ONCOGENIC PARALLELS IN ALZHEIMER DISEASE

by

ARUN K. RAINA

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Dissertation Advisors

Dr. Mark A. Smith

&

Dr. Xiongwei Zhu

Department of Pathology

Case Western Reserve University

Janurary 2005

Copyright © 2005 by Arun K. Raina All rights reserved

ii CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the dissertation of ARUN K. RAINA

candidate for the Ph.D. degree *.

(signed)___George Perry______

(chair of the committee)

____Mark A. Smith______

____Xiongwei Zhu______

____Robert B. Petersen______

____David Boothman______

(date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein.

iii Dedication

To

Bhairavah

Uma, Autar, Arvind

and

P

iv Table of Contents

List of Tables 7

List of Figures 8

Acknowledgements 11

Abbreviations 12

Abstract 16

Chapter 1 Introduction 18

1.1 Historical 18

1.2 Definition of Dementias 19

1.3 Epidemiology of AD 19

1.4 Pathology of AD 19

a. Neurofibrillary Tangles 20

b. Senile Plaques 22

1.5 Clinical Features of AD 22

1.6 Etiopathogenesis of AD 23

a. Age 23

b. Genetics 24

i. Amyloid β Protien Precursor (AβPP) 24

ii. Presenilins 1 and 2 25

iii. ApoE 25

c. Oxidative Stress 27

d. Cell Cycle and AD 28

1 Chapter 2 Oxidative Stress in AD 30

2.1 Introduction 30

a. Oxidative Stress in AD 30

b. Oxidative Events and Apoptosis in AD 32

c. Oxidative and Cell Cycle Events in AD 33

d. Oxidative Stress - An Adaptation in AD? 34

e. Oxidative Relevance to Lesions 35

f. Neuroprotection Within Vulnerable Neurons in AD 37

2.2 Experimental Hypothesis 37

2.3 Methods and Materials 40

a. Tissue Section Preparation 40

b. Antibodies/Protein Preparations 40

c. Immunocytochemistry 41

2.4 Results 42

a. NQO1 42

b. Bleomycin 42

2.5 Discussion 43

a. NQO1 43

b. Bleomycin Hydrolase 45

2.6 Conclusions 46

2.7 Relevant Publications 53

2 Chapter 3 Cell Cycle Dysfunction in AD 57

3.1 Introduction 57

a. The Cell Cycle and Control Mechanisms 57

b. Reappearance of Cell Cycle Markers in AD 59

c. Association of the Mitotic Phenotype with AD Neuropathology 60

3.2 Experimental Hypothesis 63

3.3 Methods and Materials 63

a. Tissue section preparation 63

b. Immunocytochemistry 64

c. Immunoblots 65

3.4 Results 65

a. Mrg 15 65

b. CARB 66

c. P27 67

3.5 Discussion 69

a. Mrg 15 69

b. CARB 70

c. P27 73

3.6 Conclusions 75

3.7 Relevant Publications 92

3 Chapter 4 Apoptotic Avoidance as a Feature in AD 94

4.1 Introduction 94

a. Programmed Cell Death and Apoptosis - Basic Concepts 94

b. Apoptotic Avoidance in AD 96

4.2 Experimental Hypothesis 98

4.3 Methods and Materials 99

a. Tissue Section Preparation 99

b. Immunocytochemistry 100

c. Antibodies 100

d. Adsorption Experiments 101

e. Immunoblotting and Immunodotting 101

f. Electron Microscopy 103

g. Cells 103

h. Determination of Cell Death 104

i. Transfection and Selection 104

j. RT-PCR 105

4.4 Results 105

a. Upstream 105

b. Downstream Caspases 106

c. Bcl-w 106

4.5 Discussion 110

a. Caspases 110

b. Bcl-w 113

4 4.6 Conclusions 116

4.7 Relevant Publications 125

Chapter 5 Oncogenic Parallels in AD and Conclusions 127

5.1 Introduction 127

a. Oxidative Stress, AD and Oncogenesis 127

b. Dysmitotic Mechansisms, AD and Oncogenesis 128

c. Loss of Apoptosis, AD and Oncogenesis 129

5.2 Experimental Hypothesis 131

5.3 Methods and Materials 133

a. Tissue Section Preparation 133

b. Immunocytochemistry 133

c. Immunoblots 135

5.4 Results 135

a. ADAM 135

b. BRCA-1 137

c. Phospho-retinoblastoma (pRb) 138

5.5 Discussion 138

a. ADAM 138

b. BRCA-1 141

c. Phospho-retinoblastoma (pRb) 144

5.6 Overall Conclusions 146

5.7 Future Directions 148

5 a. Apoptotic Avoidance 148

b. Functional Consequences of BRCA-1 Misexpression in Post-

Mitotic Neurons 149

5.8 Relevant Publications 163

Bibliography 164

6 List of Tables

Table 2.1 48 Quantitative comparison of NQO1 immunoreactivity in AD and control cases.

Table 3.1 77 Cell cycle markers that are seen in AD

Table 3.2 78 The AD phenotype resembles a mitotic phenotype.

Table 5.1 150 BRCA-1 (Ab-1) immunoreactive AD cases and respective predispositional factors.

Table 5.2 151 Antisera to Rb phosphorylational sites that immunoreacts with AD neuropathology.

7 List of Figures

Figure 2.1 49 NQO1 in AD.

Figure 2.2 50 Immunocytochemical localization of BH.

Figure 2.3 51 Immunoreactivity of BH-absorption analysis.

Figure 2.4 52 NQO1 and its role in AD.

Figure 3.1 79 Cell cycle control overview.

Figure 3.2A 80 The G1-S checkpoint control.

Figure 3.2B 81 Cell-cycle related proteins: overlapping functions.

Figure 3.3 82 Mrg 15 – Role in cell cycle.

Figure 3.4 83 Mrg 15 immunocytochemistry.

Figure 3.5. 84 Western analysis of Mrg 15.

Figure 3.6 85 Immunocytochemical localization of CARB.

Figure 3.7 86 CARB immunocytochemistry: Absorption studies.

Figure 3.8 87 CARB immunoblot analysis.

Figure 3.9 88 p27 immunocytochemistry.

8 Figure 3.10 89 p27/τ overlap.

Figure 3.11 90 p27 immunocytochemistry: absorption studies.

Figure 3.12 91 p27 quantification.

Figure 4.1 117 An overview of the apoptotic process.

Figure 4.2 118 8 in AD.

Figure 4.3 119 Caspase 3 in AD.

Figure 4.4 120 Bcl-w immunocytochemistry.

Figure 4.5 121 Bcl-w electron microscopy.

Figure 4.6 122 Bcl-w immunoblots.

Figure 4.7 123 Effect of Aβ on Bcl-w.

Figure 4.8 124 Bcl-w transfection.

Figure 5.1 152 Parallels between AD and oncogenesis.

Figure 5.2 153 Immunocytochemistry of ADAMs 1 and 2 in AD

Figure 5.3 154 Immunocytochemistry of ADAMs 1 and 2 in controls

Figure 5.4 155 Immunocytochemistry of ADAMs 1 and 2: Adsorption studies

9 Figure 5.5 156 Immunoblot analysis of ADAMs 1 and 2

Figure 5.6 157 BRCA-1 immunocytochemistry

Figure 5.7 158 BRCA-1 immunoblot analysis

Figure 5.8 159 BRCA-1 is absent from control NFT

Figure 5.9 160 Phospho-retinoblastoma immunocytochemistry

Figure 5.10 161 Effects of BRCA-1 overexpression – Assays

Figure 5.11 162 Effects of BRCA-1 overexpression – JNK/ERK studies

10 Acknowledgments

This work would not be possible if it were not for close support from many

people on a daily basis. I would like to offer my heartfelt appreciation to all of them for

taking the time and giving the effort to make all of this possible. I would like to thank

Mark A. Smith, my mentor, for opening the window of science for me and allowing me

to have fun in science. I would also like to thank Xiongwei Zhu and George Perry for the

constant support in both scientific and nonscientific matters. I have truly enjoyed the

teaching that I was offered. Additionally, I would like to thank Peggy Harris for her

showing me how to do my first experiment and for her constant support over the years. I

would also like to thank Sandi Siedlak for her help throughout these years. Of course

without Beth Kumar and her efficient help in so many areas I would be quite

handicapped. I have truly enjoyed their help, support, and friendship. I have great pleasure in knowing and working with my colleagues, Hyoung-gon Lee, Mike Marlatt,

Gemma Casadesus, Kate Webber, and others that have worked in the lab. I am also indebted to my committee members, David Boothman and Robert B. Petersen. Finally, I

would especially like to thank Catherine A. Rottkamp for her immense help throughout

my studies and for being a consummate friend.

11 Abbreviations

8OHG 8-hydroxyguanosine

AD Alzheimer disease

ADRDA Alzheimer disease and Related Disorders Association

AGE Advanced glycation end products

ApoE Apolipoprotein E

ATP Adenosine triphosphate

Aβ Amyloid beta

AβPP Amyloid beta protein precursor bFGF Basic fibroblast growth factor

BH Bleomycin hydrolase

BRCA1 Breast cancer 1

CAK CDK activating kinase

CARB CIP 1 regulator of cyclin B

CDK Cyclin dependent kinase

CERAD Consortium to Establish a Registry for Alzheimer's Disease

CKI CDK inhibitor

CNS Central Nervous System

CSF Cerebrospinal fluid

DAB Diaminobenzadine

DCS Donor Calf Serum

DHFR Dihydrofolate reductase

DSB Double strand breaks

12 ECM Extracellular matrix

EDTA Ethylenediaminetetraacetic Acid

EEG Electroencephalography

EGFR Epidermal growth factor receptors

G6PD Glucose 6 phosphate dehydrogenase

GSH Glutathione

GSHPx Glutathione peroxidase

GSSG-R Reduced Glutathione

GST-NQO1 Glutathione S NQO1

HAT Histone acetyltransferase

HNE Hydroxynonenal

HO-1 Heme Oxygenase-1

HPG Hypothalamic-pituitary-gonadal

IAP Inhibitor of apoptosis protein

JNK Jun N-terminal kinase

LDH Lactate dehydrogenase

LDL Low Density Lipopprotein

MCI Mild Cognitive Impairment

MEKK Mitogen-activated protein kinase kinase kinase

MMP Matrix metalloproteinases

MORF Mortality factor 4

MRG MORF4-Related Gene on 15

MRI Magnetic Resonance Imaging

13 NAD Nicotinamide Adenine Dinucleotide

NADP Nicotinamide adenine dinucleotide phosphate

NADPH Nicotinamide adenine dinucleotide phosphate reduced

NGF Nerve Growth Factor

NGS Nerve growth serum

NIA National Institute of Aging

NINCDS National Institute of Neurological & Communicative Disorders & Stroke

NO Nitrous oxide

NQO1 NADPH Quinone 1

NSAIDS Nonsteroidal anti-inflammatory drugs

PARP Poly (ADP-ribose) polymerase

PBS Phosphate buffered Saline

PCNA Proliferating cell nuclear Antigen

PCR Polymerase Chain Reaction

PET Positron Emission Tomography

PGEX-KG Bacterial Expression vector

PHF Parahelical filaments

PLK Polo like Kinase

PMSF Phenylmethanesulfonyl fluoride

PS1/2 Presenilin 1/2

PUFA Polyunsaturated fatty acid

RAGE Receptor for advanced glycation end products

ROS Reactive oxygen species

14 SAPK Stress-activated protein kinase

SDS Sodium dodecyl sulfate

SEK SAPK kinase

SOD Superoxide dismutase

SPECT Single Photon Emission Computed Tomography

SSB Single strand breaks

TAK1 TGF-β activating kinase 1

TBS Tris Buffered Saline

TGFβ Transforming growth factor-β

15

ONCOGENIC PARALLELS IN ALZHEIMER DISEASE

Abstract

by

ARUN K. RAINA

Alzheimer disease (AD), the leading cause of senile dementia, is characterized by

the selective degeneration of specific neuronal populations responsible for memory acquisition and storage. While a multitude of mechanisms have been proposed to account for such neuronal vulnerability, in this body of work we have focused on three crucial elements; two of which are amongst the earliest and most pervasive changes in the disease, namely alterations in oxidative stress and cell cycle, and one of which is central to the disease process, namely cell death. In Chapter 2, we demonstrate that NADPH quinine oxidoreductase 1 (NQO1) and bleomicine hydrolase, sensitive redox sensors, are upregulated in those neurons vulnerable to degeneration indicating that oxidative stress is

a key and early determinant of disease. In Chapter 3, we focus on cell cycle alterations

using a specific marker (Mrg 15) of cellular emergence from senescence, as well as

downstream markers of a successful passage through the cell cycle, namely p27 and CIP-

1-associated regulator of Cyclin B (CARB). Our data not only further the evidence for a

role of mitotic dysfunction in AD but also suggest that neurons exit from their normal senscent status and at least progress to the G2/M phase of cellular division. In Chapter 4,

we demonstrate that, while initiation of apoptosis is an invariant feature of neurons in the

16 brains of individual with AD, there was no evidence of later apoptotic features such as the activation of effector caspases. This novel phenomenon, we termed “Abortosis”.

Since alterations in redox balance, cell cycle control and avoidance of apoptosis are also key to successful oncogenic transformation, in Chapter 5, we investigated parallels between oncogenesis and AD. Specifically, we determined a key role in AD for two canonical cancer-related proteins, BRCA-1 and retinoblastoma protein, and also found evidence for alterations in the cancer-related matrix metaloproteases, ADAMs 1 and 2.

All of these parallels between AD and neoplasia indicate that AD could represent a cancer in the postmitotic environment and show how the postmitotic state behaves given oncogenic stimulation. Elucidation of mechanisms of this oncogenic stimulation will have broad applications beyond AD itself.

17 Chapter 1

Introduction

1.1 Historical

Progressive mental deterioration in old age has been recognized and described

throughout history. However, it was not until the early part of the 20th century that a

collection of brain neuronal abnormalities were specifically identified by a German physician, Dr. Alois Alzheimer (1864-1915). Dr. Alzheimer first saw Auguste D. in the clinic of Emil Sioli in 1901 in Frankfurt (Maurer et al., 1997). After moving to Munich in 1903, to practice in his own clinic he received the brain of Auguste D upon her demise in 1906. He performed an autopsy on her brain and described miliar foci (plaques) in the upper layers of the cortex, outside and around the nerve cells. Inside the nerve cells he noted the presence of twisted bands of fibers [neurofibrillary tangles (NFT)]. It was at the

37th regional meeting of German psychiatrists in November of 1906 in Tubingen where

he presented the case of Frau Auguste D., a woman who had died after years of experiencing severe memory problems, confusion, and difficulty understanding questions. Interestingly the first case to be correctly diagnosed as Alzheimer disease was

Dr. Alzheimer’s second patient, Johann F. The diagnosis for Johann F was written by Dr.

Alzheimer himself - “Alzheimersche Krankheit”. This was not due to ego but only

because Emil Kraepelin had already named the disease after his colleague in the most current edition of his psychiatry text (Moller and Graeber, 1998; Graeber, 1999). The observation of the plaques and tangles at autopsy is still required to obtain a definitive diagnosis of Alzheimer disease (AD).

18

1.2 Definition of Dementia

Dementia is a clinical syndrome characterized by acquired persistent disturbances in multiple areas of neuropsychological function including language, memory, visuo- spatial skills, emotion and cognition. Thus there must be multiple progressive memory- related cognitive deficits that lead to social dysfunction (American Psychiatric

Association. and American Psychiatric Association. Task Force on DSM-IV., 2000).

1.3 Epidemiology of AD

The course of AD spans about a decade on the average. The proportion of people

afflicted with AD is about 1 percent before the age of 65 years. However, after the age of

65 years this increases to 10 percent. By the age of 85 years and older this figure

increases to about 40 percent. The number of new cases that arise annually also increases

between the ages of 65 years and 85 years from less than 1 percent to about 6 percent

respectively.

There is a gender bias to AD. The disease affects females about twice as much as

males and the duration of AD is also longer in females than in males. Female first degree relatives of affected individuals are also more prone to succumb to AD than are men

during their lifetime (Fratiglioni et al., 1991).

1.4 Pathology of AD

Although the specific molecular etiology of AD remains yet to be resolved there

is consensus surrounding the central elements that form the pathology of the disease

19 namely cortical atrophy grossly and cell loss, intracellular tangles and extracellular

plaques microscopically that occur in the neocortex and its association areas,

hippocampus including the entorhinal cortex, amygdala and the basal nucleus of

Meynert. Sometimes these microscopic changes extend in AD into the medial nucleus of

thalamus, dorsal tegmentum, coeruleus, paramedical reticular area and the lateral

hypothalamic nuclei.

In the normal brain the adjacent gyri are apposed closely to each other so that the

sulci remain only as a potential space between two adjacent gyri. However, in brains of individuals with AD, due to cortical atrophy, the normal CSF pressure exerted within the ventricular system leads to reactive dilatation of these ventricles. This is visible upon a coronal section of the brain (Kumar et al., 2005).

a. Neurofibrillary Tangles

The original description (Alzheimer et al., 1995) of NFT was as follows:

“In sections which had been stained according to Bielschowsky there were very peculiar changes of the neurofibrils. Inside an otherwise still normal appearing cell, one or several fibrils become prominent because of their remarkable thickness and special impregnability. Subsequently, many of the fibrils running next to each other show the same type of change. Then, they cluster as thick bundles and slowly emerge at the surface of the cell. Finally, the nucleus and the cell disintegrate, and only a tangle of fibrils indicates the place where a nerve cell had been previously located. Since these fibrils can be stained with other dyes than normal neurofibrils, a chemical reaction of the fibril substance must have taken place. This is probably the reason why the fibrils survive the

20 demise of the cell. The transformation of the fibrils appears to go hand in hand with the incorporation of an as yet unknown pathological metabolic product into the nerve cell.

Approximately 1/4 to 1/3 of all nerve cells of the cortex show these changes. Numerous nerve cells, especially in the upper cortical layers, have completely disappeared.”

The neurofibrillary tangle is the major intraneuronal lesion in AD and is located in the primary neurons of the affected corticies and their appearance follows the pattern of AD development and cognitive debility (Braak and Braak, 1991). Electron microscopy reveals that NFT are composed of paired helical filaments (PHFs). Paired helical filaments (PHFs) are composed of two axially opposed helical filaments with a diameter of 10 nm and a half-period of 80 nm (Wisniewski et al., 1976). PHFs are formed by the pathological aggregation of the microtubule associated protein τ, which is highly phosphorylated, (Morishima-Kawashima et al., 1995). Importantly, the phosphorylation of τ interferes with its ability to assemble microtubules (Iqbal et al., 1986) which would likely compromise neuronal transport and consequently result in neuronal dysfunction. τ normally binds to microtubules when it is dephosphorylated or only partially phosphorylated. The six human isoforms of the τ are derived from a single gene on . Accumulated paired helical filaments or microtubule destabilization, or both, may disrupt axonal transport and lead to neurofibrillary tangle formation and neuronal death (Alonso et al., 1996).

21 b. Senile Plaques

“Throughout the whole cortex, especially numerous in the upper layers, one finds miliar foci, which are caused by deposition of a peculiar substance in the cortex. It is visible even before staining [the sections] but quite refractory to dyes.”

The above is partly how Alzheimer originally described the senile plaques associated with AD (Alzheimer et al., 1995). These are spherical extracellular lesions 10-

200 µm in diameter that were originally identified and classified by the use of silver based histological stains. Neuritic senile plaques contain a central core made of 6-10 nm amyloid-β (Aβ) protein filaments arranged as bundles (Smith, 1998). The core is surrounded by an argyrophilic rim of dystrophic synapses and neurites (mainly axons) primarily containing paired helical filaments. Aβ protein and its precursor, amyloid β protein precursor (AβPP), as well as τ and neurofilament proteins, are found in this peripheral region of the senile plaque (Perry et al., 1988). A cross-β sheet secondary structure has been observed by X-ray diffraction of Aβ accounts for its congophilic staining leading to birefringence under plane-polarized light after staining with Congo red (Kirschner et al., 1986).

1.5 Clinical Features and Diagnostic Criteria of AD

The presenting features of AD evolve insidiously, from impairment in memory and progress to end in debility and finally death. This march is inexorable with few pit stops along the way. The chief complaint is that of a loss of recently acquired information. In the beginning there is still memory of older events. As the disease progresses there are further losses in linguistic abilities as well as abstract reasoning.

22 Other features include depression, anorexia, agitation and hallucinations. Upon

examination only in later stages is there an involvement of extrapyramidal systems and

lead to postural changes, bradykinesia, shuffling gait and rigidity. What remains relatively unaffected are the primary motor systems (Kasper and Harrison, 2005).

