neuropil threads, and dystrophic neurites within plaque regions. Inset (D) indicates the granular nature of tau-CCP immuno-reactivity (arrowheads). (E and F) Serial hippocampal tissue sections from a severe AD case showing extensive neuronal labeling usingtau-CCP, which is absent after staining with preabsorbed antibody (panel F).



ined (Fig. 4D). Interestingly, numerous tau-CCP-pos-itive neurons displayed staining that was granular innature, and appeared to be membrane-bound (inset,arrowheads, Fig. 4D). This type of staining has alsobeen observed in a previous study demonstrating theactivation of caspase-3 in the AD brain (Stadelmann etal., 1999). Importantly, staining with the tau-CCP an-tibody was completely prevented following preab-sorption with free peptide (Fig. 4E and F), illustratingthe specificity of the tau-CCP antibody. In addition,we also tested purified tau-CCP antibody derivedfrom a second rabbit immunized with the same ami-no-terminal peptide and obtained identical results fol-lowing immunohistochemistry (data not shown).

DISCUSSION

Several recent studies have reported increased lev-els of caspase-3-cleaved substrates, including APP, fo-drin, and actin in the AD brain (Yang et al., 1998;Gervais et al., 1999; Rohn et al., 2001b). All of thesestudies relied on the use of site-directed caspase-cleav-age antibodies, which take advantage of the fact thatcaspases are specific, cleaving after aspartic residuesthereby generating caspase-cleavage products (CCPs)that are antigenically distinct. Using this approach, wehave synthesized antibodies against one of the activefragments of caspase-8, compared these results with acommercially available active-caspase-8 antibody, anddemonstrated colocalization of this antibody togetherwith caspase-3 cleaved products of fodrin in the ADbrain (Rohn et al., 2001a). The goal of the present studywas to further dissect out the pathways of apoptosisactivated in the AD brain. Presently, there are twomajor pathways leading to apoptosis: the death-recep-tor pathway in which caspase-8 plays a pivotal initia-tor role, and the mitochondrial pathway (or chemicalpathway) where caspase-9 is the key initiator caspase(Ashkenazi and Dixit, 1999). Both pathways convergeand activate the executioner caspase, caspase-3.

Data presented in the present manuscript suggests arole for the activation caspase-9 in the AD brain. Theseresults are supported by a recent study demonstratingan immunoreactive band consistent with cleavedcaspase-9 in synaptosomal preparations from five offive AD brains but only in one of five control brainpreparations using caspase-9 CSSA (Lu et al., 2000).Taken together, these results implicate the mitochon-drial pathway of apoptosis as a mechanism leading toactivation of caspase-3 in the AD brain. Furthermore,the observation of caspase-9 activation in one of five

control cases in the Lu et al., study (2000) is consistentwith our observations of a subset of control casesshowing active caspase-9 immunoreactivity.

A Link between Oxidative Damage and PathwaysAssociated with Apoptosis in AD

Oxidative damage is a key feature of normal agingand is more extensive in the AD brain (Smith et al.,2000). It is possible that neurons accumulating exten-sive oxidative insults may selectively undergo apopto-sis (Lu et al., 2000). Mitochondria are particularly vul-nerable to oxidative damage because they are the ma-jor source of oxidants produced through normalmetabolic processes. Mitochondrial dysfunction hasbeen observed in AD brain (Hirai et al., 2001), and asa consequence, stimulation of the mitochondrial path-way of apoptosis may be elicited through oxidativedamage leading to the release of cytochrome c. Re-leased cytochrome c interacts with Apaf-1 and thiscomplex recruits caspase-9 leading to its activation(Kuida, 2000). Studies have shown that if caspase-9 isinactivated or is not expressed, full activation of themitochondrial pathway does not occur (Kuida, 2000).For example, caspase-9 null mice are deficient in acti-vation of caspase-3 in vivo and cytochrome c-mediatedcleavage of caspase-3 is absent in cytosolic extracts butcan be restored after reconstitution with in vitro-trans-lated caspase-9 (Kuida et al., 1998). Thus, the observa-tion of neurons containing both oxidatively damagedDNA/RNA, most likely associated with mitochon-drial dysfunction, and the activation of caspase-9 sug-gests a link between these two events in vivo. Furtherexperiments will help clarify the strength of the asso-ciation between these two markers of cell dysfunction.