The diagnosis of AD is dependent upon criteria formulated by the joint consensus committee of National Institute of Neurological and Communicative Disorders and

Stroke (NINCDS) and the Alzheimer Disease and Related Disorders Association

(ADRDA) in 1984. These include a progressive history of cognitive loss along with

exclusion of other conditions that may mimic this cognitive loss. A definitive diagnosis is

obtained postmortem and involves the presence of a specific density of senile plaques and

NFT (Khachaturian, 1985).

CSF is normal while the EEG shows generalized slowing and MRI can show

dilatation of the ventricles and widening of the sulci in the cortex. In patients with

moderate and severe symptoms, PET and SPECT show hypometabolism in parietal and

temporal areas (Poulin and Zakzanis, 2002).

1.6 Etiopathogenesis of AD

a. Age

Aging is the most significant risk factors for developing AD. Even a genetic

predisposition to AD means that the disease will only occur after early 50’s (Smith,

1998). With increases in longevity this is indeed a significant challenge for the research

community. If there is no change in the natural history of this disease just the material

costs will easily exceed half a trillion dollars by 2040. However, what seems to be

23 dismal statistics may infact be an opportunity. Indeed a delay of 10 years in disease development would lead to its near eradication.

b. Genetics

A small subset of AD cases are due to mutations on chromosome 21, 14 and 1. As stated earlier, the first degree relatives of people affected by AD, have an increased lifetime risk of getting AD (Breitner and Folstein, 1984). A number of have been linked to familial as well as sporadic AD. The amyloid precursor protein gene has been mapped to chromosome 21. Of course chromosome 21 is the locus for the Down’s trisomy. The presenilin (PS) 1 and 2 genes on chromosome 14 and 1 respectively are also a cause of early onset AD. The PS2 mutations are largely exclusive to the German Volga ethnicity. Other loci of increasing risk susceptibility to AD include chromosome 19

[Apolipoprotein E (ApoE)] and possibly there may be susceptibility regions on 12 and perhaps 10 (Kasper and Harrison, 2005).

i. Amyloid β Protein Precursor (AβPP)

AβPP can be processed via nonamyloidogenic and amyloidogenic pathways.

The fragment of AβPP formed by the latter pathway, termed Aβ, is a soluble

secretory product composed of 39 to 43 residues. This pathway may produce Aβ at

the cell surface, in the lysosomes, or in the Golgi apparatus. The exact mechanism of

Aβ processing depends on the cell type in question. The form of Aβ that is secreted in

highest quantity is Aβ1-40, while lesser amounts are in the Aβ1-42 form. What is

interesting is that both forms are secreted in varying amounts in cognitively normal

and abnormal people, and of both types can be measured in CSF as well as plasma. It

24 is the insoluble form of amyloid namely, Aβ1-42 form that is primarily contained in

the senile plaques. The mechanism(s) by which a mutation version of AβPP

precipitates the onset of AD with very near 100% penetrance is unknown; however, mutations in AβPP gene increase the amount of Aβ produced (Selkoe, 1997).

ii. Presenilins 1 and 2

Mutations in the genes encoding PS1 and PS2 account for the majority of

the cases of familial early-onset AD. The way the familial AD (FAD)-linked

presenilins exert their pathogenicity is by selectively increasing the levels of the

highly fibrillogenic Aβ 1-42 peptides (Takasugi et al., 2003). PS1 is primarily seen in

the brain while PS2 is primarily present in the periphery such as heart, skeletal

muscle and pancreas. Functionally PS1 acts as part of a complex of proteins which

includes Aph-1, Pen-2 and nicastrin. Their γ-secretase activity is based on their

stoichiometric interactions. The endoproteolysis of presenilins is conserved

(Thinakaran et al., 1997).

iii. ApoE

ApoE4 allele is positively correlated with AD while ApoE2 expression is

negatively correlated with AD. With ApoE4 expression there is higher β amyloid

deposition in the brain (Chalmers et al., 2003).

Lipoproteins are transport vehicles for triacylgylcerols and cholesterol to

both the nervous system as well as periphery. They are essentially a core of

hydrophobic lipids surrounded by a shell that is formed by more polar lipids

combined with apoproteins. There are about 10 major apoproteins (Kumar et al.,

2005).

25 ApoE is one of the apoproteins and has three isoforms, namely ApoE2,

ApoE3 and ApoE4. The three isoforms differ by cysteine and aginine residues at

positions 112 and 158. Interaction of ApoE with two classes of cell-surface receptor

mediates the transport of cholesterol inside the cells through the LDL family receptor

and also via a scavenger receptor family member.

ApoE4 is 299 amino acids long. It has two main regions, one which is

located in the N terminal (1–191 amino acids) and the other at the C-terminal end

(216–299 amino acids). These two regions are separated by a so called hinge region

(165–215 amino acids). It is within the N-terminal domain, where the receptor

is localized. The C-terminal domain contains a lipid binding area (Rall et

al., 1982).

What is interesting is that the formation of the ApoE-Aβ complex is

increased in the presence of oxygenated buffer and is quenched during a reductive

environment (Strittmatter et al., 1993). Therefore, the oxidation of ApoE, alone or

bound with Aβ, might affect receptor affinity and/or other catabolic interactions by a

mechanism analogous to the oxidation of low–density lipoprotein in diabetic renal

disease (Gupta et al., 1992).

Only future studies will shed light on the causative role of ApoE in AD

etiopathogenesis. Does it alter cholesterol metabolism in some way, abolish a

protective role in fibril formation, or have a direct effect in increasing amyloid kinetics towards fibrillogenesis?

26 c. Oxidative Stress

Free radicals generated consequent to oxidative stress has been speculated to be important in the pathogenesis of AD and other neurodegenerative diseases (Cross et al.,

1987). There is uncontrovertable evidence that the earliest pathological changes that are characteristic of AD show evidence of oxidative damage (Perry and Smith, 1998). This leaves no doubt of oxidative stress and its role, very early if not simultaneous with the genesis of the AD phenotype. Indeed, this early role is borne out by clinical management of oxidative stress which appears to reduce the incidence and severity of AD (Stewart et al., 1997). A spectra of modifications that indicate the presence of oxidative stress are seen in the neurons, NFT and senile plaques of AD. These include advanced glycation endproducts (AGE), nitration, lipoperoxidative end products, and carbonyl-modified neurofilament protein along with free carbonyls (Castellani et al., 1998). Additionally protein crosslinking by oxidative processes may well lead to lesions being refractory to intracellular as well as extracellular degredation regardless of extensive ubiquitination

(Cras et al., 1995). This resistance of pathological end products like NFT and others to proteolytic destruction would doubtless, have significant impact in the progression of AD

(Smith, 1998).

PHF are a known source of oxidative stress (Yan et al., 1995). PHFs are also found in neuritic pathology with membrane abnormalities indicative of extensive lipid peroxidation and thus strengthens the role of oxidative stress in another key pathophysiological process in AD (Praprotnik et al., 1996). Indeed, vitamin E-deficient rats contain dystrophic neurites like those seen within the senile plaques of patients with

AD (Heslop et al., 1996).

27 Further consensus around oxidative involvement in the genesis of the AD phenotype comes from neuronal damage caused by Aβ being mediated by free radicals

(Behl et al., 1992). Indeed, plaques are surrounded by microglia, which, upon activation, produce free radicals (Colton and Gilbert, 1987). Both NFT and senile plaques contain excess iron (Good et al., 1992). Iron is crucial to the initiation of free radical formation.

Of potential importance in this regard is that oxidative stress upregulates the production of both AβPP and Aβ (Yan et al., 1995) indicating a potential positive feedback link between oxidative stress and Aβ deposition. What we see here is the early influence of oxidative stress in the genesis of signal pathology of AD.

d. Cell Cycle and AD

The primary neurons in the adult CNS do not divide and are terminally differentiated. Thus it is interesting that susceptible neurons in AD have a re-emergence of features that more closely resemble a mitotic cell rather than a terminally differentiated non-dividing cell. Indeed, a number of control proteins that regulate progress through the cell cycle are upregulated including p16, p21, CDK4, CDK5, CDK7, Plk, CARB, Ki67 and mpm-2 (Raina et al., 2000). One of the hallmarks of AD pathophysiology is the increased phosphorylation of τ which in turn leads to the formation of NFT. A similar, if not identical, form of highly phosphorylated τ is also seen in mitotic cells during development (Preuss et al., 1995).

Although terminally differentiated neurons may not be fully competent in mitosis, they may be able to proceed through to a certain point(s) prior to cell division. Thus, what one will expect to see is a ‘transitional’ neuronal phenotype which has a wide array of

28 ‘mitotic’ markers. The presence of CDK4 and CDK2 in AD neurons (McShea et al.,

1999a) indicates this mitotic phenotype. Additionally, the presence of p21, highly phosphorylated τ protein, Ki67, p107 and mpm-2 proteins (Raina et al., 2000) leads one to believe that these neurons may be attempting to activate early phases of mitosis. As to date there is no evidence of a successful nuclear division nor of chromosomal condensation in AD, which would suggest neuronal division proceeding beyond the retinoblastoma checkpoint at late G1.

Even though cell cycle arrest cannot be maintained for a long time and triggers programmed cell death, i.e., apoptosis, it is not clear that similar triggers would exist in a postmitotic environment where neuronal survival is paramount. But during the cell cycle there is an increase of mitochondrial along with other organelles. This huge mitochondrial mass would pose an oxidative threat to the neuron (James and Bohman,

1981) beyond what endogenous antioxidants could blunt. In such a scenario it is easy to account for the global oxidative damage we encounter in AD (McShea et al., 1997).

What remains to be seen is the source of this mitotic pressure. It is logical to envision a path where ectopic expression, release, or re-sensitivity to cellular differentiation factor(s) could trigger the mitotic cascade. Some common candidates that are usually held to be the culprits include nerve growth factor (NGF), transforming growth factor-β1 (TGFβ) and basic fibroblast growth factor (bFGF), all of which show increased activity in AD (Crutcher et al., 1993).

29 Chapter 2

Oxidative Stress in AD

2.1 Introduction

Aerobic respiration is inimical to eukaryotic life. The by-products of aerobic respiration include reactive oxygen species namely, superoxide, hydrogen peroxide, charged hydroxyl radical and singlet oxygen (Ames, 1991). Oxidative stress can be

defined as an imbalance between the production of reactive species (oxygen, nitrogen) and the antioxidant defenses. The sources of oxidative stress include mitochondria, endoplasmic reticulum, microbodies (peroxisomes) and plasma membranes. Although we are familiar with the common examples of oxidative stress acting through reactive oxygen species include the rusting of iron, butter becoming rancid, apples turning brown, what is more insidious is the role of these radicals in neurological systems.

a. Oxidative Stress in AD

The depth and breadth of oxidative damage in the vulnerable neurons in AD is well established for some time now (Marlatt et al., 2004). The spectrum of the damage that occurs within the vulnerable neurons in AD due to chronic oxidative assault includes, the formation of advanced glycation products (Smith et al., 1994a), attacks on lipids leading to the formation of lipid peroxidation adduction products (Montine et al., 1996;

Sayre et al., 1997), the formation of carbonyl-modified neurofilament protein along with free carbonyls (Smith et al., 1991) and nitration end products (Good et al., 1996). There are numerous sources of free radical sources in AD. Free radical generators in AD

30 include iron, which in a redox-active state, are increased in NFT as well as in Aβ deposits

(Good et al., 1992). Iron catalyzes the formation of the hydroxyl radical from hydrogen

peroxide along with the formation of glycation end products (Smith et al., 1995a). Aβ can

produce reactive species by using peptidyl radicals (Hensley et al., 1994). Peri-plaque

microglia that are activated contribute to NO as well as superoxide radical and also lead

to the formation of peroxynitrite (Good et al., 1996; Smith et al., 1997). AGE along with

Aβ interact with certain hypothesized receptors such as RAGE (receptor for AGE) as

well as a scavenger-receptor, to augment reactive oxygen species (ROS) formation (Yan

et al., 1996). Finally any abnormalities in the mitochondrial genome may be the initiating

source of reactive oxygen (Debrinski et al., 1994).

There is a shift in thinking about the earliest events known to date in AD. Indeed

oxidative stress is becoming more proximal in the evolution of the AD phenotype

(Nunomura et al., 2000). Additionally, cerebral cortical neurons in Down’s syndrome

cases in their teens collect cytoplasmic 8-hydroxyguanosine (8OHG) and nitrotyrosine decades prior to Aβ deposition (Nunomura et al., 2001). Indeed this proximal role of oxidative pathophysiology prior to the appearance of the hallmark pathological signs is borne out by the positive outcomes in the use of NSAIDs, which are thought to be efficacious due to their antioxidant actions (Stewart et al., 1997). Finally, Aβ transgenic mice provide in vivo evidence that oxidative lipid modifications precede Aβ deposition

(Pratico et al., 2001). Taken together this evidence points to oxidative stress and its pathophysiology ushering in the well known phenotype of AD.

31 b. Oxidative Events and Apoptosis in AD

There is a large increase in ROS during apoptotic cell death. Oxidative stress is known to be associated with the Bax-induced externalization of cytochrome C (Lenaz,

1998). Bax-deficient neurons showed no increase of reactive oxygen species.

Furthermore in these neurons cytochrome C release in response to growth factor deprivation was totally abrogated. The source of ROS in these sympathetic neurons was the mitochondrion since blocking complex 1 by rotenone totally blunted the ROS increase. The mechanism of mitochondrial ROS generation by Bax is not clear. Indeed, reactive oxygen species can lead to peroxidation of the mitochondrial membrane lipids especially, cardiolipin, which can lead to the dissociation of cytochrome C from its interaction with the inner membrane of the mitochondrion (Ott et al., 2002). Furthermore activated caspase-3 can lead to increased mitochondrial ROS distal to cytochrome c release through damage to complexes I and II (Ricci et al., 2003). A membrane permeable form of reduced glutathione reduced the level of ROS increase and cytochrome c release in neurons that were deprived of NGF (Kirkland et al., 2002). This points to the role of ROS in mitochondrial membrane permeability increase perhaps due in part to lipid or protein oxidation.

ROS along with loss of transmembrane potential are potent inducers of the intrinisic pathway of apoptotic cell death. A wide variety of apoptotic mediators are seen in neurons that are vulnerable in AD (Raina et al., 2001b; Raina et al., 2003). This may suggest that neuronal death in AD is mediated by programmed cell death events that culminate in the apoptotic morphology. Indeed, the environment of the AD brain is awash with proapoptotic mediators such as Aβ, oxidative stress, hydroxynonenal (HNE)

32 oxidants and metabolic alterations with concomitant energy failures. But, the phenotype

that defines the terminal events that are pathogonomic of apoptosis, such as chromatin

condensation, apoptotic bodies and membrane blebbing, are not seen in AD. Although

AD presents with a proapoptotic environment, apoptosis does not proceed to completion.

In this regard, as discussed in Chapter 4, we found that while the initiator phases of

apoptosis were engaged, this does not lead to the activation of the terminal commitment

phase necessary for apoptotic cell death. In other words, in AD, there is a lack of

effective apoptotic signal propagation to distal effectors. Hence the signature

morphology that is apoptosis is not reflected in the AD phenotype. This represents an

inhibition of apoptosis at the post-initiator stage in neurons that survive in AD (Raina et

al., 2001b).

c. Oxidative and Cell Cycle Events in AD

Oxidative stress is known to lead to pleotropic events from DNA damage, cell

cycle initiation and arrest to apoptosis. Additionally oxidative stress may also induce

reactivation of cells to begin cycling. Indeed there could be an ancestral role of

reinitiating the cell cycle as an adaptive response to stress and ROS-induced DNA

damage. This is true of both cycling as well as postmitotic cells. However there is a

problem in postmitotic cells that have sacrificed global DNA repair for repair of genes that are expressed (Nouspikel and Hanawalt, 2003). Hence during all of these years defects have accumulated in the genome spontaneously and essentially left unrepaired.

Add to this picture that DNA lesions such as oxidized bases, single-strand DNA breaks

(SSB), and double-strand DNA breaks (DSB) occurring secondary to oxidative stress

33 (Benitez-Bribiesca and Sanchez-Suarez, 1999) have also accumulated minimally during

the course of AD and now the task of initiating DNA repair mechanisms upon a cell cycle

initiation signal becomes nearly impossible. This can account for the pause at the G1/S

boundary. Indeed a pause is triggered if there are problems encountered in the DNA repair at the G1/S boundary. However in AD we suspect that this pause is permanent and

leads to cell cycle arrest (see Chapter 3 for further details). Indeed, cell cycle inhibitors

like p21 are also up in these vulnerable neurons in AD (McShea et al., 1997). An arrest at

this checkpoint leads to signals that initiate apoptosis. Indeed DNA damage can lead to

the recruitment of p53 to sequester Bcl-XL and release Bax to help externalize

cytochrome C to initiate apoptosis. Additionally other mitogenic loci in AD include

AGEs. AGEs are known to activate mitogenic pathways (Munch et al., 2003).

d. Oxidative Stress - An Adaptation in AD?

What is the adaptive value of phosphorylation and changes in oxidative dynamics,

the two proximal events in these vulnerable primary neurons? Indeed the τ formations

themselves are subject to attack by HNE. Furthermore it is the phosphorylated but not the

nonphosphorylated τ that is the substrate for HNE attack (Takeda et al., 2000).

Any changes in the redox equation in the neuron are strictly monitored by various

systems but especially systems that generate energy (in terms of reducing equivalents).

These systems would sense any changes in reducing equivalents and would respond early

and quickly. Hence we find that the pentose phosphate shunt is induced in AD (Martins et

al., 1986). Indeed the early phosphorylation cascade coincides with induction of heme

oxygenase-1 (HO-1) (Takeda et al., 2000), an that regulates the rate limiting step

34 in the conversion of heme to bilirubin. HO-1 needs reduced NAD as the energy source

(Smith et al., 1994a). Heme is a potent prooxidant while bilirubin is an antioxidant.

Upregulation of pathways that contribute to generating reducing equivalents also leads to

increase in glutathione and free sulfhydryls. Free sulfhydryls which contribute to

quenching the effects of reactive oxygen are increased in vulnerable neurons in AD

(Russell et al., 1999). All of this is in addition to the array of antioxidant responses that

are seen in AD, namely the recruitment of SOD, GSSG-R, catalase and others (Zhu et al.,

2004a).

The picture so far is that some type of stress activates the phosphorylational

cascade, changes in oxidative dynamics and induction of antioxidant responses. Naturally

more needs to be done to understand the adaptive value of these early responses. The early responses of induction of phosphorylation along with oxidative adduction and

antioxidant responses seem to be under selective pressure. These responses may be

present in other diseases, and even during aging but their recruitment to selective set of

neurons makes this a tailored adaptive response.

e. Oxidative Relevance to Lesions

The effects of the oxidative dynamics in AD are numerous and many are long-

lived. What we know so far is that the sequelae of oxidative stress in the near term

include characteristic modifications of macromolecules such as carbohydrates, lipids and

nucleic acid pools.

In AD, we have carbonyl modifications in NFT as well as plaques (Smith and

Perry, 1996). Furthermore these lesions accumulate AGE as well as continue to be the

35 loci for further glycation (Castellani et al., 2001). Assessing the temporal profile we find

that lipid peroxidation (increased isoprostane levels) is present early on in mild cognitive

impairment (MCI) cases (Pratico et al., 2002). Oxidative damage markers of lipid

peroxidation, reactive carbonyls, nitration, and oxidation of nucleic acids as well as their

pools is increased in vulnerable neurons irrespective of whether they contain NFT

(Nunomura et al., 1999a). Additionally, accumulation of 8OHG and nitrotyrosine, in the cytoplasm of frontal cortical neurons from Down’s syndrome early on in their teens,

precedes Aβ deposition by decades and the same accumulation is seen in AD brains

(Nunomura et al., 2000). In fact, Tg2576 AβPP transgenic mice show that oxidative

stress precedes Aβ deposition (Pratico et al., 2001). Antioxidant treatment decreases the

overall levels of Aβ (Lim et al., 2001).

Phosphorylation of τ decreases its binding ability to microtubules and leads to

destabilization of microtubules in neurons (Johnson and Bailey, 2002). Molecules that

tend to induce oxidative stress such as homocysteine and Aβ peptides increase τ

phosphorylation in neuronal cells (Williamson et al., 2002). On the other hand HNE

inhibits τ dephosphorylation (Mattson et al., 1997). HNE attack on phospho-τ generates

the τ conformation that defines the Alz50 epitope that is found in PHF-τ (Takeda et al.,

2000). Indeed it also is associated with the in vitro formation of filaments (Perez et al.,

2002). Additonally, polyunsaturated fatty acids (PUFAs) that are sensitive to redox

changes promote oxidation of other molecules also facilitate τ polymerization (Wilson

and Binder, 1997). This effect is prevented by free radical scavenger molecules (Gamblin

et al., 2000). Hence oxidative stress plays a central role in the development of AD

lesions.

36

f. Neuroprotection within Vulnerable Neurons in AD

Neuroprotection in response to oxidative stress is provided for by a number of

redundant systems. The neuronal antioxidant defenses have been well documented. The

induction of HO-1, Cu/Zn superoxide dismutase, GSHPx, GSSG-R, peroxiredoxins and

several heat shock proteins and their association with intracellular neurofibrillary

pathology (Aksenov et al., 1998) can only mean that vulnerable neurons are mobilizing

antioxidant defenses in the face of increased oxidative stress from an ever increasing

number of sources. However there is limitation to their efficacy. What we see is that,

even after the actions of superoxide dismutase (SOD), hydrogen peroxide still remains in the cytoplasm. These pose a significant threat to biological markers in the presence of

redox-active metals. It is not by happenstance that there is upregulation of antioxidant

systems starting with HO-1 which converts heme to biliverdin and in the process over

time provide free iron which makes the disease worse. Is there some mechanism(s) that

can sense early on the changing oxidative dynamics in these vulnerable neurons?

2.2 Experimental Hypothesis

When the cellular oxidant defense networks are overcome by a chronic oxidative

assault then the resultant is what we call oxidative stress. Oxidative stress and it’s

sequelae are a well known early feature of AD. The targets of oxidative assault are the

building blocks of the cell, namely lipids, proteins, carbohydrates and nucleic acids along

with their pools. Defense systems to combat oxidative stress must be recruited

immediately in order to be efficacious and result in the minimal amount of long term

37 cellular damage. Given all of the above considerations, I chose to look at three things that

can occur within an oxidative environment. First is there a marker that can sense early on

the process of a changing redox environment? Furthermore is there any genetic evidence that can tie together oxidative stress and AD.

Given that the primary neuron is a site of high energy metabolism and that these oxidatively stressed neurons live for a long time, there must be multiple types of systems/proteins that may be recruited into this environment in AD either as early sensors or antioxidants or with multiple functions. In this study, I chose to focus on two of these,

NADPH quinine oxidoreductase (NQO1) and blemoycin hydrolase (BH). NQO1 is upregulated under conditions of oxidative stress (Riley and Workman, 1992). NQO1 is a coordinately transcriptionally regulated Phase II enzyme that has an important role in quinone detoxification by a two-electron transfer, redox-sensitive chemo-protection in

some neoplastic models, activation of prodrugs, carboxylation by vitamin K (Cadenas,

1995). NQO1 is also involved in the upstream regulation of the MEKK > SEK > SAPK >

JNK stress activated cascade (Cross et al., 1999) which is activated in AD (Zhu et al.,

2003). Additionally, it is postulated that the adaptive significance of NQO1 may be to

regenerate the reduced form of membrane-bound Coenzyme Q, which is a potent

antioxidant within membranes in addition to the role of Coenzyme Q in mitochondrial

oxidative phosphorylation (Beyer, 1994). Indeed, Coenzyme Q is ubiquitous in the

various compartments of the cell and especially in the cell membranes. As discussed

earlier, mitochondria are a potent source of oxidative stress. Therefore, NQO1 acts as a

sensitive indicator of the redox state of cells especially in cells which are oxidative load

and need reductive compensations. NQO1 could signal the dynamics of reductive

38 compensations in the vulnerable neurons of AD. Indeed, the importance of NQO1 as a

monitor of redox balance together with the proposed importance of oxidative stress in

AD provided the impetus for us to explore the status of NQO1 in AD.

BH is a ubiquitous, evolutionary-conserved intracellular peptidase. The only known

function of BH, to date, is the inactivation of the anti-neoplastic glycopeptide, bleomycin,

which produces a DNA-lytic effect by generating superoxide. Naturally, this function is

not a result of selective pressure. There are other selection-dependent forces that would induce the recruitment of BH. Since BH is acutely induced in this oxidative environment,

we suspected that increased oxidative stress in AD would not only trigger BH induction

but also would effectively date this induction proximally in the course of disease

pathophysiology. Naturally, the variable susceptibility to AD dependent on

polymorphisms at the BH-PHEN locus gave us further impetus to explore BH expression

in AD (Raina et al., 1999b).

Below are the main hypotheses that would be addressed.

1) Neurons that develop NFT survive prolonged exposure to oxidative stress and

show reduced levels of oxidative stress. Hence there must be a mechanism(s) which acts

as not only a sensor of this redox imbalance but whose selective recruitment allows for

neuronal survival in these neurofibrillary tangle-contaning neurons. A candidate sensor is

NQO1 which is intimately involved in generation of antioxidant forms of Coenzyme Q

and vitamin E. We will expect to see NQO1 in the vulnerable neurons of AD. The

profile of NQO1 in the AD brain should parallel the vulnerability profile in AD as seen

using known markers of oxidative stress and τ.

39 2) Since oxidative stress is an early feature of AD, then there may be polymorphisms associated with the response to oxidative stress in AD. Indeed the known genetic polymorphisms in AD namely AβPP, PS1/2 and Apo E4 are associated with either increased oxidative stress or elevated vulnerability to oxidative stress. Such a locus is BH-PHEN, where polymorphisms could give rise to changes in increased susceptibility to AD. This could be a genetic marker of oxidative stress induction in AD.

2.3 Materials and Methods

a. Tissue Section Preparation

Hippocampal tissue samples were obtained postmortem from patients (n = 22, ages 68-95) with histopathologically confirmed AD, as well as from young and aged- matched controls (n = 26, ages 3-85) for the NQO1 study. In the BH study, CNS tissue samples were obtained postmortem from patients (n = 12, ages 69-95) with histopathologically-confirmed AD, as well as from young and aged-matched controls (n

= 13, ages 17-86). From the clinical reports available to us, we could discern no obvious differences in agonal status between the groups. Tissue was fixed by methacarn

(methanol: chloroform: acetic acid in a 6:3:1 v/v/v) immersion for 16 h at 4°C. Some sections were fixed in buffered formalin. Tissue was subsequently dehydrated through graded ethanol and xylene solutions and embedded in paraffin. Six-micron thick microtome sections were prepared and placed on saline-coated slides.

b. Antibodies/Protein Preparations

For the NQO1 study, a cDNA clone encoding human NQO1 was isolated during a

40 yeast two hybrid screen (Deak and Templeton, in preparation) from a HeLa cDNA

library. This cDNA sequence was completely identical to that of Genbank entry J03934.

The complete NQO1 coding region from this cDNA was expressed using the bacterial expression vector pGEX-KG. The GST-NQO1 fusion protein grown in E. coli was purified by affinity chromatography on GSH-agarose beads. Polyclonal antisera was developed in rabbits reared at Cocalico Biologicals (Reamstown, PA) by immunization of animals using SDS-denatured and acetone precipitated protein were used to confirm the

GST-NQO1 fusion protein.

In the BH study, we used affinity purified mouse or rabbit polyclonal antibodies to BH (gift of Stephen A. Johnston). To verify the specificity of immunolabeling, adsorption experiments were performed by incubating the antibody with 1mg/ml of recombinant BH protein (gift of Stephen A. Johnston and Lemoor Joshua-Tor).

c. Immunocytochemistry

Following hydration, sections were immunostained by the peroxidase- antiperoxidase procedure. Adjacent sections were also immunostained with antibodies to

τ, rabbit antiserum and monoclonal antibody, AT8, to locate pathological changes.

Additionally, counterstaining with Congo red was used to label NFT and senile plaques.

To verify the specificity of immunolabeling, adsorption experiments were performed by incubating the antibody with 1mg/ml of relevant peptide at 4°C for 16 hr or by omission of the primary antibody. Immunoreactivity was markedly improved by the use of formic acid (70%, 5 min) pretreatment, presumably due to the unmasking of a previously occult epitope.

41

2.4 Results

a. NQO1

NQO1 immunolabeling in AD was present in two distinct forms. First, within intracellular NFT, neuropil threads and other neurofibrillary pathology, and second, as cytoplasmic staining in hippocampal pyramidal neurons (Figure 2.1). Quantification of

NQO1 immunoreactivity in comparison to τ immunocytochemistry (rabbit antiserum or

AT8) and Congo red histochemistry is summarized in Table 2.1. In brief, there is significant overlap of the immunostaining profiles of NQO1 and τ-positive neurofibrillary pathology. However, importantly, while labeling with a marker to early τ abnormalities (AT8) showed essentially complete overlap with NQO1. This was significantly less using a marker that recognizes τ (rabbit antiserum) and Congo red histochemistry. In fact, NQO1 was only evident in a subpopulation (~6-39%) of all intracellular NFT (Table 2.1).

Although NQO1 can be weakly detected in all cases (control and AD) after 43 years of age, AD cases show markedly greater numbers of neurons displaying NQO1 (Table

2.1). Additionally, the few NFT noted in aged controls were also NQO1 immunoreactive. Absorption using NQO1 protein abolished the NQO1 labeling (not shown).

b. Bleomycin Hydrolase

Immunocytochemistry by using the rabbit and mouse anti-BH sera showed high levels of cytoplasmic staining only within a subpopulation of the hippocampal pyramidal

42 neurons in AD cases (Figure 2.2A). There was no staining in the control sections (Figure

2.2B). The rabbit anti-hBH sera generated a stronger immunoreactive profile compared to the mouse antibody to hBH; furthermore, it was immunolocalized, possibly to subcellular organelles. Nonetheless, the overall match of the immunoprofiles from the two related antisera underwrites the inference of an increase in cytoplasmic BH levels.

Attesting to the specificity of our findings, absorption of each antisera with purifed BH protein abolished completely the BH immunoreactivity (Figure 2.3A,B). Localization of

τ in adjacent sections showed considerable overlap with BH but significantly, BH did not recognize any neurofibrillary pathology (NFT and neuropil threads), or other end-stage manifestations of AD. Thus BH upregulation could signal early events in AD pathophysiology.

2.5 Discussion

a. NQO1

In this study, we found increased NQO1 immunolabeling in AD neurons compared to young and age-matched controls. The increased association of NQO1 with neurons displaying NFT (an end-stage manifestation in AD) and also with neurons in cases of AD with pre-NFT, as well as those with no obvious histopathological manifestations of AD pathology, suggests that NQO1 is upregulated early in AD and is of chronic duration. Such a notion is supported by labeling, albeit at lower levels, of control cases over age 43. The lack of immunolabeled neurons or pathology in AD, upon adsorption with NQO1 protein confirms the specificity of the immunolabeling.

Furthermore, antibodies to NQO1 do not recognize τ protein isolated from the human

43 brain and antibodies to τ do not recognize NQO1.

In vivo, NQO1 may be responsible for reducing membrane-bound Coenzyme Q

(CoQ), besides its other roles, e.g., as a Phase II detoxification enzyme. This allows

Coenzyme Q to function as an antioxidant by preventing lipid peroxidation and protein oxidation. Indeed, NQO1 upregulation may signal an early change in the redox-state of

AD neurons, as well as being a marker of oxidative stress. Other results consistent with an altered redox potential include an increase in the level of glucose-6-phosphate dehydrogenase (G6PD) in AD cases compared to normal controls. An increase in G6PD, a rate-controlling enzyme for the hexose monophosphate pathway which generates

NADPH, may lead to a more reducing intraneuronal environment via a greater NADPH to NADP ratio (Balazs and Leon, 1994). Normally, damage by ROS is kept in check by a redundant array of antioxidant , but in pathological conditions the oxidant- antioxidant equilibrium is altered, leading to an overabundance of ROS. Among the findings suggestive of oxidative stress in AD is (i) the presence of defective energy

metabolism (Beal, 1995; Bowling and Beal, 1995), (ii) compensatory upregulation of

antioxidant enzymes (Smith et al., 1994a; Premkumar et al., 1995), (iii) protein changes due to oxidation (directly or indirectly) such as crosslinked and non-crosslinked AGEs-, malonyldialdehyde- and HNE-adducts in NFT and senile plaques (Smith et al., 1994a),

(Sayre et al., 1997), induction of an oxidative response; (iv) AGE- in neuroblastoma cells

(Yan et al., 1994), (v) oxidized neurofilament heavy subunit in NFT (Smith et al., 1995b)

(vi) peroxynitrite-mediated damage (Smith et al., 1997) and (vii) greater lipid peroxidation (Lovell et al., 1995; Subbarao et al., 1990). Oxidative stress-induced

upregulation of NQO1 may be mediated through the binding of various redox-sensitive

44 transcriptional regulators to the NQO1 gene promoter site. Among these are NFκB, AP-

1, Nrf1 and Nrf2, which share the commonality that they are activated by ROS

(Venugopal and Jaiswal, 1996; Schreck and Baeuerle, 1991).

b. Bleomycin hydrolase

BH is induced in vulnerable neuronal populations in AD. That BH is generally not associated with end-stage cellular events (i.e., neurons containing neurofibrillary pathology) coupled with the well-known acute induction of BH, strongly suggests that

BH induction is an early event in neuronal pathophysiology in AD. Additional support for this derives from recent findings showing that oxidative stress is also reduced as neurons degenerate (Nunomura et al., 2000). Whether chronic BH upregulation contributes to neuronal degeneration in AD awaits further study, however, given the select location in vulnerable neurons, known acute upregulation of BH together with the genetic association, it appears likely that reactive oxygen in concert with BH do contribute to disease pathogenesis. Indeed, BH, a potent cysteine is a candidate

β-secretase to cleave AβPP (Abraham C., personal communication) and therefore could play a role in Aβ deposition, a cardinal pathological feature of AD.

BH is an evolutionarily conserved, broad spectrum peptidase that is found both within the cytoplasm and, given its DNA-binding activity, in the nucleus as well. The only well-documented physiologic role of BH is the pharmacokinetic inactivation of the

anti-mitotic agent bleomycin. Other putative functions assigned to BH center around its

involvement in the breakdown of proteins (Ferrando et al., 1996) including, as mentioned

above, AβPP. Interestingly, BH also binds to the human homolog of ubiquitin-

45 conjugating enzyme 9 (Koldamova et al., 1998), and it is notable that extensive

ubiquitination is noted in the diseased brain (Perry et al., 1987).

It is intriguing to note that bleomycin produces DNA lysis secondary to the

formation of superoxide, with cell cycle arrest occurring in G2. Bleomycin generates superoxide by binding DNA and chelating iron and copper. In the presence of molecular oxygen, this complex of bleomycin-DNA-Fe/Cu serves effectively as an oxidase forming superoxide (Lazo and Pitt, 1995). Therefore, it is perhaps not surprising that the location of BH closely parallels that of redox-active iron. A diverse number of studies have indicated that ROS generate cytoprotection by upregulation of multiple anti-oxidant systems many of which are found in AD. Given that bleomycin generates ROS, it is possible that ROS serve as secondary messengers to upregulate BH and in support of this, an AP-1 site upstream of the human BH gene may serve as the target site (Zheng and

Johnston, 1998). Such an oxidative-inductive pathway is quite consistent with, and explains, the otherwise remarkable coincidence between vulnerable neurons in the labeling of BH and markers of oxidative damage in AD.

2.6 Conclusions

We found that increased NQO1 is associated with susceptible neuronal populations in AD and is likely an early change in the pathological process. While we cannot definitively demonstrate that NQO1-immunoreactiive neurons go on to die, given the effective function of NQO1 as a redox sensor and a cytoprotective enzyme, the findings presented here do suggest that oxidative and redox imbalance is a significant proximal feature that is intimately associated with the pathogenesis of AD. Furthermore,

46 the proximal induction of BH in AD, as reported here, is significant in several regards.

First, given the known genetic association, together with the oxidative induction of BH,

these findings support genetic evidence for oxidative stress in AD. Second, that BH

cleaves AβPP not only links BH to disease pathogenesis but by association, also links

oxidative stress to AβPP metabolism.

NQO1 also is associated with donating reductive equivalents, a process whose

induction would occur early on during the onset of oxidative stress while BH is know to

be induced in only an oxidative environment. Both of these novel findings underline the

early onset of oxidative stress ss well as its intimate association with AD neuropathology.

All of the above lead to the conclusion of oxidative stress as being intimately involved in

the early pathogenesis of AD (Figure 2.4).

47 Table 2.1 Quantitative comparison of NQO1 immunoreactivity in AD and control cases.

NFT identity was based on Congo red labeling for determination of NQO1 immunoreactive neurons containing NFT and τ-immunoreactivity of determination of number of NFT/mm2 alternatively. The number of NQO1-positive neurons are those assessed above with greater intensity than formed in young (< 40 years) control individuals.

Control Cases

Age Number of NQO1- Number of NFT/mm2 Percent NQO1-positive positive neurons/mm2 neurons containing NFT 43 0 -- -- 53 0 -- -- 54 0 -- -- 69 0.8 3.0 2.3 70 0.5 1.0 28.0

AD Cases

Age Number of NQO1- Number of NFT/mm2 Percent NQO1-positive positive neurons/mm2 neurons containing NFT 79 25.13 22 6.3 79 5.26 11 28.0 81 17.5 52 26.0 83 33.5 49 19.0 88 9.07 19 26.0 89 5.16 2.3 39.6

48

Figure 2.1. NQO1 in AD. Neurons containing tangles are strongly recognized by the antibody to NQO1 (A) while in the brains of control individuals (B) no specific staning was noted. Scale bar = 50µm (Raina et al. 1999).

49

Figure 2.2. Immunocytochemical localization of bleomycin hydrolase (using the rabbit antisera) in AD (A) and control (B) brain tissue. Neurons known to degenerate during the disease are selectively labeled in AD cases only. Scale bar = 50µM.

50

Figure 2.3. Immunoreactivity is prominent in vulnerable neurons in AD (A) and is almost completely abolished by absorption with purified BH protein (B). * indicates a blood vessel in adjacent serial sections. Scale bar = 50µm.

51

NQO1 Oxidation NQO1

SAPK JNK/p38 Normal Pre-NFT NFT X-NFT Pre-Tau Stage . . Phosphorylation . . . HNE Adducts . of τ . of τ Form

10-20 Years!!!!

Figure 2.4 The contribution of NQO1 to the early pathophysiology in AD.

52 2.8 Relevant Publications

1. Raina, A.K., Templeton, D.J., Deak, J.C., Perry, G. and Smith, M.A. (1999)

Quinone reductase (NQO1), a sensitive redox indicator, is increased in

Alzheimer’s disease. Redox Report, 4, 23-27.

2. Raina A.K., Takeda, A., Nunomura, A. Perry, G. and Smith, M.A. (1999)

Genetic evidence for oxidative stress in Alzheimer’s disease. NeuroReport, 10,

1355-1357.

3. Russell, R.L., Siedlak, S.L., Raina, A.K., Bautista, J.M., Smith, M.A. and Perry,

G. (1999) Increased neuronal glucose-6-phosphate dehydrogenase and sulfhydrl

levels indicate reductive compensation to oxidative stress in Alzheimer disease.

Arch. Biochem. Biophys., 370, 236-239.

4. Perry, G., Roder, H., Nunomura, A., Takeda, A., Friedlich, A.L., Zhu, X., Raina,

A.K., Holbrook, N., Siedlak, S.L., Harris, P.L.R. and Smith, M.A. (1999)

Activation of neuronal extracellular receptor kinase (ERK) in Alzheimer disease

links oxidative stress to abnormal phosphorylation. NeuroReport, 10, 2411-2415.

5. Perry, G., Hirai, K., Nunomura, A., Raina, A.K. and Smith, M.A. (1999)

Neuronal oxidative damage in Alzheimer disease may be brought about by a

fundamental shift in redox balance and oxidative metabolism. Recent Res. Devel.

Neurochem., 2, 277-285.

6. Raina, A.K., Perry, G., Nunomura, A., Sayre, L.M. and Smith, M.A. (2000)

Histochemical and immunocytochemical approaches to the study of oxidative

stress. Clin. Chem. Lab. Med., 38, 93-97.

53 7. Kim, H.T., Russell, R.L., Raina, A.K., Harris, P.L.R., Siedlak, S.L., Zhu, X.,

Petersen, R.B., Shimohama, S., Smith, M.A. and Perry, G. (2000) Protein

disulfide in Alzheimer disease. Antiox. Redox Signal., 2, 485-489.

8. Perry, G., Raina, A.K., Nunomura, A., Wataya, T., Sayre, L.M. and Smith, M.A.

(2000) How important is oxidative damage? Lessons from Alzheimer’s disease.