Stimulation of Apoptotic Pathways by MultipleInsults May Ultimately Lead to the Activationof Caspase-3

Another outcome of the present study was the dem-onstration of colocalization of activated caspase-9 andcaspase-8 within individual neurons in the AD brain.This suggests the possibility of concurrent activationof both major pathways of apoptosis within the sameneuron possibly by different stimuli. Caspase-8 hasbeen implicated in Fas/CD95 or tumor necrosis factor(TNF) receptor cell death program. In this regard,caspase-8 is thought to be the most apical member ofthe family of caspases that is recruited by adapterproteins (e.g., Fas-associated death domain, FADD)and converted to an active form by auto proteolysis

350 Rohn et al.


(Muzio et al., 1996). Cleavage of caspase-8 results intwo active fragments of 11 and 18 kDa, both of whichrepresent the activated form of the enzyme. Caspase-8in turn, is thought to activate downstream caspases,particularly caspase-3 (Takahashi, 1999).

What mechanism occurring in the AD brain couldpossibly lead to the activation of both pathways ofapoptosis within the same neuron? In a previousstudy, we presented a model proposing that �-amy-loid may induce neuronal cell death by inducing ap-optosis following cross-linking of death-receptors andconcomitant activation of caspase-8 and caspase-3(Rohn et al., 2001a). Several in vitro studies support thisas a possible mechanism for �-amyloid-induced neu-rotoxicity (Ivins et al., 1999a,b). In addition to thepossible activation of death receptors by a cross-link-ing mechanism, evidence suggests that �-amyloidmay also induce significant oxidative stress withinneurons by the formation of reactive oxygen species(Miranda et al., 2000; Varadarajan et al., 2000). Theformation of reactive oxygen species and subsequentoxidative damage may lead to the stimulation of themitochondrial pathway of apoptosis. Evidence givenin the present report, indicating the activation ofcaspase-9 and caspase-8, suggests that activation ofboth pathways of apoptosis may occur within neuronsin the AD brain.

However, the current study indicates a closer asso-ciation between NFT formation rather than �-amyloiddeposition and the extent of caspase-9 activation. Thesimplest interpretation is that �-amyloid does notserve as a stimulus for the activation of either caspase.This may, however, be a difficult hypothesis to testusing post mortem tissue because �-amyloid is a long-lived protein and will be present in the brain at muchlonger time intervals than would be expected for theactivity of any single enzyme. Experiments using neu-ronal cultures and in vitro methodologies would bebetter able to address this question directly.

An Association between Caspase Activation andtau Pathology?

Regardless of the exact pathway of apoptosis stim-ulated, the end result will be the activation of execu-tioner caspases, such as caspase-3, and the cleavage ofcritical cellular proteins. One susceptible protein thatmay be a target for caspase-3 cleavage relevant to ADis the microtubule-associated protein, tau. In AD, taubecomes hyperphosphorylated leading to the forma-tion of NFTs and destabilization of the cytoskeletonnetwork (Bramblett et al., 1993). NFTs consist of paired

helical filaments (PHFs) resulting from the hyperphos-phorylation of tau in select neuronal populations. Inthe brains of patients with AD, the cytoskeleton isgradually damaged and replaced by bundles of PHFs,which are largely composed of abnormally processedforms of tau (Wischik et al., 1988; Alonso et al., 1996).Until recently, it was thought that an abnormal phos-phorylation of tau proteins was responsible for theiraggregation in AD. However, normal tau proteins arealso phosphorylated in fetal and adult brain, and theydo not aggregate to form filamentous inclusions (Bueeet al., 2000). Moreover, nonphosphorylated recombi-nant tau proteins form filamentous structures underphysiological conditions in vitro (Buee et al., 2000). Inlight of these observations, it has been proposed thatmodifications of tau independent of phosphorylationmay also be involved in the formation of pathologicaltau filaments.

One recurrent theme demonstrated in the presentstudy along with our previous studies is activation ofcaspases within tangle-bearing neurons (Rohn et al.,2001a,b; Su et al., 2001). Thus, we have previouslydemonstrated a strong positive correlation (R � 0.84)between the presence of CCPs of fodrin and NFTs(Rohn et al., 2001b). In the present study we actuallyobserved a negative correlation between the activationof caspase-9 and NFTs in the CA1, parahippocampalregions, and the frontal cortex of the AD brain. Thereare a number of possible interpretations of this result:(1) the stimuli to drive mitochondrial-dependentcaspase-9 activation are no longer present in tanglebearing neurons; (2) caspase-9 activation is indepen-dent of tangle formation; (3) caspase-9 activation pre-cedes tangle formation. We favor the last interpreta-tion because it is clear that the activation of caspase-9and tangle formation is related based on the resultsfollowing quantitative analysis. These results are con-sistent with the appearance of activated caspase-9,followed by increasing numbers of neurons develop-ing tangles.

The negative correlation in the present study be-tween caspase-9 activation and NFTs and the positivecorrelation between caspase cleavage of fodrin andNFTs in our previous report (Rohn et al., 2001b) wouldseem to be contradictory. However, it is important torealize the distinction between the two antibodiesused for each of these studies. The fodrin CCP anti-body recognizes cleavage products of fodrin that arestable, which accumulate over time and are presenteven when the neuron dies. It seems plausible thatonly when sufficient amounts of these CCPs of fodrinaccumulate, does the fodrin CCP antibody give a



strong signal following immunohistochemical analy-sis. We hypothesize that the accumulation of CCPs offodrin may take many years to develop and this iswhy it is so strongly correlated with PHF-1 immuno-labeling. On the other hand, an antibody against theactive fragment of caspase-9 would be expected toprovide a maximum signal as soon as it was activatedin the AD brain, and might actually be attenuated inlate stages of AD as the neurons die.

Based on these results and other evidence in theliterature, it is possible that the activation of caspaseswithin neurons in the AD brain may lead to tau cleav-age and its modification may contribute to the forma-tion of PHFs and NFT formation. In order for this to betrue, it must be first demonstrated that tau is cleavedby caspases in vivo. Data in the present study indicatethe labeling of neurons with a site-directed caspase-cleavage antibody to tau in the AD brain, supportingthe conclusion that tau is cleaved by caspases in theAD brain. The implications of these data are thatcaspases play an earlier role in the pathology associ-ated with AD then originally described and do notrepresent simply end stage processes associated withneuronal cell death. Therefore, pharmacological inter-vention using caspase inhibitors may be a potentialbeneficial therapeutic target for the treatment of AD.

In some instances, staining with the tau-CCP anti-body revealed labeling of neurons that was granularand vesicular in nature (Fig. 4D, inset). Previous stud-ies have shown that tau is not only a substrate forcaspase cleavage, but in turn, this produces a taufragment that is an effecter of apoptosis generating apositive-feedback loop that is self-propagating (Fasuloet al., 2000; Chung et al., 2001). Therefore, the compart-mentalization of caspase-cleaved fragments of tauwould suggest that these fragments are somehowtoxic to the neuron and are being segregated as apossible means of protection. These results are sup-ported by a previous study, which demonstrated theactivation of caspase-3 in autophagic granules in sin-gle neurons in the AD brain (Stadelmann et al., 1999).The authors concluded that activation of caspases inneurons is counteracted by the seclusion of damagedareas into autophagic vacuoles (Stadelmann et al.,1999).

A Hypothetical Sequence of Events Linking�-Amyloid Deposition and NFT Formationin the AD Brain

We propose that individual neurons within the ADbrain may have a number of caspases that have been

activated by several possible stimuli (Fig. 5). The dep-osition of �-amyloid may lead to caspase-8 activationthrough cross-linking of death receptors such as Fas.In parallel, oxidative damage accumulating with bothage and with disease may lead to mitochondrial dys-function, the release of cytochrome c, and finally,