Free Radic. Biol. Med., 28, 831-834.

9. Rottkamp, C.A., Nunomura, A., Raina, A.K., Sayre, L.M., Perry, G. and Smith,

M.A. (2000) Oxidative stress, antioxidants, and Alzheimer disease. Alzheimer

Dis. Assoc. Disord., 14 (Suppl. 1), S62-S66.

10. Perry, G., Kaminski, M.A., Nunomura, A., Raina, A.K. and Smith, M.A. (2000)

What oxidative stress tells us about therapeutic targets for Alzheimer’s disease.

Current Opinion in CPNS Investigational Drugs, 2, 423-426.

11. Smith, M.A., Rottkamp, C.A., Nunomura, A., Raina, A.K. and Perry, G. (2000)

Oxidative stress in Alzheimer’s disease. Biochim. Biophys. Acta, 1502, 139-144.

12. Raina, A.K., Zhu, X., Rottkamp, C.A., Monteiro, M., Takeda, A. and Smith,

M.A. (2000) Cyclin’ toward dementia: cell cycle abnormalities and abortive

oncogenesis in Alzheimer disease. J. Neurosci. Res., 61, 128-133.

13. Perry, G., Nunomura, A., Friedlich, A.L., Boswell, M.V., Brazdil, L., Jones, P.K.,

Rottkamp, C.A., Zhu, X., Raina, A.K., Hirai, K., Friedland, R.P., Shi, J., Aliev,

G., Cash, A.,Russell, R.L., Wataya, T., Shimohama, S., Atwood, C.S. and Smith,

M.A. (2000) Factors controlling oxidative damage in Alzheimer disease: metals

and mitochondria. In Free Radicals in Chemistry, Biology and Medicine,

54 Yoshikawa, T., Toyokuni, S., Yamamoto, Y. and Naito, Y. (Eds). OICA

International (UK) Limited, London, UK, pp 417-423.

14. Rottkamp, C.A., Raina, A.K., Zhu, X., Gaier, E., Bush, A.I., Atwood, C.S.,

Chevion, M., Smith, M.A., Drew, K.L., Nunomura, A., Takeda, A., Hirai, K.,

Zhu, X., Atwood, C.S., Perry, G. and Smith, M.A. (2001) Redox-active iron

mediates amyloid-ß toxicity. Free Radic. Biol. Med., 30, 447-450.

15. Raina, A.K., Rottkamp, C.A., Sayre, L.M., Friedland, R.P. and Perry, G. (2002)

Mitochondrial dysfunction in Alzheimer disease: the chickens or the eggs?

Neurochem. Int., 40, 527-531.

16. Cash, A.D., Perry, G., Ogawa, O., Raina, A.K., Zhu, X. and Smith, M.A. (2002)

Is Alzheimer’s disease a mitochondrial disorder? Neuroscientist, 8, 489-496.

17. Lee, H.G., Zhu, X., Ghanbari, H.A., Ogawa, O., Raina, A.K., O’Neill, M.J.,

Perry, G. and Smith, M.A. (2002) Differential regulation of glutamate receptors in

Alzheimer disease. Neurosignals, 11, 282-292.

18. Aliev, G., Seyidova, D., Raina, A.K., Obrenovich, M.E., Neal, M.L., Siedlak,

S.L., Lamb, B.T., Vinters, H., LaManna, J.C., Smith, M.A. and Perry, G. (2003)

Vascular hypoperfusion, mitochondria failure and oxidative stress in Alzheimer

disease. Proc. Indian Natn. Sci. Acad., Part-B, 69, 209-238.

19. Chao, M., Zhu, X., Raina, A.K., Aliev, G., Takeda, A., Nunomura, A., Tabaton,

M., Perry, G. and Smith, M.A. (2003) Sources contributing to the initiation and

propagation of oxidative stress in Alzheimer disease. Proc. Indian Natl. Sci.

Acad., Part-B, 69, 251-260.

55 20. Zhu, X., Raina, A.K., Lee, H.G., Chao, M., Nunomura, A., Tabaton, M.,

Petersen, R.B., Perry, G. and Smith, M.A. (2003) Oxidative stress and neuronal

adaptation in Alzheimer disease: the role of SAPK pathways. Antiox. Redox

Signal., 5, 571-576.

21. Liu, Q., Raina, A.K., Smith, M.A., Sayre, L.M. and Perry, G. (2003)

Hydroxynonenal, toxic carbonyls, and Alzheimer disease. Mol. Aspects Med., 24,

305-313.

22. Perry, G., Raina, A.K., Nunomura, A., Lee, H.G., Zhu, X., Ogawa, O., Liu, Q.,

Perez, M., Cash, A.D., Avila, A., Castellani, R.J., Atwood, C.S., Aliev, G.,

Takeda, A. and Smith, M.A. (2003) Oxidative stress and the pathogenic

mechanism of Alzheimer disease. In Alzheimer’s Disease and Related Disorders:

Research Advances, Iqbal, K. and Winblad, B. (Eds). Ana Aslan International

Academy of Aging, Bucharest, Romania, pp 425-432.

23. Lee, H.G., Zhu, X., O’Neill, M.J., Webber, K., Casadesus, G., Marlatt, M.,

Raina, A.K., Perry, G. and Smith, M.A. (2004) The role of metabotropic

glutamate receptors in Alzheimer’s disease. Acta Neurobiol. Exp., 64, 89-98.

24. Zhu, X., Raina, A.K., Lee, H.G., Casadesus, G., Smith, M.A. and Perry, G.

(2004) Oxidative stress signaling in Alzheimer’s disease. Brain Res., 1000, 32-39.

25. Dwyer, B.E., Raina, A.K., Perry, G. and Smith, M.A. (2004) Homocysteine and

Alzheimer’s disease: a modifiable risk? Free Radic. Biol. Med., 36, 1471-1475.

56 Chapter 3

Cell Cycle Dysfunction in AD

3.1 Introduction

a. The Cell Cycle and Control Mechanisms

The cell division cycle is a mechanism for growth. Precise control is important for

appropriate growth and its dysfunction can have untoward consequences. It is composed

of four phases that occur sequentially. The next phase begins only after the previous

phase has finished. The sequential order is G0/G1 > S > G2 > M > G1/G0. There are two

growth phases G1 and G2 (needed to accumulate constituents for the subsequent phase), a

DNA replication phase (S) and a division phase called M. Although the time period of G1 and G2 phases can change, that of the S and M phases does not change in general. The S phase typically lasts for about eight hours while the M (mitosis) phase lasts about an

hour. The G2 lasts for about 2 hours while the G1 can last from a few hours, days to finally even a lifetime; the latter state is defined as G0 or terminal differentiation. This

state is marked by functionality without proliferation. Entry into G0 does not prohibit re-

entry back into G1. The time course of these phases is contracted significantly during

embryogenesis and cancer (Figure 3.1).

The regulation of the cell cycle is by two means. First, by the sequential order of

each of the phases, in that the subsequent phase does not start until all of the events of the

prior phase have not been completed and second by a dedicated system of proteins

(activated cyclin-cdk complexes) that are regulated intrinsically and extrinsically at

multiple levels, with each level consisting of control by synthesis, degradation,

57 activation, inhibition and activating and deactivating modifications such as

phosphorylation/dephosphorylation cycles. Needless to say, the cyclin dependent kinases

are under selectional pressure (Alberts, 2004). A brief overview of the cell cycle is given

in Figure 3.2.

The combination of an inactive kinase catalytic subunit and the phase dependent

synthesis of a specific cyclin (A, B, D, E) form the fully active form of the kinase.

Optimal activity of this complex enzyme is dependent upon another enzyme called CDK

activating kinase (CAK). It is the CAK (composed of cyclin H and CDK7) that

phosphorylates the Cdk-cyclin complex on theronine residues. This type of phophorylation is activating. The activity of the Cdk-cyclin complex is terminated by phosphoyrylation on the ATP binding sites of this complex by dual specificity (threonine and tyrosine. Reactivation of a specific Cdk-cyclin complex is possible using Cdc25A, B, and C, a family of phosphatases. Further, inhibitory regulation of Cdk-cyclin complexes comes from the CKI (CDK inhibitors) family of proteins. This family is classified according to their structure and Cdk affinity. They have two broad classes namely the

Cip/Kip and the INK4 families. The Waf/Kip family is composed of p21WAF1/CIP1, p27

KIP1 and p57KIP2, while the INK4 family is composed of INK4a/p16, INK4b/p15,

INK4c/p18 and INK4d/p19. The Cip/Kip families are wide spectrum in the acquisition of

their target substrate, while the INK4 family members share a common structural and

mechanistic inhibitory profile that is circumscribed to Cdk4 and Cdk6. Thus the

regulation of CKIs in terms of their synthesis, proteolysis and CKI exchange are among

the important methods of their regulation of Cdk-cyclin activity. The CKIs themselves

are subject to control from both mitotic and anti-mitotic signals. Finally it is the

58 ubiquitin-mediated proteolytic system that is responsible for the degradation of the many

classes of these cell cycle regulators which then is the final go-ahead signal for entry into

the next phase (Lodish, 2004).

b. Reappearance of Cell Cycle Markers in AD

Primary neurons of the adult brain do not divide and are terminally differentiated.

However the hippocampal pyramidal neurons in AD have features reminiscent of a cell

that is attempting to prepare for cell division (Figure 3.2B). Indeed, the reappearance of a

whole host of the cell cycle markers (Listed in Table 3.1) like cyclin D, cdk4, and/or

Ki67 in all the cases reported suggests that the neurons in AD are no longer quiescent.

The appearance of cyclin E/cdk2 complex reflects that neurons may have emerged from

G0. Perhaps even more importantly the presence of cyclin E/cdk2 complex indicates that neurons have passed G1 and are therefore committed to division or death without the

possibility of de-differentiation. In support of this, the presence of coordinated DNA

replication suggests that the susceptible neurons may complete a nearly full S phase.

Additionally, the aberrant expression of cyclin B1/cdc2 complex indicates that

degenerating neurons in AD may even, in some cases, reach G2 phase. However, while neurons in AD do appear to re-enter into the cell cycle, mitotic figures have never been

observed leading us to speculate that there is a “mitotic catastrophe”. Indeed, the highly

unorganized nature of the cell cycle in AD neurons is pointed to by the concurrent

expression and aberrant localization of PCNA and cyclin B; the concurrent appearance of cdk4 and p16; and the presence of cyclin E and cyclin B but absence of cyclin D and cyclin A. All these abnormalities point to an inadequate or a failed control of cell cycle in

59 these neurons the reason of which is still unclear but it appears that oxidative stress caused damage and cell cycle arrest may play a role. Although it is believed that the aberrant re-entry of cell cycle in susceptible neurons may contribute to their eventual death in AD, mitotic proteins are not exclusively associated with end stage of

neuropathology but rather with the very earliest neuronal changes to occur in the disease.

Indeed cell cycle markers occur prior to the appearance of gross cytopathological changes and the proximal nature of mitotic events has been shown in pre-AD patients with MCI which represents a prodromal stage of AD.

c. Association of the Mitotic Phenotype with AD Neuropathology

All of the major genetic and protein elements, including τ formations, Aβ, AβPP,

PS, and, possibly, ApoE, that are dysregulated in AD and/or confer increased susceptibility to AD are altered during the cell cycle progression. Indeed there are significant similarities between AD and mitotic cells (Table 3.2).

The major protein component of NFT is a highly phosphorylated form of the microtubule associated protein, τ (Iqbal et al., 1984; Grundke-Iqbal et al., 1986).

Hyperphosphorylation of τ renders it unable to stabilize microtubular dynamics and results in neuronal dysfunction (Lindwall and Cole, 1984; Alonso et al., 1996). However, hyperphosphorylation of τ also occurs when cells are mitotically active, in which case, phosphorylation is driven by cyclin-dependent kinases (CDKs) (Brion et al., 1985;

Kanemaru et al., 1992; Goedert et al., 1993; Pope et al., 1994). Notably, CDKs have been localized in vivo to lesions in AD and phosphorylate τ in in vitro assays in a manner similar to that found in AD in vivo (Arendt et al., 1995), (Vincent et al., 1996), (Nagy et

60 al., 1997). Since increased phosphorylation and altered microtubule stability are coincident during progression through the cell cycle, it is predictable that there are microtubular abnormalities associated with AD (Terry et al., 1964).

AβPP is a single pass membrane protein expressed at the cell surface, the proteolytic processing of which gives rise to the Aβ peptide found in senile plaques and a soluble N-teminal fragment (sAPP). In vitro studies demonstrate that AβPP and its proteolytic fragments (i.e., Aβ peptide and sAβPP) are mitogenic. Notably, Aβ can promote the activation of the mitotic cycle in cultured differentiated neurons which enter the S phase and start the replication of DNA. sAPP has been shown to have epithelial growth factor activity, inducing a two- to three-fold increases in the rate of cell proliferation and cell migration. While the function of AβPP has yet to be elucidated, several adaptor proteins, including Fe65 and AβPP-BP1, that interact with its C-terminus have been identified and may transduce its downstream signals. Interestingly, these protein have been shown to function as regulators of the cell cycle. AβPP-BP1 is a cell cycle protein that normally negatively regulates the progression of cells into the S phase and positively regulates progession into mitosis. Strikingly, the other adaptor, Fe65, is a nuclear protein and also regulates negatively G1 to S phase cell cycle progression by inhibiting the key S phase enzymes (Bruni et al., 2002). It is therefore conceivable that

AβPP may act as a cell surface receptor to relay cell cycle-related signals. It therefore follows that overexpression or mutation of AβPP may push neurons into an aberrant cell cycle and in support of this hypotheses, FAD mutants of AβPP have a greater capacity to drive DNA synthesis than expression of wild AβPP.

61 The other known mutations that have been linked to FAD are in the human PS 1 and 2 proteins which required for the proteolytic processing of Notch and the AβPP, molecules that play pivotal roles in cell-fate determination during development and AD pathogenesis, respectively. It has been shown that the overexpression of both PS1 and

PS2 proteins resulted in G1 phase arrest of the cell cycle (Janicki and Monteiro, 1999;

Janicki et al., 2000) which may be due to the decrease in Cdk4 activity and phosphorylation of the retinoblastoma tumor suppressor protein. Overexpression of FAD

PS1/2 mutants further increase cell cycle arrest compared to wild-type PS1/2 and the degree to which the different FAD PS1 mutants inhibits cell cycle progression correlates somewhat with the age of AD onset induced by the mutations. Conversely, PS1 deficiency results in accelerated entry into the S phase of the cell cycle. Furthermore, impaired cell cycle control of neuronal precursor cells in the neocortical primordium with significantly prolonged length of S-phase is demonstrated in PS1-deficient mice. This compelling evidence suggest that PS1/2 plays an important role in cell cycle control and thus, the disruption of PS1/2 function caused by FAD mutants could affect the regulation of cell cycle.

The ApoE4 alleles confer increased susceptibility both to AD and prostate cancer, suggestive of an association between the E4 allele and a propensity toward developing a dysregulated cell cycle (Slooter and van Duijn, 1997). In support of such a notion, ApoE is itself associated with the amyloid deposits found in pituitary adenomas (Steusloff et al.,

1998).

62 3.2 Experimental Hypothesis

In AD, there is significant evidence of reentry into the cell cycle by the primary

neurons which are supposed to be postmitotic in affected regions throughout the cortex as

well as subcortical structures. However, the significance of such cell cycle markers is

clouded by duplicity of function in that many such proteins namely cyclin-cdk complexes

are also involved in apoptosis and/or DNA repair following oxidative damage. To

solidify the connection between neuronal vulnerability and cell cycle reentry there should be selective cell cycle related protein(s) that would indicate G1 entry in these neurons.

One such protein is MORF4-Related Gene on chromosome 15 (Mrg 15) that is present

transiently during G1. Its presence in AD neurons would be an independent line of

evidence for G1 entry (Figure 3.3). Furthermore, the presence of novel cell cycle regulators such as CIP-1-associated regulator of cyclin B (CARB), an indicator of G2/M

progression and p27, a negative regulator of the cell cycle, would add to the evidence of a

dysregulated entry into the cell cycle.

3.3 Methods and Materials

a. Tissue Section Preparation

Hippocampal tissue samples were obtained postmortem from patients. All clinical

and pathological diagnoses were according to standardized criteria (Khachaturian, 1985),

(Mirra et al., 1991). From the clinical reports available to us, we found no obvious

differences in agonal status or other potential confounders between the groups. Tissue

was fixed in methacarn (methanol: chloroform: acetic acid in a 6:3:1 v/v/v) or buffered

formalin for 16 h at 4°C. Tissue was subsequently dehydrated through graded ethanol and

63 xylene solutions and embedded in paraffin. Microtome sections 6 µm thick were prepared and placed on silane-coated slides (Sigma, St. Louis, MO). The cases used in each study were as follows: Mrg 15: AD cases n = 8 (age 68-95), control cases: n = 8 (ages 63–85);

CARB: AD cases n = 9, ages 59–96 years, , 2–47 hr; Control cases: n = 8, ages 62–81,

4–30 hr; p27: AD cases n = 14, ages 69-96 years, 5.5-25 hours, Control cases n = 11, ages 71-93, 6-27 hours.

b. Immunocytochemistry

Following immersion in xylene and hydration through graded ethanol solutions, endogenous peroxidase activity was eliminated by incubation of the sections of 3% hydrogen peroxidase for 30 min. To reduce non-specific binding, sections were incubated for 30 min at room temperature in 10% normal goat serum (NGS) in Tris-buffered saline

(TBS; 50 mM Tris-HCl, 150 mM NaCl, pH 7.6). After rinsing briefly with 1% NGS, sections were incubated overnight in the respective antisera (1/100) or a mouse monoclonal antibody, AT8, to altered cytoskeletal τ protein (Smith et al., 1994b). The sections were then incubated in either goat anti-rabbit or goat antimouse antisera (ICN

Biomedicals, Costa Mesa, CA), followed by species-specific peroxidase anti-peroxidase complex (Sternberger Monoclonals Inc., Lutherville, MD and ICN Biomedicals, Costa

Mesa, CA). Antibodies were localized using 3, 3 diaminobenzidine (DAB) as a chromogen (DAKO Corporation, Carpinteria, CA). To ensure the specificity of the reactivity observed, immunostaining was also performed in parallel using antibody that had been pre-incubated with the purified immunizing peptide (Bertram et al., 1999).

64 The primary antisera utilized were against the anti-chromodomain region of the

MORF4 family Mrg 15 peptide antibody (Bertram et al., 1999), CARB, p27 (antigen was against phosphorylated p27 (Thr187).

c. Immunoblots

Tissue samples were solubilized in RIPA buffer, protein concentration determined and 20 µgs of protein loaded on SDS PAGE gels. The proteins were transferred to membranes and probed with 1:1000 of our primary antibody (anti Mrg 15, CARB, or p27) and 1:2000 secondary rabbit polyclonal antibody and visualized with ECL according to the manufacturer’s instructions (Santa Cruz, Santa Cruz, CA).

Quantification of the results where needed was performed using a computer- assisted image analysis system (KS300, Zeiss). The data obtained were expressed as optical densities and analysed statistically using Students t-test.

3.4 Results

a. Mrg 15

The antisera made against the MORF4 family chromodomain region specifically

labels large pyramidal neurons in AD cases, but not age-matched controls (Figure 3.4).

There was significant overlap of the immunostaining profiles of MORF4-related protein and phosphorylated τ in neurofibrillary pathology assessed with the antibody AT8, which recognizes τ protein only when serine 202 and threonine 205 are phosphorylated. In

marked contrast, age-matched normal controls did not show any immunoreactivity. The

specificity of this select immunolocalization of MORF4- related protein was confirmed

65 by the almost complete abolition of immunostaining using antibody pre-incubated with

the immunizing peptide (data not shown).

Immunoblot analyses of homogenates of AD and control cortex reveals a single

52 kDa band in 3/3 AD cases and 1/3 control cases recognized by the antibody to

MORF4 (Figure 3.5). We determined that these bands were specifically eliminated in

immunoblot analyses performed with the antibody pre-incubated with the peptide used to

obtain the antiserum, indicating that the antisera is specific to the chromodomain region

of this family of proteins. Recently, two chromodomain proteins with molecular weights

corresponding to about 52 kDa and with homology to the peptide region against which

our antibody was raised were cloned, namely MOF (Neal et al., 2000), (Eisen et al.,

2001) and an alternatively spliced form of MSL3L1 (Prakash et al., 1999). MOF is a

histone acetyl transferase (HAT) and MSL3L1 was cloned during studies to identify the

microopthalmia gene.

b. CARB

In the brains of individuals affected by AD, CARB showed intense colocalization

with NFT throughout the sections of medial temporal tissue, including hippocampus,

subiculum, entorhinal cortex, and isocortex (Figure 3.6). Dystrophic neurites within neuritic plaques were occasionally stained. Immunostaining for CARB differed slightly from AT8 (phospho-τ) immunocytochemistry, in that neuropil threads were generally

unstained with CARB (not shown). Extracellular amyloid deposits within senile plaques

also did not stain with anti-CARB immunocytochemistry. CARB staining of

pathologically unaffected neurons, reactive astrocytes, and blood vessels with and

66 without mural amyloid deposits was present; however, the intensity of staining

approached background levels encountered in control tissue. Also noteworthy was the

intense anti-CARB immunolabeling of neuronal cytoplasmic granules of granulovacuolar

degeneration (Figure 3.6A). Not surprisingly, this latter finding was most conspicuous

within Sommer’s sector (CA-1), where granulovacuolar degeneration is generally most

pronounced (not shown). In aged-matched control cases (Figure 3.6B), CARB was found

only at background levels in similar pyramidal neurons, but CARB did label NFT in

those control cases that contained NFT. No nuclear staining was seen in either the AD

cases or the controls. Of importance is that there is no difference between AD and age-

matched control cases in cerebellum (data not shown), an area that is unaffected in AD.

Adsorption of anti-CARB with CARB protein prior to application to the section abolishes

the immunolabeling, demonstrating the specificity of the antibody (Figure 3.7A,B).

Immunoblotting revealed an anti-CARB reactive band at the appropriate molecular

weight of 27 kD in AD and control (Figure 3.8A), which further demonstrated the

specificity of the antibody. Quantification of the blot, which was normalized with the

constitutively expressed HO-2 protein, confirms our immunocytochemistry findings and

shows that CARB protein, is significantly increased more than tenfold in AD compared

with control cases (P < 0.02; Figure 3.8B).

c. p27

p27 was increased in hippocampal pyramidal neurons in AD in comparison with

control (Figure 3.9A,B). Importantly, p27 was increased in both tangle-bearing as well as histologically unremarkable neurons, indicating that this increase in p27 is not an

67 epiphenomena of neuronal degeneration. As expected for a marker of a cellular

proliferation, p27 was also found in astrocytes (data not shown). Phosphorylated p27

(Thr187) was increased in the cytoplasm of hippocampal neurons in AD compared with

control cases (Figure 3.9C,D) and was markedly increased in association with neuronal pathological alterations, including NFT, dystrophic neurites and neuropil threads. The

distribution of phosphorylated p27 (Thr187) was virtually identical to that of τ protein phosphorylated at residue 202 (serine) and residue 205 (threonine) as shown by adjacent serial sections (Figure 3.10A,B) or double-immunostaining (Figure 3.10C). No significant p27 or phosphorylated p27 (Thr187) was noted in the cerebral vasculature aside from background. To demonstrate the specificity of phosphorylated p27 (Thr187) detection, several control experiments were performed in parallel. Absorption of the phosphorylated p27 (Thr187) antibody with the immunizing peptide essentially abolished

immunostaining (Figure 3.11) whereas no effect was observed by the absorption of (i) the

antibody to phosphorylated histone H3 with phosphorylated p27 (Thr187) peptide; or (ii)

the antibody to phosphorylated p27 (Thr187) with phosphorylated histone H3 peptide

(results not shown). To exclude the possibility that the antibodies used in our experiment

were cross-reacting with other factors, especially phosphorylated τ protein, we performed

dot blot analysis to show that phosphorylated τ protein extracted from NFT in AD reacts

with antiphosphorylated τ antibodies (AT8 and PHF-1) but not with the

antiphosphorylated p27 antibody used in our experiment (data not shown). Confirming our immunohistochemical findings, immunoblot analysis revealed a major anti-p27 and antiphosphorylated p27 (Thr187) immunoreactive bands with an approximate molecular weight of 27 kDa from brain homogenates from AD and control cases (Figure 3.12A).

68 Quantification of p27 and phosphorylated p27 (Thr187) immunoblots normalized to actin

showed a significant intensity increase of both proteins in AD as compared to control

cases (P< 0.0001) (Figure 3.12B).

3.5 Discussion

a. Mrg 15

Given the scenario where a postmitotic cell is trying to re-enter the cell cycle, one

would predict that novel markers, whose presence is obligatory in the G1 phase of the cell

cycle, would also be seen in the vulnerable neurons of AD. In this study, we found that a

52 kDa protein MORF4-related cell cycle protein is associated with intraneuronal neurofibrillary pathology in select neuronal population that are vulnerable in AD, while all age matched controls have no immunostaining. Immunoblot analysis confirmed these findings although one of three control individuals did express detectable levels of the

protein, which may be an indicator of pathology that might arise in the future. Two such

proteins, 52 kD in size, have recently been cloned. One gene, MOF, encodes a HAT and

the other being MSL3L1, which was cloned during studies localizing the microopthalmia

gene. These, together with Mrg 15, all encode chromodomain regions. Chromodomains

are known to be involved in protein–protein interactions and form complexes that

remodel chromatin, making transcription factors accessible to gene promoters. Such

complexes often contain HATs. The best-studied chromodomain protein is Drosophila

msl3 (male specific lethal) that is involved in dosage compensation and up-regulation of transcription on the X chromosome in males. The chromodomain regions of msl3, Mrg

15, MOF, and MSL3L1 are very similar and the prediction is that these proteins will be

69 present in transcriptional activating complexes that most likely will include a HAT

activity. The presence of a positive transcriptional controlling protein in association with

intraneuronal neurofibrillary pathology in vulnerable neurons of AD is consistent with

the many changes in that occur and adds credence to the role of proximal

mitotic mechanisms in AD. This return to mitotic competence in AD can have profound

implications for both screening and therapeutics to modify the nature history of this

disease.

b. CARB

We show that CARB, a G2/M phase regulator, is specifically increased in the

cytosolic compartment of vulnerable neurons in AD, in particular those neurons affected

by intracellular NFT formation and granulovacuolar degeneration. Similar neuronal

populations in control cases showed only background levels of immunoreactivity.

Confirming our immunocytochemical studies, immunoblot analyses revealed statistically significantly (P < .02) higher levels of CARB in homogenates of AD compared with control brain. Given the role of CARB in cyclin B regulation and G2/M phase transition

(McShea et al., 2000a), our findings lend further credence to the hypothesis that

susceptible neurons may be arrested at the G2/M phase, which ultimately leads to cell

death. Indeed, CARB affects the steady-state levels of cyclin B1 by preventing

proteolytic degradation, and CARB overexpression leads to increased levels of cyclin B1

in CARB transfected cells (McShea et al., 2000). Therefore, the elevation of CARB

observed in AD neurons in this study may contribute to the aberrant elevation of cyclin

B1 observed in susceptible neurons in AD (Vincent et al., 1997). Insofar as translocation

70 of cyclin B into the nucleus is a prerequisite for the activation of cdc2 and the onset of the

prophase of mitosis, CARB sequestration of cyclin B in the cytoplasm likely regulates the transit through the G2/M phase of the cell cycle. Therefore, it is conceivable that elevated

levels of CARB in AD may cause excessive sequestration of cyclin B1 in the cytoplasm, as was found in many susceptible neurons in AD (Vincent et al., 1997), and contribute to

cell cycle arrest. That cyclin B1 was also found in the nucleus in some AD cases may be

due to the concurrent increase of p21 CIP-1 in AD (McShea et al., 1999b), (Engidawork

et al., 2001), which releases cyclin B1 from interaction with CARB.

Of significance is the observation that most of the CARB-positive NFT were

intraneuronal rather than extraneuronal. A similar observation was reported previously

for the active cdc2, presumably associated with cyclin B, and MPM-2-reactive phosphoepitopes (Vincent et al., 1996; Vincent et al., 1997). Given that cyclin B1 indirectly associates with CARB in a multiprotein complex (McShea et al., 2000), this

colocalization suggests that cdc2 may also be associated with CARB as part of the

multiprotein complex through the interaction with cyclin B1. Therefore, it is conceivable

that CARB can further provide docking sites to sequester active mitotic kinase (cdc

2/cyclin B1 pairs) from entering the nucleus. This not only directly leads to cell cycle

arrest but also results in the generation of MPM epitopes observed in the cytoplasm,

which may contribute to neurofibrillary tangle formation (Vincent et al., 1996; Vincent et

al., 1997). The presence of a growing number of control elements of cell cycle

machinery in degenerating neurons of AD suggests a crucial role for temporally ectopic

cell cycle reentry driving proximal pathophysiology in AD. Given that successful mitosis

has not been reported for neuronal populations in AD, it is plausible that neurons become

71 arrested at some point between G1 and M phases. Depending on where cells are arrested,

they can either return to G0 and redifferentiate or die via apoptosis. For example, if a

terminally differentiated cell, for any reason, reaches late G1 or even later, when cyclin A

is expressed, arrest leads to cell death. Therefore, the characterization of the exact phase

when arrest occurs will help to determine the mechanism by which neurons degenerate.

The presence of cyclin E/cdk2 complex indicates that neurons have passed G1.

Coordinated DNA replication suggests that the susceptible neurons may complete S phase, and the aberrant expression of cyclin B1/cdc2 complex indicates that degenerating neurons in AD may reach G2 phase. Our finding that CARB is aberrantly elevated in

degenerating neurons in AD not only supports the latter notion that degenerating neurons

in AD reach G2 phase but also suggests that neurons may arrested at G2/M phase, in that

active mitotic kinases (cdc2/cyclin B1 pair) were sequestered by CARB in cytoplasm.

Arrest at G2/M phase will inevitably lead to neuronal degeneration. What factor or

factors stimulate the neurons to reenter the cell cycle remains an important unanswered

question. Insofar as AD-specific forms of hyperphosphorylated τ are also highly

expressed in the brain during fetal life, it is notable that fetal development is under the

control of numerous growth factors, including many hypothalamic-pituitary-gonadal

(HPG) hormones. Notably, two of the HPG hormones, leutenizing hormone and activins,

become elevated during senescence because of the age-related decline in reproductive

function (Neaves et al., 1984), and leutenizing hormone is more highly expressed in AD

brain (Bowen et al., 2000) and colocalizes to the brain areas most susceptible to AD

neuropathology (Bowen et al., 2002). Activins and bone morphogenetic protein 4 bind to

the same receptors, and the latter has been shown to regulate p21 in neuronal stem cells

72 (Gomes and Kessler, 2001; Israsena and Kessler, 2002). Taken together, these findings suggest that, during senescence, changes in HPG hormone concentrations, resulting from the body’s attempt to maintain reproductive function, cause dysregulation of cell cycle events and that this cell cycle dysregulation is responsible for the neuropathological changes associated with AD.

c. p27

Both p27 and phosphorylated form of p27 (Thr187) are localized to the cytoplasm of vulnerable neurons in individuals with AD. This finding is consistent with other studies that indicate the cell cycle suppressors are up-regulated in AD neurons (Smith and

Lippa, 1995). Moreover, phosphorylated p27 (Thr187) shows a virtually identical distribution to that of phosphorylated τ, indicating a proximal role for phosphorylated p27

(Thr187) in disease pathogenesis and parallel findings made with other cell cycle markers. The functions of p27 are multifarious and occur both in the nucleus and the cytoplasm. What is interesting is that even though most tumor suppressor genes are recessive at the cellular level, p27 is haploinsufficient for tumor suppression. This means that the level of p27 is critical for tumor suppression. This is not a fit for Knudson’s two hit hypothesis. The increase in both p27 and phosphorylated p27 (Thr187) in AD neurons may reflect an effort to stop the cell cycle. Indeed, p27 can inhibit cdk4 activity at high concentrations (Olashaw and Pledger, 2002).

Additionally, Cip/Kip proteins bridge cyclin D1 and CDK in the cytoplasm. p27 also associates with mNPAP60, a nuclear pore associated protein and CRM1 the exportin both of which can lead to activation of D family of cyclins (Alt et al., 2002).

73 Additionally, p27 can have an antiapoptotic influence (Philipp-Staheli et al., 2001) and

thus be involved in cell growth and differentiation and fate.

The cytoplasmic localization of p27 gives a poorer prognosis in subtypes of breast

and colon cancer. The presence of p27 in the cytoplasm in AD can lead to its degradation

and also activation of cdk2. Indeed cdk2 is known to phosphorylate τ protein (Baumann

et al., 1993). This scenario would then lead to G1 arrest after proteolytic degradation of

p27. But there may be other mechanisms that may be involved that modulate cell death

processes possibly through the mitochondrial pathway possibly via a negative regulator

of cell death, mcl-1 (Eymin et al., 1999). Indeed, ced 3 and ced 9 homologues are found in the AD cortex (Desjardins and Ledoux, 1998). In this way the first response would be to constrain oncogenic mechanisms in AD and then later on this mechanism would finally lead to overriding of apoptosis as is seen during transformative processes. p27 can also interact with AP1, and AP1 interacts with the -203 to +55 bp of the human AβPP promoter sequence (Lukiw et al., 1994).

The cytoplasmic translocation of p27 from the nucleus inactivates p27 by sequestration from DNA targets (Reynisdottir and Massague, 1997). Therefore, p27 sequestration in the cytoplasmic compartment in neuronal populations in AD, plays a role

in may be inactive and play a role in driving mitotic re-entry and oncogenic signalling in

AD (Zhu et al., 2000). Another major mechanism to inactivate p27 occurs through phosphorylation, which can be a general which is a general signal for proteasomal degredation. Since certain parts of the proteasome pathway is impaired in AD (Master et al., 1997; Lopez Salon et al., 2000), p27, once phosphorylated, p27 may remain undegraded.

74 Although cyclin E/CDK2 complexes play a major role in the phosphorylation-

mediated degradation of p27, several different kinase signal transduction pathways are also able to phosphorylate p27, resulting in its degradation. For example, oncogenic Ras confers anchorage independence by accelerating p27 degradation through activation of

MAPK pathway (Kawada et al., 1997). In fact, p27 contains several sites that fit the minimal MAPK consensus sequence and in vitro assays have confirmed that Thr187 is a target for MAPK-induced phosphorylation and subsequent degradation of p27 (Lenferink et al., 2001). Importantly, we and others have shown that MAPK pathways are up- regulated in vulnerable neuronal populations in AD with close association to neuronal pathological alterations (Zhu et al., 2001) similar to p27, suggesting that MAPK may participate in phosphorylation of p27 at Thr187 and contribute to dysregulation of cell cycle machinery in AD.

Hence, p27 (and phosphorylated p27 (Thr187)), central regulator in cell cycle control, are localized to the cytoplasm in neuronal populations in AD. These findings suggest that the dysregulation of the cell cycle plays a crucial early role in the evolution of AD pathogenesis.

3.6 Conclusions

We have addressed a number of issues with respect to the re-emergence of cell cycle markers in AD neurons. First, a nagging critique of the presence of these numerous reentrant cell cycle markers was that they were also seen during apoptosis and during

DNA repair following oxidative stress and hence are not specific to the etiopathogenesis of the disease and may represent a secondary phenomena such as oxidative and/or

75 excitotoxic stress (Elledge, 1996; Park et al., 2000). In order to address this question, we

wanted to select a marker of G1 entry that was very specific just to G1 entry. We looked

for the presence of Mrg 15, a member of the mortality factor 4 (MORF 4) families of

proteins. This marker is present only transiently during G1 and is indicative of emergence from senescence. Indeed we did find that the vulnerable neurons in AD but not control brains, have increased MORF4-related proteins indicating re-entry into the cell cycle

(Raina et al., 2001a). We thus believe that this reentry does indeed represent exit from senescence for these vulnerable neurons.

Additionally, we found that the neuron is trying to resist the completion of the cell cycle by upregulating the inhibitors of various phases of the cell cycle. We localized

CARB as well as p27 in these vulnerable neurons. What this means in totality is that there is a significant dynamic that is going on with respect to re-entry in AD and that this dynamic is very early in the disease process because the regulation (synthesis, activation, destruction) of these mitotic regulators is very quick. Indeed this is not by happenstance that all there is activation and attempted inhibition of the cell cycle when the cell cannot go ahead with the cell cycle.

76 Table 3.1. Cell cycle markers that are seen in AD (Obrenovich et al., 2004).

MARKER ROLE IN MARKER ROLE IN CELL CYCLE CELL CYCLE

Cyclin A S to G2 PP2A or PP2B Phosphatase (Cdk5, Cyclin B Late G2 PP-1 cdc2) Cyclin C G0/G1/ lateG1 Cdc25A Phosphatase G2/M Cyclin D (D3) G to G /S 1 1 PKC/ Wnt path Translation control Cyclin E

p34cdc2/ cdk 1 Late G2/M PKA Kinase cdk2 -like kinase Cell division PKN Kinase Cdk4/Cdk6 G / G /S 1 1 PI3K Kinase Cdk5/p25/p35 G2 Nclk cdc2-like Cyclin kinase (A) AKT/PKB/RAC Kinase kinase Cdk5/p67 TGFBeta/ TAK Kinase Cdk7/MPM2 Cyclin-kinase p44/p42 MAPK Kinase Cdc42/rac G protein/cell division (ERK1/2) p21ras G protein/division p38 MAPK JNK/ MAP Kinase Mrg 15 M phase regulator (SAPK-2/3) -alpha (Stress Activated) Ki-67 LateG1,S,G2,M gamma p105/pRb G1/S TF MEK MAPK Kinase pCNA non cell-cycle specific MAPK Proline-dep kinase antigen p107/pRb Cdk2/4/6, check pt GSK-3 and Proline Dependent

c-myc S to G2 checkpoint beta Catenin- Protein Kinase (PDPK) p53/MDM2 Repressor complex P120/E-cadherin Adhesion complex ATM Check-point c-fos TF / regulator Raf/Raf-1 Check point kinase 14-3-3/14-3-3zeta Adapter protein p16INK4a CyclinD/cdk4/6inhibitors p18p15p19 (M) c-jun/p39, AP-1 TF component p27/Kip1 Cyclin H/cdk7 inhibitor Fyn Transcription Factor WAF-1/ p21/Cip1 Multi-Cyclin /cdk- Rho G-protein inihibitor (G1 and S) Rap Rab G-protein Plk1/cdc5 G2/M M check point Polo-like kinase Sos-1, Grb-2 Ex. factor, receptor

77 Table 3.2. The AD phenotype resembles a mitotic phenotype. Adapted from (Lu et al.,

2003).

Mitotic cells Adult neurons AD neurons Activation of mitotic kinases + - + Mitosis specific immunoreactivity + - + Alzheimer specific immunoreactivity + - + Phosphorylation of cytoskeletal proteins + +/- + Microtubule destabilization + - + Phosphorylation of APP + - + Amyloidogenesis + - + Availability of Pin 1 + + - Consequence Cell division Normal function Redox dysregulation No cell division Cell cycle dysregulation DNA repair defects? Oncogenic events? Tangles, Plaques, Dysfunction

78

CyclinB/CDK1

M G2 Growth factors

Rb G1 CyclinD/ CyclinA/ S CDK4,6 CDK2 R Rb-

CyclinE/CDK2

Figure 3.1. Cell cycle control overview.

79

Transcription repression Transcription activation of S-phase genes of S-phase genes

P P P

R TF R TF R + TF

p16INK4a

Mitogens CyclinD/ CyclinE/ CDK4 CDK2

DNA damage ATM P

p53 p53 p21CIP1

p19ARF mdm2

Figure 3.2A. The G1-S checkpoint control.

80

pRb cdk1 Differentiation Cdc25A Plk1/cdc5 Mitosis G1

2 cdk4 G2 cyclinD 4n S

p21-CIP1 Cyclin A C-myc

Figure 3.2B. Cell-cycle related phenomena found ectopically expressed in AD may also occur secondarily to processes which include apoptosis, trophic-deprivation and DNA repair.

81 G0 Senescence

Mitosis G1 Mrg 15 is seen expressed transiently in G1 phase of cells emerging from senescence (Smith O P 1999). 2n

G2 4n S

Figure 3.3. Is there a unique signature of re-entry into the cell cycle in the vulnerable AD neurons?

82

Figure 3.4. Antisera to the chromodomain region of MORF4-related proteins specifically recognize intraneuronal neurofibrillary pathology in cases of AD (A), whereas, conversely, there is little specific localization in control cases (B).

83

Lanes: 1 2 3 4 5 6 7 8 MW kDa

52

37

Figure 3.5. Western analysis of AD (lanes 1-3) and normal samples (lanes 4-6).