FIG. 5. A hypothetical sequence of events leading to caspase acti-vation and possibly tau cleavage. Two parallel pathways, activatedby different stimuli, may operate within individual neurons. Theaccumulation of �-amyloid may lead to the cross-linking and acti-vation of receptors resulting in caspase-8 activation. An additionalmechanism is that �-amyloid may modify lipids by oxidation in thecell membrane that may also increase or exacerbate existing oxida-tive damage within neurons. Mitochondrial dysfunction is also aconsequence of oxidative damage that can be detected by oxidativemodifications to DNA/RNA, which may accumulate within neu-rons and in turn lead to further mitochondrial dysfunction. Cyto-chrome c is released by dysfunctional mitochondria and activatescaspase-9. Both pathways may then converge by activating theeffecter enzyme, caspase-3. Tangle formation may be occurringindependently or as a consequence of caspase-3 activation andcleavage of tau. Asterisks indicate markers used in the currentstudy.

352 Rohn et al.


caspase-9 activation. Thus, it is possible that the cu-mulative events of �-amyloid accumulation and oxi-dative damage may converge and become additivewhen levels of caspase-3 rise. In this model, tangleformation would be a relatively late event and is moreclosely associated in time with caspase-3 activationand the accumulation of caspase-3 cleavage products,such as fodrin. Caspase cleavage of proteins that areconstituents of the cytoskeleton, such as fodrin andtau, may result in the disassembly of the cytoskeleton,leading to disruption of synaptic vesicle transport aswell as altering cell morphology.

In conclusion, we have demonstrated the activationof caspase-9 together with caspase-8 and products ofcaspase-3 cleavage in the AD brain. The colocalizationof caspase-9 together with caspase-8 within individualneurons would suggest the activation of multiplepathways of apoptosis within neurons in the ADbrain. It is possible that �-amyloid may initiate theactivation of both apoptotic pathways through eithercross-linking of death receptors or through the gener-ation of oxidative damage. The activation of eitherpathway within neurons of the AD brain may lead tothe concomitant activation of caspase-3 and subse-quent cleavage of cellular proteins, including tau.Caspase-cleavage of tau may induce modifications oftau that facilitate the formation of PHFs and ulti-mately NFTs. However, the exact role of caspasecleavage of tau, in vivo, and the formation of NFTs willrequire additional studies.

ACKNOWLEDGMENTS

This work was supported in part by NIA grants; R15AG19642-01and R03AG19386-01 (T.T. Rohn), a grant from the Aging Federationof Aging Research (T.T. Rohn), and NIA grant P50 AG16573 toCWC.

REFERENCES

Aliev, G., Smith, M. A., Seyidov, D., Neal, M. L., Lamb, B. T.,Nunomura, A., Gasimov, E. K., Vinters, H. V., Perry, G., La-Manna, J. C., & Friedland, R. P. (2002) The role of oxidative stressin the pathophysiology of cerebrovascular lesions in Alzheimer’sdisease. Brain Pathol. 12, 21–35.

Alonso, A. C., Grundke-Iqbal, I., & Iqbal, K. (1996) Alzheimer’sdisease hyperphosphorylated tau sequesters normal tau into tan-gles of filaments and disassembles microtubules. Nat. Med. 2,783–787.

Ashkenazi, A., & Dixit, V. M. (1999) Apoptosis control by death anddecoy receptors. Curr. Opin. Cell Biol. 11, 255–260.

Bramblett, G. T., Goedert, M., Jakes, R., Merrick, S. E., Trojanowski,J. Q., & Lee, V. M. (1993) Abnormal tau phosphorylation at Ser396in Alzheimer’s disease recapitulates development and contributesto reduced microtubule binding. Neuron 10, 1089–1099.

Buee, L., Bussiere, T., Buee-Scherrer, V., Delacourte, A., & Hof, P. R.(2000) Tau protein isoforms, phosphorylation and role in neuro-degenerative disorders. Brain Res. Brain Res. Rev. 33, 95–130.

Canu, N., Dus, L., Barbato, C., Ciotti, M. T., Brancolini, C., Rinaldi,A. M., Novak, M., Cattaneo, A., Bradbury, A., & Calissano, P.(1998) Tau cleavage and dephosphorylation in cerebellar granuleneurons undergoing apoptosis. J. Neurosci. 18, 7061–7074.