HeLa cells transfected with Mrg 15 driven by the CMV promoter demonstrate the position of the Mrg 15 protein at 37 kDa (lanes 7, 8) and the cross reactivity of the antibody with a protein in 3/3 AD cases and 1/3 control casea at 52 kDa (lane 8).

84

Figure 3.6. Immunocytochemical localization of CARB in hippocampal neurons in AD patients (A) and control patients with background levels of immunoreactivity (B). CARB is localized to NFT and senile plaques (inset), with some neurons where CARB is restricted to granulovacuolar degeneration in AD (arrows). Scale bars = 50 µm.

85

* *

Figure 3.7. Neuronal staining in the hippocampus with CARB antisera in AD (A) is completely abolished completely with CARB protein (B). Scale bar = 100 µm.

* represents blood vessel in adjacent sections.

86

A B

Figure 3.8. A) Representative results of immunoblots of cortical gray matter from AD and control patients probed with antisera against CARB show a strong band at the expected molecular weight of about 27 kDa in AD (AD) and weaker in control (con) samples. The purified recombinant CARB protein detected by the same antibody is shown as positive control (PC). B) Quantification of CARB, normalized to the

constitutive HO-2 protein, shows a significant increase of CARB intensity in AD (*P <

0.02). Result is shown as average +/-SEM.

87

Figure 3.9. p27 immunoreactivity is localized to the cytoplasm of pyramidal neurons in

AD cases (B) as compared to neurons in age-matched controls (A). Phosphorylated p27

(Thr187) is localized to the vulnerable neuronal populations in AD (D) and not in age- matched controls (C). Scale bars = 100 µm.

88

Figure 3.10. The immunohistological analysis of the distribution of phosphorylated p27

(Thr187) and phosphorylated τ protein in an AD case. Adjacent serial hippocampal sections show near complete overlap (arrowheads) between phosphorylated p27 (Thr187)

(A) and phosphorylated τ (B) profiles. Double-labeling of pyramidal neurons in the hippocampus further shows the co-localization of phospho p27 (brown) with τ immunoreactivity (blue, AT8) (C). Scale bars in B and C = 100 µm.

89

Figure 3.11 Absorption verifying the specificity of antibody binding. Immunolabeling of phosphorylated p27 (Thr187) (A) in AD neurons is almost completely abolished by absorption of the antibody with the immunizing peptide (B). The asterisk represents landmark blood vessel in adjacent sections. Scale bars = 100 µm.

90

Figure 3.12 Quantification of p27 and phosphorylated p27 (Thr187) in control and AD brains. (A) Representative results of immunoblots of cortical grey matter, probed with

antisera against p27 and phosphorylated p27 (Thr187), shows a strong band at the expected molecular weight of 27 kDa in AD (AD) (n = 6) and much weaker in control

(CON) (n = 6) samples. (B) Quantification which is normalized by an actin blot, of p27 and phosphorylated p27 (Thr187) immunoblot. Results show increase of protein intensity in AD (P < 0.0001). Results are shown as mean ± SEM.

91 3.8 Relevant Publications

1. Raina, A.K., Monteiro, M.J., McShea, A., and Smith, M.A., (1999) The role of

cell cycle-mediated events in Alzheimer’s disease. Int. J. Exp. Pathol., 80, 71-76.

2. Zhu, X., Raina, A.K., and Smith, M.A. (1999) Cell cycle events in neurons:

proliferation or death? Am. J. Pathol., 155, 327-329.

3. Raina, A.K., Takeda, A. and Smith, M.A. (1999) Mitotic neurons: a dogma

succumbs. Exp. Neurol., 159, 248-249.

4. Zhu, X., Rottkamp, C.A., Raina, A.K., Brewer, G.J., Ghanbari, H.A., Boux, H.

and Smith, M.A. (2000) Neuronal CDK7 in hippocampus is related to aging and

Alzheimer disease. Neurobiol. Aging, 21, 807-813.

5. Raina, A.K., Zhu, X., Rottkamp, C.A., Monteiro, M., Takeda, A. and Smith,

M.A. (2000) Cyclin’ toward dementia: cell cycle abnormalities and abortive

oncogenesis in Alzheimer disease. J. Neurosci. Res., 61, 128-133.

6. Raina, A.K., Pardo, P., Rottkamp, C.A., Zhu, X., Pereira-Smith, O.M. and Smith,

M.A. (2001) Neurons in Alzheimer disease emerge from senescence. Mech.

Ageing Dev., 123, 3-9.

7. Ogawa, O., Zhu, X., Lee, H.G., Raina, A.K., Obrenovich, M.E., Bowser, R.,

Ghanbari, H.A., Castellani, R.J., Perry, G. and Smith, M.A. (2003) Ectopic

localization of phosphorylated histone H3 in Alzheimer’s disease: a mitotic

catastrophe? Acta Neuropathol., 105, 524-528.

8. Raina, A.K., Zhu, X. and Smith, M.A. (2004) Alzheimer’s disease and the cell

cycle. Acta Neurobiol. Exp., 64, 107-112.

92 9. Zhu, X., Casadesus, G., Raina, A.K., Perry, G. and Smith, M.A. (2004) Neuronal

cell cycle re-entry: a doomed journal in Alzheimer disease? In Frontiers in Aging

and Neurodegenerative Disorders: Fundamental Aspects, Clinical Perspectives

and New Insights, Özben, T. and Chevion, M. (Eds.). IOS Press, Amsterdam, pp

200-206.

10. Zhu, X., Raina, A.K., Perry, G. and Smith, M.A. (2004) Alzheimer’s disease: the

two-hit hypothesis. Lancet Neurol., 3, 219-226.

11. Zhu, X., McShea, A., Harris, P.L.R., Raina, A.K., Castellani, R.J., Funk, J.O.,

Shah, S., Bowen, R., Bowser, R., Morelli, L., Perry, G. and Smith, M.A. (2004)

Elevated expression of a regulator of the G2/M phase of the cell cycle, neuronal

CIP-1 associated regulator of cyclin B, in Alzheimer disease. J. Neurosci. Res.,

75, 698-703.

12. Obrenovich, M.E., Raina, A.K., Ogawa, O., Atwood, C.S. and Smith, M.A.

(2004) Alzheimer disease – a new beginning, or a final exit? In The Cell Cycle

and Neuronal Cell Death, Copani, A. and Nicoletti, F. (Eds.), Landes Bioscience,

Georgetown, Texas, in press.

93 Chapter 4

Apoptotic Avoidance as a Feature in AD

4.1 Introduction

a. Programmed Cell Death and Apoptosis - Basic Concepts

Cell death can be classified into programmed cell death that entails a global/extrinsic program of cell death and a cell death program(s), which entails a local/intrinsic/cellular death program (Ratel et al., 2001). The latter can present itself in a

variety of phenotypes ranging from necrosis to apoptosis or even as a combination

phenotype (Sperandio et al., 2000), while the former is seen primarily during

development and presents with an apoptotic phenotype. These ideas are useful when we

come across novel phenomena or in situations where there is ambiguity as to the nature

of cell death, i.e., AD.

A cell death program can lead to the apoptotic phenotype. This is mediated by a

growing number of cysteine called caspases in a number of species. To date,

about fourteen of these proteases have been described which exist initially as zymogens

called procaspases. activity can be detected in all cells undergoing

apoptosis, regardless of their origin or the death stimulus.

Caspase classification is according to how they are activated (Figure 4.1). Those

that are activated first are known as initiator caspases and those that are activated later are

called executioner caspases. Caspases 8 and 9 are included in the former while caspases

3, 7 and 6 are included in the latter category. Although the initial enzymatic activity of a

caspase is negligible, activation confers an amplification of the catalytic activity. In brief,

94 caspase activation mechanistically, involves recruitment that then leads to local

clustering, dimerization and subsequent proteolysis including autoproteolysis, as has been

fully described by the induced proximity model (Salvesen and Dixit, 1999). Initiator caspases then proteolyze the executioner caspases. The executioner caspases then destroy specific cellular constituents and produce a signature morphology commonly called apoptosis (Figure 4.1).

The mechanisms know to date that block this caspase mediated cell death are the inhibitor of apoptosis proteins (IAP) family members. XIAP, a human form inhibits caspases 3, 7 and 9. They prevent procaspase activation and also bind to caspases and inhibit their activity. Other putative members also include cIAPs (Verhagen and Vaux,

2002). Other proteins that also prevent apoptosis belong to the Bcl-2 family (e.g., Bcl-2,

Bcl-XL, Bcl-w). These proteins act at both the mitochondrial stage (by interfering with

cytochrome c release) and also at downstream stages. The end result in this cascade includes the formation of caspase activated DNase (CAD) which causes double strand

breaks in the DNA. Additionally, Poly (ADP-ribose) polymerase (PARP) an NAD-

dependent enzyme that mediates DNA repair, is cleaved in this effector phase and thus

prevents both its nuclear entry and repair. Therefore the morphology that presents at the

end of this cell death program includes cell shrinkage, chromatin condensation, cytoplasmic blebbing, apoptotic bodies and phagocytosis of the apoptotic product leaving only normal tissue architecture.

95 b. Apoptotic Avoidance in AD

The nature of cell death in AD is such that even though the majority of evidence

points to neuronal death as the endpoint in AD, the precise death mechanisms underlying this pathway continue to elude us. The AD environment is rife of numerous presumptive apoptogenic sources that either by themselves or synergistically act to lead to cell death.

These apoptogenic stimuli include reactive oxygen species, Aβ, energy failure and HNE oxidants.

Even though internucleosomal end-labeling techniques along with DNA fragmentation have been used as absolute indicia of apoptotic cell death in AD (Anderson et al., 1996; Cotman and Su, 1996) it is apparent that DNA fragmentation also occurs secondary to oxidative stress (Tsang et al., 1996; Smith and Perry, 1997) and postmortem autolysis. Furthermore, DNA fragmentation is inferentially related to apoptosis (the phenotype of the end stage of a cell death program) whereas the hallmark signs of cell death program such as nuclear chromatin condensation and apoptotic bodies, know as apoptosis, are not seen in AD (Perry et al., 1998). Arguments have been put forth that account for the absence of apoptotic morphology in AD based upon the required rarity of these events especially given a slowly progressive disease (Allen et al., 2001; Nakagawa et al., 2000). However, the stereotypical presentation that defines the terminal phases of cell death programs, such as chromatin condensation, apoptotic bodies, and blebbing, has never been observed in AD. Added to this mechanistic and biochemical ambiguity of

neuronal death mechanisms in AD is the already mentioned temporal dichotomy between the acuteness of an apoptotic cell death program and the chronicity of AD. The end stage

phenotype (apoptosis) of a cell death program requires up to 24 hours for completion and

96 therefore, in a chronic disease like AD with an average clinical duration of almost 10 years, less than one in about 4000 cells should be undergoing apoptosis at any given time.

This means that the apoptotic events should be rare in AD in order to explain the absence of end-stage apoptotic morphology. However, the number of neurons demonstrating early apoptotic features and the amount of DNA fragmentation is extensive. Indeed, if all the neurons reported with changes that are presumed to undergo apoptosis did so, then the hippocampus, the area of the brain most vulnerable in AD, would rapidly be stripped of neurons leading to an acute onset and disease progression. This is most certainly not the case in AD where the onset is insidious and the progression is chronic. What is more astounding is that it is only the pyramidal neurons that seem to be vulnerable in AD. This greatly reduces the number of neurons that would have to die before triggering clinical symptoms in AD. It is therefore unlikely that the cell death program present in AD is apoptosis. Perhaps the greatest source of misinterpretation of data is that the criteria used for apoptosis have primarily relied on DNA fragmentation. Specifically, the laddering pattern of fragmentation does not indicate the presence of apoptosis since histones afford protection to DNA from a variety of insults including oxidative insults. Using the criterion of nuclear condensation, few neurons can be said to be undergoing apoptosis in

AD.

Additionally, since the putative apoptotic population is not synchronous, we should be able to detect a partial if not a complete spectrum of the ongoing apoptotic process at any given time, especially given the chronic nature of AD. We have no difficulty in detecting mitotic or even meiotic substages in dividing tissues. These are events that take only a few hours to complete. For these reasons the complete failure to

97 detect apoptosis especially given its signature morphology is quite puzzling. Hence,

neuronal cell death in AD suggests distinct mechanisms from the classical apoptotic

process.

The only way to address these concerns is to systematically look at the molecules

that are involved in cell death programs. Hence, we undertook a systematic study of the main proteins that are involved in the apoptotic cascade in AD by the evaluation of initiator (caspases 8 and 9) as well as the executioner proteins (caspases 3, 6 and 7) of apoptosis.

4.2 Experimental Hypothesis

In AD, if apoptosis proceeds normally, then the affected region should be stripped of neurons within weeks and AD symptomatology should ensue within months. As we know the time course of just the clinical phase of AD lasts up to a decade. Hence apoptosis as we know it (caspase dependent, signature morphology, time limited) must have been overridden in these vulnerable neurons. It is logical to assume that this override mechanism may be operative in a postmitotic niche.

If caspase-dependent apoptosis is not completed in vulnerable neurons in AD then we predict:

1) There will be a block in the activation and/or transmission of the caspase

mediated apoptotic signal within these vulnerable neurons.

2) There will be induction of anti-apoptotic proteins such as the anti-apoptotic

members of the Bcl and the IAP family.

3) There will be induction of the IAP family members.

98

4.3 Materials and Methods

a. Tissue Section Preparation

Temporal lobe samples, including hippocampal formation, subiculum, entorhinal

cortex and neocortex, were obtained postmortem from patients (n = 10, ages 72-95 years)

with histopathologically confirmed AD, as well as from younger (ages 31 and 46 years)

and aged-matched controls (n = 8, ages 56-81 years) for the caspase study and Cases used

in this study included AD (n = 23; ages 60-91 years; post-mortem interval 1-23 h),

younger control (n = 5; ages 3-56 years; post-mortem interval 2-23 h) and age-matched

control (n = 16; ages 65-91 years; post-mortem interval 3-24 h) cases based on clinical

and pathological criteria established by CERAD and an NIA consensus panel

(Khachaturian, 1985; Mirra et al., 1991).

No significant differences in agonal status between the groups were apparent from

the available medical records. Tissue was fixed in methacarn

(methanol:chloroform:acetic acid in a 6:3:1 v/v/v) or buffered formalin by immersion for

16 h at 4°C. Tissue was subsequently dehydrated through graded ethanol and xylene solutions and embedded in paraffin. Microtome sections (6 µm thick) were prepared and

placed on silane-coated slides. Tissue removed at biopsy from a diffuse large B cell

lymphoma of humans was obtained to serve as a positive control and was fixed and processed in parallel.

99 b. Immunocytochemistry

Following hydration and incubation of the tissue sections with the antisera,

sections were immunostained by the peroxidase-antiperoxidase procedure with 3,3’-

diaminobenzidine as chromogen. Adjacent serial sections were also immunostained with

antibodies to τ, rabbit antiserum and monoclonal antibody AT8 (Innogenetics,

Zwijndrecht, Belgium), to locate pathological changes. To verify the specificity of

immunolabeling, adsorption experiments were performed by incubating the antibodies of

the various caspases with their specific immunogen (StressGen Biotechnologies) at 1

µg/ml at 4°C for 16 h. Sections were pretreated with 70% formic acid to retrieve antigen.

c. Antibodies

Caspase: Rabbit antisera to caspases 3, 6, 7, 8, and 9 were obtained from

StressGen Biotechnologies Corporation, Inc. (Victoria, B.C., Canada). These antisera

recognize both active and zymogen forms of the enzymes. Antisera to the active forms of

caspases 3, 7 and 9 were also used.

Bcl-w: Immunoaffinity purified rabbit polyclonal antibody to Bcl-w (StressGen

Biotechnologies Corporation, Victoria, British Columbia, Canada), (ii) immunoaffinity purified rabbit polyclonal antibody to Bcl-w (Chemicon International, Temecula, CA,

USA) or (iii) mouse monoclonal AT8 antibody (Innogenetics, Ghent, Belgium) to phosphorylated τ protein.

100 d. Adsorption Experiments

Briefly, the purified Bcl-w immunizing peptide (StressGen Biotechnologies

Corporation), as well as an irrelevant peptide [TGF-b activating kinase 1 (TAK1) peptide

(100 lg/mL) (StressGen Biotechnologies Corporation)], was coupled to cyanogen bromide activated 4B beads (Sigma, St Louis, MO, USA) according to the manufacturer’s instructions. The immunostaining protocol was repeated using absorbed antibody produced by an overnight incubation at 4°C of primary antibody with beads coupled to the Bcl-w immunizing peptide followed by centrifugation. In parallel, as a

control against artifactual absorption, absorption of the Bcl-w antibody with beads

coupled to an irrelevant peptide (i.e. TAK1) and absorption of an irrelevant antibody (i.e.

TAK1 antibody) with beads coupled to Bcl-w peptide was performed.

e. Immunoblotting and immunodotting

Samples from gray matter of temporal cortex of AD (n = 6) and control cases (n =

6) were homogenized in 10 volumes of TBS containing 0.02% sodium azide, 0.5% sodium deoxycholate, 0.1% SDS, 1% NP-40, 1 mM phenylmethanesulfonyl fluoride

(PMSF), 1 µg/mL aprotinin and 1 µg/mL antipain (lysis buffer). Brain homogenates were centrifuged for 10 min at 16 000 g and the supernatant fluids were then transferred to new tubes. Protein concentration was determined by the bicinchoninic acid assay (BCA Kit) method (Pierce, Rockford, IL, USA) and an equal amount of protein from each sample was loaded. Proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electroblotted onto Immobilon-P (Millipore, Bedford,

MA, USA) by standard procedures as previously described. Blots were incubated

101 sequentially with blocking agent (10% nonfat milk in 0.1% Tween-20 in TBS), rabbit

anti-Bcl-w antibody (1: 200, StressGen Biotechnologies Corporation) and affinitypurified goat anti-rabbit immunoglobulin peroxidase conjugate preabsorbed to eliminate human cross-reactivity. Blots were developed by the ECL technique (Santa Cruz Biotechnology,

Inc., Santa Cruz, CA, USA) according to the manufacturer’s instructions. Parallel blots were probed with antibodies against actin or HO-2 (1:500, StressGen Biotechnologies

Corporation), which are constitutively expressed in neuronal cells, as a control to demonstrate equivalent loading.

Blots were scanned at high resolution and the immunoreactive bands were quantitated with KS300 image analysis software (Zeiss, Thornwood, NY, USA) as follows. The optical density was determined for the entire band in each case by outlining the immunoreactive area and subtracting the background density of the surrounding unlabeled blot area with similar size. The values for the AD and control cases were averaged and the Student’s t-test used to determine whether the differences were significant (p < 0.05). Dot blots were prepared by applying 5 µg of τ protein (0.5–1.0

mg/mL), immunizing peptide of Bcl-w (1 mg/mL), 5 µg of insoluble PHF or a control

fraction directly onto Immobilon (Millipore) membrane and then air-dried. Human s from

a normal human brain and PHF-enriched fractions from AD brain were prepared by

previously described methods (Takeda et al. 2000). The membrane was incubated

sequentially with blocking agent (10% non-fat milk in TBS-Tween), rabbit anti-Bcl-w

and goat anti-rabbit peroxidase conjugate. AT8 antibody that detects PHF-τ and 5E2

antibody that detects total s were used as a positive control, and rabbit anti-Ras was used

102 as a negative control. Dot blots were developed using ECL technique (Santa Cruz

Biotechnologies).

f. Electron Microscopy

Brain tissue was obtained at autopsy and fixed in 2% paraformaldehyde/ 0.5% glutaraldehyde and sectioned at 60 µm using a Vibratome. Sections were rinsed with

PBS, incubated in 10% NGS for 1 h, followed by incubation with antibody against Bcl-w for 16 h at 4°C. After rinsing with 1% NGS, goat-anti-rabbit conjugated to 17 nm colloidal gold was applied. After incubating for 4 h, sections were rinsed overnight in

PBS, followed by fixation in 2.5% glutaraldehyde for 1 h and rinsed with PBS. After incubation in 1% osmium tetraoxide for 1 h, sections were dehydrated with acetone and embedded in Spurr embedding media (Electron Microscopy Sciences, Hatfield, PA,

USA). Ultrathin sections were stained with uranyl acetate and lead citrate, and examined in a Zeiss CEM 902 electron microscope. The primary antibody was omitted on one section as a negative control.

g. Cells

The human neuroblastoma cell line M17 was maintained in serum free Opti-

MEM media (Life Technologies, Gaithersburg, MD, USA) supplemented with 5% donor calf serum and 1% penicillin/ streptomycin with fungizone (Life Technologies). Cells were serum-starved overnight before Aβ treatment. Preparation of fibrillar Aβ1-42

(fAβ). Human Aβ peptide (Aβ1-42) was synthesized and purified by HPLC, and characterized by amino acid analysis and mass spectroscopy by W. M. Keck Foundation

103 Biotechnology Resource Laboratory (Yale University, New Haven, CT). The peptide was identified as a single peak upon HPLC and showed no chemical modification. A 1 mM stock solution of Aβ1-42 in deionized water was allowed to fibrillize for 7 days at 37°C. fAβ was stored at -80°C until use.

h. Determination of Cell Death

A lactate dehydrogenase (LDH) release assay (Roche Molecular Biochemicals,

Indianapolis, IN, USA), combined with immunoblot analysis of cleaved caspase 3 and cleaved PARP were used to determine cell death and apoptosis, respectively. For immunoblot, after each treatment, both detached and attached cells were spun down, harvested together and washed with PBS, then lysed with lysis buffer (Cell Signaling

Technology, Beverly, MA, USA). Antibody against cleaved caspase 3 (Cell Signaling

Technology) only recognizes cleaved casapse 3 at 17 kDa, while antibody against human

PARP (Pharmingen) recognizes both full length and cleaved PARP. The LDH concentration in the media was determined according to manufacturer’s instructions.

i. Transfection and Selection

Plasmids containing the cDNAs encoding Bcl-w were kind gifts from Dr Suzanne

Cory (Gibson et al. 1996). M17 cells were grown in OPTI-MEMI supplemented with 5%

DCS and 1% penicillin/ strptomycin. After serum-starving overnight, they were transfected with 5 µg of mammalian expression vectors containing Bcl-w using Superfect transfection reagent (Qiagen, Valencia, CA, USA). Stablytransfected cells were selected in 600 µg/mL G418 (Life Technologies) for 4 weeks. Pools of 50 or more colonies were

104 used to avoid bias of single cell colonies. Protein expression of Bcl-w was confirmed by

immunoblot and RT-PCR.

j. RT-PCR

Total RNA was extracted from cells using RNeasy Midi kit (Qiagen). cDNAs

were prepared with random hexamers (Firststrand cDNA synthesis kit; Amersham

Pharmacia Biotech, Piscataway, NJ, USA) and used as templates for subsequent PCR.