Chung, C. W., Song, Y. H., Kim, I. K., Yoon, W. J., Ryu, B. R., Jo,D. G., Woo, H. N., Kwon, Y. K., Kim, H. H., Gwag, B. J., Mook-Jung, I. H., & Jung, Y. K. (2001) Proapoptotic effects of tau cleav-age product generated by caspase-3. Neurobiol. Dis. 8, 162–172.

Cummings, B. J., Head, E., Afagh, A. J., Milgram, N. W., & Cotman,C. W. (1996) Beta-amyloid accumulation correlates with cognitivedysfunction in the aged canine. Neurobiol. Learn. Mem. 66, 11–23.

Fasulo, L., Ugolini, G., Visintin, M., Bradbury, A., Brancolini, C.,Verzillo, V., Novak, M., & Cattaneo, A. (2000) The neuronalmicrotubule-associated protein tau is a substrate for caspase-3and an effector of apoptosis. J. Neurochem. 75, 624–633.

Gervais, F. G., Xu, D., Robertson, G. S., Vaillancourt, J. P., Zhu, Y.,Huang, J., LeBlanc, A., Smith, D., Rigby, M., Shearman, M. S.,Clarke, E. E., Zheng, H., Van Der Ploeg, L. H., Ruffolo, S. C.,Thornberry, N. A., Xanthoudakis, S., Zamboni, R. J., Roy, S., &Nicholson, D. W. (1999) Involvement of caspases in proteolyticcleavage of Alzheimer’s amyloid-beta precursor protein and amy-loidogenic A beta peptide formation. Cell 97, 395–406.

Hirai, K., Aliev, G., Nunomura, A., Fujioka, H., Russell, R. L.,Atwood, C. S., Johnson, A. B., Kress, Y., Vinters, H. V., Tabaton,M., Shimohama, S., Cash, A. D., Siedlak, S. L., Harris, P. L., Jones,P. K., Petersen, R. B., Perry, G., & Smith, M. A. (2001) Mitochon-drial abnormalities in Alzheimer’s disease. J. Neurosci. 21, 3017–3023.

Ivins, K. J., Ivins, J. K., Sharp, J. P., & Cotman, C. W. (1999a) Multiplepathways of apoptosis in PC12 cells. CrmA inhibits apoptosisinduced by beta-amyloid. J. Biol. Chem. 274, 2107–2112.

Ivins, K. J., Thornton, P. L., Rohn, T. T., & Cotman, C. W. (1999b)Neuronal apoptosis induced by beta-amyloid is mediated bycaspase-8. Neurobiol. Dis. 6, 440–449.

Kerr, J. F., Wyllie, A. H., & Currie, A. R. (1972) Apoptosis: A basicbiological phenomenon with wide-ranging implications in tissuekinetics. Br. J. Cancer 26, 239–257.

Kuida, K. (2000) Caspase-9. Int. J. Biochem. Cell Biol. 32, 121–124.Kuida, K., Haydar, T. F., Kuan, C. Y., Gu, Y., Taya, C., Karasuyama,

H., Su, M. S., Rakic, P., & Flavell, R. A. (1998) Reduced apoptosisand cytochrome c-mediated caspase activation in mice lackingcaspase 9. Cell 94, 325–337.

Lu, D. C., Rabizadeh, S., Chandra, S., Shayya, R. F., Ellerby, L. M.,Ye, X., Salvesen, G. S., Koo, E. H., & Bredesen, D. E. (2000) Asecond cytotoxic proteolytic peptide derived from amyloid beta-protein precursor. Nat. Med. 6, 397–404.

Marx, J. (2001) Neuroscience. New leads on the ‘how’ of Alzhei-mer’s. Science 293, 2192–2194.

Mattson, M. P., Pedersen, W. A., Duan, W., Culmsee, C., & Caman-dola, S. (1999) Cellular and molecular mechanisms underlyingperturbed energy metabolism and neuronal degeneration in Alz-heimer’s and Parkinson’s diseases. Ann. NY Acad. Sci. 893, 154–175.

Miranda, S., Opazo, C., Larrondo, L. F., Munoz, F. J., Ruiz, F.,Leighton, F., & Inestrosa, N. C. (2000) The role of oxidative stress



in the toxicity induced by amyloid beta- peptide in Alzheimer’sdisease. Prog. Neurobiol. 62, 633–648.

Muzio, M., Chinnaiyan, A. M., Kischkel, F. C., O’Rourke, K.,Shevchenko, A., Ni, J., Scaffidi, C., Bretz, J. D., Zhang, M., Gentz,R., Mann, M., Krammer, P. H., Peter, M. E., & Dixit, V. M. (1996)FLICE, a novel FADD-homologous ICE/CED-3-like protease, isrecruited to the CD95 (Fas/APO-1) death-inducing signalingcomplex. Cell 85, 817–827.

Nunomura, A., Perry, G., Aliev, G., Hirai, K., Takeda, A., Balraj,E. K., Jones, P. K., Ghanbari, H., Wataya, T., Shimohama, S.,Chiba, S., Atwood, C. S., Petersen, R. B., & Smith, M. A. (2001)Oxidative damage is the earliest event in Alzheimer disease.J. Neuropathol. Exp. Neurol. 60, 759–767.

Rohn, T. T., Head, E., Nesse, W. H., Cotman, C. W., & Cribbs, D. H.(2001a) Activation of caspase-8 in the Alzheimer’s disease brain.Neurobiol. Dis. 8, 1006–1016.

Rohn, T. T., Head, E., Su, J. H., Anderson, A. J., Bahr, B. A., Cotman,C. W., & Cribbs, D. H. (2001b) Correlation between caspase acti-vation and neurofibrillary tangle formation in Alzheimer’s dis-ease. Am. J. Pathol. 158, 189–198.

Rohn, T. T., Ivins, K. J., Bahr, B. A., Cotman, C. W., & Cribbs, D. H.(2000) A monoclonal antibody to amyloid precursor protein in-duces neuronal apoptosis. J. Neurochem. 74, 2331–2342.

Smith, M. A., Rottkamp, C. A., Nunomura, A., Raina, A. K., & Perry,G. (2000) Oxidative stress in Alzheimer’s disease. Biochim. Biophys.Acta 1502, 139–144.

Stadelmann, C., Deckwerth, T. L., Srinivasan, A., Bancher, C., Bruck,W., Jellinger, K., & Lassmann, H. (1999) Activation of caspase-3 insingle neurons and autophagic granules of granulovacuolar de-generation in Alzheimer’s disease. Evidence for apoptotic celldeath. Am. J. Pathol. 155, 1459–1466.

Su, J. H., Zhao, M., Anderson, A. J., Srinivasan, A., & Cotman, C. W.(2001) Activated caspase-3 expression in Alzheimer’s and agedcontrol brain: Correlation with Alzheimer pathology. Brain Res.898, 350–357.

Takahashi, A. (1999) Caspase: Executioner and undertaker of apo-ptosis. Int. J. Hematol. 70, 226–232.

Thornberry, N. A., & Lazebnik, Y. (1998) Caspases: Enemies within.Science 281, 1312–1316.

Varadarajan, S., Yatin, S., Aksenova, M., & Butterfield, D. A. (2000)Review: Alzheimer’s amyloid beta-peptide-associated free radicaloxidative stress and neurotoxicity. J. Struct. Biol. 130, 184–208.

Wischik, C. M., Novak, M., Thogersen, H. C., Edwards, P. C.,Runswick, M. J., Jakes, R., Walker, J. E., Milstein, C., Roth, M., &Klug, A. (1988) Isolation of a fragment of tau derived from thecore of the paired helical filament of Alzheimer disease. Proc. Natl.Acad. Sci. USA 85, 4506–4510.

Yang, F., Sun, X., Beech, W., Teter, B., Wu, S., Sigel, J., Vinters, H. V.,Frautschy, S. A., & Cole, G. M. (1998) Antibody to caspase-cleavedactin detects apoptosis in differentiated neuroblastoma andplaque-associated neurons and microglia in Alzheimer’s disease[see comments]. Am. J. Pathol. 152, 379–389.

354 Rohn et al.


Documents

Caspase-9 Activation and Caspase Cleavage of tau in the Alzheimer's Disease Brain