For Bcl-w cDNA amplification, samples were heated at 94°C for 5 min and then subjected to thermocycling (30 cycles of 1 min at 94°C, 1 min at 55°C and 1 min at

72°C) using the specific primers 5’-TTATGTCTGTGGAGCTGGCC-3’ and 5’-

TCTCCAGGTAGGCCACCATC-3’ (Kitamura et al. 2000). The PCR products were separated on a 2% agarose gel in 1% Tris borate-EDTA buffer and visualized with UV light after ethidium bromide staining.

4.4 Results

Sections of medial temporal lobe from patients with AD demonstrated numerous neuritic plaques and NFT throughout the temporal neo- and allocortices. The younger controls showed no pathology, while the age matched controls showed only isolated NFT in the entorhinal pre-α layer in one of the cases.

a. Upstream Caspases

Caspase 8 immunohistochemistry demonstrated strong immunolabeling in nearly all of the NFT in each of the AD cases, with no differences in regional distribution within

105 the medial temporal lobe (Figure 4.2A). Non-NFT-containing neurons in AD cases

(Figure 4.2A) and neurons in control cases (Figure 4.2B) showed only weak cytoplasmic staining that was similar between AD cases and controls. Double-immunostaining using antibodies to and τ (AT8) further demonstrated caspase 8 labeling of a subset of neuropil threads, while dystrophic neurites showed caspase 8 staining only rarely. No staining of amyloid cores of cored plaques, or blood vessels involved by amyloid angiopathy could be discerned.

NFT-containing neurons also demonstrated caspase 9 immunoreactivity (Figure

4.2C), while, again, the level of caspases in neurons not containing NFT was similar between AD and control cases (Figure 4.2D). was found in association with senile plaques in almost all cases of AD (Figure 4.2E) and was present at low levels in neurons in both AD and control cases (Figure 4.2F).

b. Downstream Caspases

Downstream caspases 3 and 7 showed weak, uniform immunolabeling of neuronal cytoplasm in both AD and control cases (Figure 4.3A-D). Two out of ten AD cases also showed immunolabeling of isolated NFT and neuritic plaques (less than 1% of total tangles and plaques).

c. Bcl-w

Bcl-w was increased and associated with pathological alterations in AD. In situ localization of Bcl-w protein was performed on hippocampal and cortical sections from patients with AD and from cognitively normal age-matched individuals, using two

106 polyclonal antibodies against Bcl-w from different sources, and identical results were found. In 14 out of 16 age-matched control cases and all younger control cases, Bcl-w was found at background low levels in the cytoplasm of neurons (Figure 4.4A). However, conversely, Bcl-w was markedly increased in large pyramidal neurons in the hippocampus of all individuals with AD, i.e. in those neurons vulnerable to degeneration during the disease (Figs. 4.4B-E). Most strikingly, in AD, Bcl-w was increased in association with intracellular NFT, neuritic senile plaques and neuropil threads, i.e. the

classic neuropathology of the AD brain (Figure 4.4B), as well as with punctate structures

(Figure 4.4C and at higher magnification in Figure 4.4D). Extracellular NFT are also

occasionally stained. Indeed, there was an obvious overlap of the immunostaining

profiles of Bcl-w and τ-positive neurofibrillary pathology in AD cases as assessed by

staining of adjacent sections (result not shown) and double staining (result not shown)

using AT8, an antibody which only recognizes τ protein when serine 202 and threonine

205 are phosphorylated. Quantification revealed that neurons are either Bcl-w positive or

both Bcl-w and AT8 positive, with no neurons being only stained by AT8. As often

found in normal aging, a few pyramidal neurons contained NFT and in these cells, Bcl-w

is increased and also localized to tangles. Importantly, in the cerebellum, an area that is

unaffected by AD, there was no difference in the staining pattern between AD and

control cases (result not shown). Extending the light level studies, electron microscopy

showed that Bcl-w was localized to the paired helical filaments in NFT (Figure 4.5a, see

figure legend) and to dystrophic neurites (Figure 4.5B). Additionally, in neurons

containing neurofibrillary pathology, Bcl-w was also found localized to mitochondrial

structures (Figure 4.5C). In the control section, in which the primary antibody was

107 omitted, no gold particles were detected (data not shown). To confirm the specificity of

Bcl-w immunocytochemistry, several control experiments were performed in parallel.

Absorption of the Bcl-w antibody with the immunizing peptide of Bcl-w, conjugated with

Sepharose Cl-4B beads, almost completely abolished immunostaining. No effect was observed by the absorption of: (i) antibody to TAK1 with Bcl-w peptide (free or conjugated); (ii) antibody to Bcl-w with sepharose Cl-4B beads conjugated to TAK1 peptide; or (iii) antibody to Bcl-w with amyloid b protein precursor (results not shown).

To exclude the possibility that the Bcl-w antibody was crossreacting with s protein or

PHF-s, no immunoreactivity was found on dot blots of enriched human s protein from normal patients or of purified PHF from AD brain (result not shown). Furthermore, there was no significant between Bcl-w and τ protein (Meg Align,

DNAStar).

Immunoblot analysis of cortical brain homogenates revealed a major band at around 21 kDa corresponding to the known molecular weight of Bcl-w (Gibson et al.,

1996). In accordance with our immunocytological findings, this band was much less intense in the control brain homogenates when compared with that from AD cases

(Figure 4.6A). Statistical analysis, normalized to actin level, showed an increase greater than twofold in the expression of Bcl-w in AD compared with age-matched control cases

(p < 0.05) (Figure 4.6B). Exposure to fibrillized Aβ up-regulates Bcl-w protein expression. Aβ, a critical protein in the pathophysiology of AD, has been demonstrated to be neurotoxic in cell cultures by an apoptotic mechanism. Therefore, given the elevated levels of Bcl-w in susceptible neurons in AD and the neuroprotective role that Bcl-w may play in vivo, we explored the role that Bcl-w may play in response to Aβ stress.

108 Cultured M17 human neuroblastoma cells were exposed to fibrillized Aβ1-42 (f

Aβ) and Bcl-w expression was analyzed by immunoblot experiments. After 24 h of treatment with fAβ peptides, expression levels of Bcl-w protein were examined. A 24 h exposure of M17 cells to various concentrations of fAβ1-42 (0.1-10 µM) led to significantly increased expression of Bcl-w protein compared with vehicle-treated controls (Figure 4.7).

To investigate the role of Bcl-w in Aβ neurotoxicity and apoptotic cell death in susceptible neurons in AD, we generated M17 cell lines overexpressing Bcl-w by transfection with the plasmid pcDNA-Bcl-w and subsequent selection of stable transfectants. M17 cells stably transfected with Bcl-w construct expressed higher levels of Bcl-w mRNA expression (Figure 4.8a) compared with cells transfected with the empty vector (mock). The effect of Bcl-w overexpression on the induction of neuronal death was first investigated by exposing cells to apoptosis-inducing agent, staurosporine (STS).

Exposure to 100 nM STS for 24 h led to more than 90% cell detachment and death in non-transfected and mock-transfected cells but only around 15% detachment in Bcl-w- transfected cells. Activation of caspases, reflected by the cleavage of caspase 3 and its substrate, PARP, were determined in cell lysates prepared from STS-exposed M17 cell cultures (Figure 4.8b). Exposure of mock-transfected M17 cells to 100 nM STS led to a significant increase in cleaved caspase 3 and PARP products (Figure 4.8b). In contrast, cleavage products in Bcl-w-transfected M17 cells exposed to STS were significantly decreased compared with mock-transfected cells (Figure 4.8b). The effect of Bcl-w overexpression on fAb-induced neuronal death was also investigated. Cell death after exposure to 1 µM and 10 µM fAβ was quantified as LDH release. At a sub-lethal

109 concentration of 1 lM fAβ, no significant cell death was induced in mock- or Bcl-

wtransfected cells. However, a 48 h exposure of 10 µM fAβ caused significant cell death in mock-transfected cells (Figure 4.8c). In contrast, exposure to 10 µM fAβ failed to induce a significant toxicity in cells overexpressing Bcl-w (Figure 4.8c). To investigate the activation of caspases, the levels of cleaved caspase 3 and its substrate, PARP, were determined in cell lysates prepared from fAβ-exposed M17 cell cultures (Figure 4.8d).

Cell lysates from vehicle-treated cultures showed a negligible cleavage rate of both caspase 3 and PARP (not shown). Exposure of mock-transfected M17 cells to 10 µM Aβ led to a significant increase in cleaved caspase 3 and PARP products. In contrast, cleavage products in Bcl-w-transfected M17 cells exposed to 10 µM Aβ were not significantly higher than the baseline activity of vehicle treated controls.

4.5 Discussion

a. Caspases

In an attempt to resolve the issue of programmed cell death and apoptosis in AD, in this study we systematically investigated the entire apoptotic cascade. Our results indicate that, while upstream caspases including 8 and 9 are clearly elevated in NFT- containing neurons in AD, downstream caspases, or so called effector caspases, such as 3 and 7, remain at control levels. While caspase 6, which likewise has been described as an effector caspase, is also found in AD neurons, its localization differs significantly from the upstream caspases. Indeed, the restriction of caspase 6 to senile plaques, and not neurons, suggests that extracellular pathophysiology (i.e., Aβ deposition as senile plaques) differs from the neuronal pathophysiology in AD. Although all of the antibodies

110 used in our study recognize both zymogen and cleaved caspase forms, the results

presented are significant given that the upstream caspases are localized specifically to

susceptible neurons in AD. Additionally, given that clustering of zymogen caspase is

sufficient for substrate proteolysis (Salvesen and Dixit, 1999), the condensed localization

we observe here may be an indicator of active enzymes. However, the lack of effector

caspases clearly indicates a lack of propagation of the initial apoptotic signal. The

presence of caspase 8, a downstream component of the Fas-FADD pathway and a

specific activator of caspase 3 (Stennicke et al., 1998), along with caspase 9, which

through the down-stream mitochondrial pathway also leads to caspase 3 activation,

argues for the recruitment of the initiator components of the caspase pathway in neurons

of AD. However, the lack of increased caspase 3 and 7 in the neuropathology of AD

(Selznick et al., 1999) and the sporadic incidence of other crucial downstream events of the caspase cascade (Cohen, 1997) indicates incomplete or effectively absent

amplification of the upstream apoptotic signal. Indeed, since such downstream caspases

and their proteolytic products are recognized as markers of apoptotic irreversibility

(Trucco et al., 1998), their avoidance or sporadic appearance in AD indicates an absence

of effective distal propagation of the caspase-mediated apoptotic signal(s). Indeed,

caspase 3 and 9 activity is important pro-apoptotic regulators of postmitotic neuronal

homeostasis (Roth et al., 2000).

Recently, activated caspase 3 has been shown to be present within autophagic

granules and rarely within neurons in AD and Down syndrome cases (Selznick et al.,

1999), (Stadelmann et al., 1999). However, the localization of caspase 3 activity, like that

of caspase 6 in our study, within a select subcellular compartment, i.e., autophagic

111 granules, does not necessarily indicate global, effective caspase amplification. Moreover,

modulation of distal substrates of activated caspase 3 may lead to further modification of

this cell death pathway and may explain the lack of an apoptotic death pathway. This may

explain the lack of evidence demonstrating the acute end stages, including nuclear

compensation and blebbing, in AD susceptible neurons (Perry et al., 1998).

Apoptosis can be prevented by anti-apoptotic members of the Bcl-2 family, many

of which have been reported in AD, including Bcl-XL (Pike, 1999) and Bcl-w (Zhu et al.,

2004b). Additional evidence for the antiapoptotic nature of the intraneuronal environment

in AD includes the expression of GADD 45, a growth arrest DNA damage-inducible protein in select neurons in AD where its early expression is associated with the expression of Bcl-2. Thus, this may in turn confer survival advantages in AD.

Furthermore, the hyperphosphorylated intraneuronal state that exists in AD acts to inhibit downstream substrate proteolysis and thus can promote neuronal survival. Indeed, accumulating evidence points to dephosphorylation being associated with increased capability to cleave PARP (Martins et al., 1998). Additionally, chronic oxidative stress, a major component of early AD pathophysiology, would further inhibit downstream propagation of caspase-mediated apoptotic signals (Hampton et al., 1998).

The upregulation of individual caspases, while seemingly not leading to apoptosis, could have great significance related to the pathogenesis of AD. For example, caspase 6, found here in association with Aβ senile plaques, cleaves AβPP resulting in a

6.5 kDa amyloid fragment and is proposed as an alternate AβPP processing pathway

(LeBlanc et al., 1999). The lack of caspase 6, a downstream effector caspase, in neuronal pathology, however, argues against a specific role in the apoptotic cascade. Given that

112 execution of apoptosis requires amplification of the caspase-mediated apoptotic signal;

our results indicate that in AD there is a lack of effective apoptotic signal propagation to

downstream caspase effectors. AD represents that first in vivo situation reported in which the initiation of apoptosis does not directly lead to apoptotic cell death, an interpretation consistent with earlier reports (Lucassen et al., 1995), (Sheng et al., 1998).

Naturally, other mechanisms do exist for apoptotic cell death to occur in other tissues. This includes the family of proteases and granzyme b. The role of calpain family of calcium proteases during acute injury in the nervous system has been well known (Nixon et al., 1994). Additionally, granzyme perforin pathway again leads to apoptotic cell death. The presence of p25 which is produced from the cleavage of p35 a

Cdk5 activator, leads to prolonged activation of Cdk5 (Tseng et al., 2002) confirms that the calpain system in the setting of an adult neuron has more to do with alternative functions such as cell cycle regulation. Neither of these pathways has been shown to specifically activate cell death programs in AD. Additionally, similar time arguments apply here also. Even if both of these systems were in effect in the AD environment in their apoptotic role, then the time course of AD would contract and symptomatology would evolve within weeks if not months. This is not the picture that presents itself in

AD.

b. Bcl-w

We demonstrated an increased level of Bcl-w protein in the susceptible neuronal populations of AD brain compared with age-matched controls. While the mechanism of activation of Bcl-w is poorly understood, Bcl-w translocates to the mitochondrial

113 membrane following ischemia and inhibits cytochrome c release, presumably by forming heterodimers with Bax (Yan et al. 2000b). The increased Bcl-w protein with mitochondrial localization in susceptible neurons found in our study may therefore indicate an active role for Bcl-w in these cells in response to apoptotic stimuli found in

AD. In this regard, we also demonstrated that exposure of cultured M17 human neuroblastoma cells to Ab increased the expression of Bcl-w, and that overexpression of

Bcl-w protected neurons against apoptosis induced by Aβ. It is therefore conceivable that the increased Bcl-w expression in response to Aβ is a stress response that may help to protect neurons against cell death. Given that Aβ plays a critical role in AD pathogenesis, the increased expression of Bcl-w in AD suggests that a similar stress response may also play a role in vivo and lead to neuronal survival. Strikingly, our studies show, for the first time, that the pro-survival member, Bcl-w, is not only increased in AD but is also intimately associated with the lesions of the disease, i.e., NFT, senile plaque neurites and neuropil threads. This is in marked contrast to other survival factors such as Bcl-2 and

Bcl-XL, which, while up-regulated in AD, show no clear association with lesions. In fact,

Bcl-2 is elevated in neurons without pathology and decreased in neurons with neurofibrillary pathology (Satou et al. 1995; O’Barr et al. 1996; Tortosa et al. 1998), which was suggested to make the neurons with neurofibrillary pathology particularly vulnerable, especially in light of the elevation of Bax, a pro-apoptotic regulator, in such neurons (MacGibbon et al. 1997; Su et al. 1997). However, this must be viewed in contrast with studies showing that activation of caspase 3 occurs at a much lower level

(less than one in a thousand) (Stadelmann et al. 1999) and that neurons with NFT survive for decades (Morsch et al. 1999), indicating that these neurons are probably mobilizing a

114 protective mechanism. Although it is not clear whether Bcl-w protein associated with

neuro-fibrillary pathology is functional or not, our study at least indicates that these

neurons significantly up-regulate the prosurvival Bcl-w, part of which translocates to

mitochondria membrane and may serve a neuroprotective role. Preliminary studies show

that Bcl-w is also increased and localized to neurofibrillary pathology in other

tauopathies, such as progressive superanuclear palsy (PSP) and Pick’s disease (results not

shown), but not in Parkinson’s disease and Diffuse Lewy Body disease (DLBD),

suggesting that Bcl-w may play an important neuroprotective role in those neurofibrillary

pathology-containing neurons. Indeed, it is likely that Bcl-w may, in part, serve as a

compensatory protective mechanism for neurons assaulted by apoptogenic stimuli in AD

and contribute to their survival. Such an assertion is further strengthened by the finding

that Bcl-w plays an important role in determining neuronal cell survival after cerebral

ischemia or induced seizure (Minami et al., 2000). In this regard, it is interesting to note

that, compared with Bcl-XL, Bcl-w becomes more important in regulating sensory neuron survival with age (Middleton et al., 2001). While Bcl-w knockout mice showed no abnormalities in brain (Print et al., 1998), this does not exclude a role for Bcl-w in stressed or age-related conditions because the oldest mice examined were only 1 year old, corresponding to a human age of around 40 years, at which time point there are few AD- related alterations (Nunomura et al., 2000). It will be of interest in future studies to determine whether older Bcl-w knockout mice show AD-related changes.

In conclusion, Bcl-w, a pro-survival, anti-apoptotic factor, is upregulated in AD and intimately associated with neurofibrillary pathology. Exposure to fibrillized Aβ leads to increased Bcl-w protein levels, and overexpression of Bcl-w results in the protection

115 against Aβ- and STS-induced apoptosis. These findings indicate that Bcl-w may play an

important protective role in neurons in the face of the various pro-apoptotic signals present in the brains of individuals with AD. How Bcl-w exerts its protective role in

susceptible neurons certainly merits further investigation and may reveal novel

neuroprotective strategies.

4.6 Conclusions

Obviously, in certain brain areas in AD, while many neurons do degenerate, it is

unclear whether this is through successful propagation of the apoptotic cascade or by

another pathway such as paratosis (Sperandio et al., 2000). However, in those surviving

neurons, it is clear that neuronal viability in AD is, in part, maintained by the lack of

distal transmission of the caspase-mediated apoptotic signal(s). This novel phenomenon,

which we term apoptotic avoidance or abortosis, represents an exit from the caspase-

induced apoptotic program that, given the robust survival of abortotic neurons with NFT

(Morsch et al., 1999), leads to prolonged neuronal survival in AD.

116

Receptor-mediated Cascade ( FAS)

Caspase 8 Initiator Caspases Caspases 3 and 7 Effector Caspase 9 Caspases

Apoptotic Morphology

Cell Death Mitochondrial-mediated cascade

Figure 4.1. An overview of the apoptotic process.

117

Figure 4.2. Activated upstream caspase 8 (A) and caspase 9 (C) are associated with NFT in Sommer’s sector (CA-1) neurons of AD cases. Control cases showed no histopathological abnormalities and no induction of caspases 8 (B) and 9 (D) in such neurons. Caspase 6 associates to neocortical plaques in AD (E), but showed no increase in control cases (F).

118

Figure 4.3. Activated downstream caspases 3 (A, B) and 7 (C, D) are not increased in

AD cases (A,C) in comparison to controls (B,D). A-D Sommer’s sector (CA-1) of hippocampus are depicted. Bar 100 µm

119

A B FF

C D E G

Figure 4.4. Immunocytochemical localization of Bcl-w reveals diffuse and low levels of

Bcl-w in hippocampal neurons from control cases (A) but, by marked contrast, Bcl-w is

increased in AD and localized to NFT and dystrophic neurites (B) as well as punctate neuronal staining (C). A representative neuron containing granular structure (D) and a representative neuron with increased Bcl-w but without pathology at a higher magnification (E) are shown. Neuronal immunostaining in the hippocampus with Bcl-w antibody in AD (F) is completely abolished by absorption with immunizing peptide (g) but not by irrelevant peptide (not shown). *Indicates landmark blood vessel in adjacent sections in f-h. Scale bars: A,B,C,F,G = 50 µm, D,E = 100 µm.

120

A B C

Figure 4.5. Immunogold labeling of Bcl-w in AD hippocampus is detected in the paired helical filaments of NFT and in the dystrophic neurites: (A) longitudinal section; (B) cross section. Additionally, Bcl-w is also detected in mitochondrial membrane (C). Scale bar: A = 1 µm; B = 0.5 µm, C = 0.1µm.

121 A B 66 kDa * 3 45 kDa 2.5 29 kDa 2 Bcl-w 1.5

1

0.5

Actin immunodensityRelative (%) 0 Control AD (n=6) (n=6) C1 C2 AD1 AD2 AD3

Figure 4.6. (A) A representative result of immunoblots of cortical gray matter, homogenized in lysis buffer and probed with antisera against Bcl-w, shows a strong band around the expected molecular weight of 21 kDa in AD and weaker in control samples.

(B) Quantification of the Bcl-w reactive band, normalized to actin levels, shows that Bcl- w is significantly increased in AD (*p < 0.05).

122 A Bcl-w

HO-2 B 3 300% 5 * 250% * 200%2 150%5 100%

Relative intensity Relative 50%5 0%0 0 0.1 1 10 Fibrillized amyloid-ß conc. (µM)

Figure 4.7. Exposure to fibrillized Aβ1-42 increases Bcl-w protein expression in M17 human neuroblastoma cells. (A) A representative immunoblot analysis. (B)

Quantification of the intensity of Bcl-w bands, normalized with constitutively expressed

HO-2 levels and setting the intensity of controls as 100%. The experiments were repeated four times with comparable results (*p < 0.05).

123

Figure 4.8. (a) Characterization of the Bcl-w transfectants by RT-PCR confirms the overexpression of Bcl-w in Bcl-w-transfected M17 cells. (b) Bcl-w-transfected cells are resistant to staurosporine-induced cell death. (c,d) Bcl-w-transfected cells are resistant to fAb-induced cell death. (c) LDH release assay; (d) Immunoblot with apoptosis markers.

The arrow points to the cleaved PARP. Experiments were repeated three times with comparable results (*p < 0.05).

124 4.8 Relevant Publications

1. Raina, A.K., Hochman, A., Zhu, X., Rottkamp, C.A., Nunomura, A., Siedlak,

S.L., Boux, H., Castellani, R.J., Perry, G. and Smith, M.A. (2001) Abortive

apoptosis in Alzheimer’s disease. Acta Neuropathol., 101, 305-310.

2. Zhu, X., Raina, A.K., Rottkamp, C.A., Aliev, G., Perry, G., Boux, H. and Smith,

M.A. (2001) Activation and redistribution of c-Jun N-terminal kinase/stress

activated protein kinase in degenerating neurons in Alzheimer’s disease. J.

Neurochem., 76, 435-441.

3. Raina, A.K., Sayre, L.M., Atwood, C.S., Rottkamp, C.A., Hochman, A., Zhu, X.,

Obrenovich, M.E., Shimohama, S., Nunomura, A., Takeda, A., Perry, G. and

Smith, M.A. (2002) Apoptotic and oxidative indicators in Alzheimer disease. In

Neuromethods, Vol. 37, Apoptosis Techniques and Protocols, LeBlanc, A.C.

(Ed). Humana Press Inc., Totowa, New Jersey, pp 225-246.

4. Raina, A.K., Rottkamp, C.A., Zhu, X., Ogawa, O., Hochman, A., Shimohama, S.,

Takeda, A., Nunomura, A., Perry, G. and Smith, M.A. (2002) Neuronal survival

and death in Alzheimer disease. In Advances in Behavioral Biology, 51, Mapping

the Progress of Alzheimer’s and Parkinson’s Disease, Mizuno, Y., Fisher, A. and

Hanin, I. (Eds). Kluwer Academic/Plenum Publishers, New York, pp 49-57.

5. Raina, A.K., Hochman, A., Ickes II, H., Zhu, X., Ogawa, O., Cash, A.D.,

Shimohama, S., Perry, G. and Smith, M.A. (2003) Apoptotic promoters and

inhibitors in Alzheimer’s disease: who wins out? Prog. Neuropsychopharmacol.

Biol. Psychiatry, 27, 251-254.

125 6. Raina, A.K., Zhu, X, Shimohama, S., Perry, G. and Smith, M.A. (2004) Tipping

the apoptotic balance in Alzheimer’s disease: the abortosis concept. Cell Biochem.

Biophys., 39, 249-255.

7. Zhu, X., Wang, Y., Ogawa, O., Lee, H.G., Raina, A.K., Siedlak, S.L., Harris,

P.L.R., Fujioka, H., Shimohama, S., Tabaton, M., Atwood, C.S., Petersen, R.B.,

Perry, G. and Smith, M.A. (2004) Neuroprotective properties of Bcl-w in

Alzheimer disease. J. Neurochem., 89, 1233-1240.

126 Chapter 5

Oncogenic Parallels in AD

5.1 Introduction

As outlined in the preceeding three Chapters, looking broadly at the early AD environment we see many parallels with early processes that give birth to oncogenic events (Figure 5.1). What is clear is that in both disease processes share a dysregulated cell cycle, imbalance in redox homeostasis and an overriding of apoptosis.

a. Oxidative Stress, AD and Oncogenesis

As outlined in Chapter 2, in AD, oxidative stress is one of the earliest known pathological phenomenon (Smith, 1998). This oxidative pathophysiology is also paralleled during the birth of oncogenic processes (Seril et al., 2003; Levine, 1997). Free radical generators are known to convert benign papillomas into carcinoma (Athar, 2002).

Dietary iron supplementation enhances carcinoma development in various carcinoma models (Seril et al., 2002). This effect is due to an iron-catalyzed increase in oxidative damage. Iron supplementation was also associated with a rise in nitrotyrosine expression.

The role of oxidative damage certain carcinogenesis models was also suggested by the effectiveness of the antioxidant N-acetylcysteine (NAC) in inhibiting carcinoma development and nitrotyrosine-positive cell number (Seril et al., 2002). Indeed with the presence of oxidative stress in neoplastic cells, the inhibition of SOD and other antioxdants forms an actively investigated method for novel antineoplastic chemotheraputic selectivity (Hileman et al., 2004). However, the most important link

127 between these to diseases states may involve dysregulation of DNA repair. In searching for a connection between oxidative stress and its sequelae in AD, we decided to investigate factors involved in the repair of oxidative lesions to DNA and try to determine whether they were involved in AD pathogenesis. In this search, we came upon BRCA-1 which was first identified as the gene that is mutated in approximately 50% of familial breast and ovarian cancer cases. BRCA-1 has a large number of physiologic roles including but not limited to transcriptional control, cell-cycle checkpoint control and

DNA damage repair. The possible role of BRCA-1 in DNA repair was first suggested after the observation that BRCA-1 is phosphorylated in response of cellular stresses including oxidative stress and ionizing radiation (Gilmore et al., 2003). Furthermore, the hyperphosphorylated BRCA-1 is relocated to sites of active DNA repair, including replication forks (Scully et al., 1997), (Thomas et al., 1997). Additionally, BRCA-1 has been shown to be associated with a large number of proteins known to be involved in the

DNA repair machinery. These include a number of those known to be involved in transcription coupled repair (TCR), including RNA polymerase II and the transcriptional cofactors TFIIF, TFIIE and TFIIH (Scully et al., 1997). Hence for these reasons alone it is logical to look at the presence of BRCA-1 in AD.

b. Dysmitotic Mechanisms, AD and Oncogenesis

As outlined in Chapter 3, cell cycle dysfunction is a key pathogenic event in AD.

In a given local environment, the cell division cycle possesses unique spatial and temporal controls. This is especially true in the environment where cells have acquired specialized sets of functions. A cancer develops when there is growth that is not restricted

128 in a specific spatial and temporal area. Postmitotic neurons are associated with an escape from sensitivity to mitotic signaling. What is unclear is the selective necessity for this mechanism. Indeed there are no primary or secondary (metastatic) tumors of the brain that arise from or affect the primary neurons. Therefore, any re-emergence of sensitivity to signals, i.e., neurotrophic factors, may initiate an attempt to re-enter into the cell division cycle, with progression being limited by the degree of mitotic competence. The reappearance of these cell cycle markers especially of those that imply re-entry into the

G1 phase of the cell cycle in the neurons that are vulnerable to dysfunction and death in

AD seems akin mechanistically to the processes that lead to cancer in tissues that normally proliferate. Successful dysregulation of the cell division cycle equilibrium fulfills both the sufficient and necessary criteria for the initiation of oncogenic transformation (Raina et al., 1999a).

c. Loss of Apoptosis, AD and Oncogenesis

As outlined in Chapter 4, in AD, neurons do not die by traditional apoptotic mechanisms. In oncogenesis, there are two events that need to take place for cells to grow without restraint. One is that there should be some relaxation of the brakes that are normally part of the cell cycle (see Chapter 3 for further discussion). The other is that the cell should be to some degree refractory to apoptosis (see Chapter 4 for further discussion). Having both of these events taking places leads a cell to propagate with little if any hindrance.

Mutations that lead to loss of control of apoptotic checkpoints leads cancer cells to keep growing with essentially the blood supply as a limit on the rate and extent of the

129 growth. Genes that are involved in this include, p53, c-myc, Bcl-2, c-fos among others.

Proteins derived from adenovirus such as ν-Abl and Ras can also restrain apoptotic cell death (Kumar et al., 2005).

Bcl-2 family is a conserved family of proteins is an essential component of mitochondrial membrane and is composed of pro and antiapoptotic members. Interaction between the two subtypes is what controls the decision for or against apoptosis. Some of the members of this family of proteins includes Bcl-2, Bcl-w, Bcl-XL, Bfl-1, Mcl-1, and

Bcl-B all of which are anti-apoptotic and Bax, Bid, Bcl-XS, Bcl-Gs, Bak, Bad, and

PUMA to name a few. Any change due to varied expression of the ratios of pro vs anti- apoptotic members of this family will tip the decision one way or the other. For example if there is increased expression of an anti-apoptotic member of the Bcl-2 family then what is likely to happen is that the cell will be less likely to trigger apoptosis (Raina et al.,

2001b).

An example of this process is by the overexpression of Bcl-2, in a subset of lymphocytes which leads to a B cell lymphoma. This overexpression occurs because of a translocation of the Bcl-2 gene. This translocated Bcl-2 gene comes under the control of a regulatory sequence which leads the amplification of this gene. Hence B cells that would have died due to apoptosis do not due to this translocation (Kasper and Harrison, 2005).

The p53 protein is important as a regulator at G1/S and G2/M levels. Many cancers are associated with the loss or dysregulation of p53 function in ways that impinge upon its functions in cell cycle regulation and apoptosis. There are many stresses that increase p53 levels including UV, IR and gamma rays, hypoxia and genotoxic agents.

Increasing p53 can be accomplished by slowing down its destruction through

130 posttranslational modification. It is p53 that induces a pause in the cell cycle in order to repair the DNA (Hofseth et al., 2004). One of the ways to do this is via increasing levels of p21 which interacts with the cyclin E to stop progression into the S phase. Depending on the circumstances the cell will continue dividing or go into apoptosis depending on if the DNA is repaired. So a cell that is lacking this normal function of p53 will allow for the cell to progress into the S phase and carry with it defective DNA without apoptosis occurring (Lloyd and Hanawalt, 2002). Hence this cell survives and with further replication accumulates more mutations. This genomic instability is a hallmark feature of a cancer cell and allows for the well know features of cancer cells.

From the above, we can see that, in cancer, escape from apoptosis is a necessary feature. Therapeutic regimens that are geared towards regaining sensitivity to apoptosis are beneficial for mitotic lesions and are currently available. They are however limited by their nonselectivity.

5.2 Experimental Hypothesis

Given the above background, it is easy to recognize a similarity between AD and cancer (Table 5.1). The idea behind this is that we can look at the early emergence of the mitotic phenotype in these vulnerable neurons just as we view the same phenomena during ongogenesis. Thus this analogy between AD and cancer is epistemic. This implies that the means by which we know that a cell is cancerous are the same by which we can view this phenotype in AD. Hence in order to solidify this connection between events in

AD and cancer, I sought to look for and identify other markers of transformation from cell cycle markers that are a unique to G1 entry to changes in matrix proteins in order to

131 elaborate upon the dysmitotic geography within these vulnerable neurons as well as confirm this similarity between AD and cancer. Naturally therapeutics directed towards modulating or circumventing this phenotype may be useful in changing the natural history of AD. Taken together I propose that these two disease states have significant overlaps in terms of recruitment of similar mechanisms that deserve investigation and will not only yield novel insights to AD as well as neurodegenerative pathogenesis but also will allow for new therapeutic regimens. Based on this, we formulated the following hypotheses.

1) Neoplastic transformation requires changes in cell-cell and cell-matrix

interactions. A prominent group of these proteins are the ADAM (protein with a

disintegrin and a metalloprotease domain) proteins. The presence of alterations in the

ADAM proteins in AD neurons would indicate not only cell cycle reentry but would

also solidify the comparison to oncogenic lesions.

2) The sequelae of oxidative stress includes damage to nucleoproteins especially

DNA. Oxidative lesions of DNA include 8-Oxoguanine (8-oxoG), whose

accumulation can lead to somatic mutations, is repaired primarily through the base

excision repair pathway (BER) as well as transcription coupled repair pathway

(TCR). Efficient TCR requires BRCA-1. We should expect to find BRCA-1 induction

in the vulnerable neurons of AD. Any changes in BRCA-1 localization within the

nuclear compartment could mean inefficient or absent TCR repair. Hence we will

profile the presence of BRCA-1 in the vulnerable neurons of AD.

3) Finally, if there is indeed G2/M progression then G1 entry must be complete.

The completion of G1 will involve the hyperphosphorylation of the retinoblastoma

132 protein (pRb). Hence we should at a minimum expect to see pRb in these same

vulnerable neurons in AD.

5.3 Methods and materials

a. Tissue Section Preparation

Hippocampal tissue samples were obtained postmortem from patients. All clinical and pathological diagnoses were according to standardized criteria (Khachaturian, 1985),

(Mirra et al., 1991). From the clinical reports available to us, we found no obvious differences in agonal status or other potential confounders between the groups. Tissue was fixed in methacarn (methanol: chloroform: acetic acid in a 6:3:1 v/v/v) or buffered formalin for 16 h at 4°C. Tissue was subsequently dehydrated through graded ethanol and xylene solutions and embedded in paraffin. Microtome sections 6 µm thick were prepared and placed on silane-coated slides (Sigma, St. Louis, MO). The cases used in each study were as follows: ADAM: AD cases n = 6 (age 77-95), control cases: n = 5 (age 53-82); phosphorylated Retinoblastoma where AD cases are n = 5 for each of the sites and control cases are also n=5 for each of the sites. Cases used for BRCA-1 are listed in

Table 5.2

b. Immunocytochemistry

Following immersion in xylene and hydration through graded ethanol solutions, endogenous peroxidase activity was eliminated by incubation of the sections of 3% hydrogen peroxidase for 30 min. To reduce non-specific binding, sections were incubated for 30 min at room temperature in 10% NGS in Tris-buffered saline (TBS; 50 mM Tris–

133 HCl, 150 mM NaCl, pH 7.6). After rinsing briefly with 1% NGS, sections were incubated overnight in the respective antisera (1/100) or a mouse monoclonal antibody, AT8, to altered cytoskeletal τ protein (Smith et al., 1994b). The sections were then incubated in either goat anti-rabbit or goat antimouse antisera (ICN Biomedicals, Costa Mesa, CA), followed by species-specific peroxidase anti-peroxidase complex (Sternberger

Monoclonals Inc., Lutherville, MD and ICN Biomedicals, Costa Mesa, CA). Antibodies were localized using 3, 3 DAB as a chromogen (DAKO Corporation, Carpinteria, CA).

To ensure the specificity of the reactivity observed, immunostaining was also performed in parallel using antibody that had been pre-incubated with the purified immunizing peptide (Bertram et al., 1999).

The primary antisera utilized were against ADAM (rabbit antisera were raised against two peptides SSSWSDSDSQANDEVQEVVSPPSSESS) and a peptide

(CFSSEEQFESESESKD) corresponding to the cytoplasmic tails of ADAM-1 and 2, respectively (Wolfsberg et al., 1995); pRb (pSpT249/252, pS612, pS780, pS807, pS811, pT821). For the BRCA-1 study the following antibodies used were as follows 1) Ab-1

(mouse monoclonal), whose immunogen is the N-terminus 304 amino acid residues of the BRCA-1 protein (commercially available from Oncogene Research Products; Cat No.

OP92). The eluted protein concentration was about 0.1 mg/ml. while the uneluted protein concentration was about three times more. Absorption was done using both fractions. This antibody has been shown to be specific for BRCA-1 and does not react with blood group antigens or growth factor receptors such as EGFR. 2) Phospho BRCA-1 from Upstate Cell Signaling Solutions (Cat no 07-007) whose immunogen is KLH- conjugated, synthetic phospho-peptide corresponding to residues 1489-1500 of human

134 BRCA-1 (C-KEPGVERS [pS] PSK). This rabbit polyclonal IgG recognizes phosphorylated (Ser 1497) BRCA-1 at a molecular weight of 250kDa.

c. Immunoblots

Tissue samples were solubilized in RIPA buffer, protein concentration determined and 20 µgs of protein loaded on SDS PAGE gels. The proteins were transferred to membranes and probed with 1:1000 of our primary antibody and 1:2000 secondary rabbit polyclonal antibody and visualized with ECL according to the manufacturer’s instructions (Santa Cruz, Santa Cruz, CA).

Quantification of the results where needed was performed using a computer- assisted image analysis system (KS300, Zeiss). The data obtained were expressed as optical densities and analysed statistically using Students t-test.

5.4 Results

a. ADAM

Both ADAM-1 (Figure 5.2A) and ADAM-2 (Figure 5.2B) proteins are localized specifically in the cytoplasm pyramidal neurons and NFT in all cases of AD and overlap in the pathological profile to that of ubiquitin (Figure 5.2C). In many of the cells, lysosomal recognition is also seen (Figure 5.3A andC, insets), and in a few AD cases,

Hirano bodies were also immunolabeled. In contrast, ADAM proteins were not detected in young control tissue (Figure 5.3B and D), whereas older control cases found to exhibit pathology as revealed by τ immunostaining (Figure 5.4C) did show limited ADAM-1 localization (Figure 5.4A). Notably, ADAM-2 was found in neurons of aged-matched

135 controls but with a much stronger intensity in those few cells with neurofibrillary pathology (Figure 5.4B). At high magnification, nuclear localization in the cytoplasm of young controls, without any τ involvement (Figure 5.4F) was detected with ADAM-2

(Figure 5.4E) but not ADAM-1 (Figure 5.4D).

The differential staining pattern for ADAM was independent of postmortem interval or agonal status. The close overlap in immunoreactive profiles between antisera to ADAM and other antigenic markers of NFT led us to conduct an amino acid homology search for possible cross-reactive proteins. Using a protein sequence database (Swiss and

PIR and Translated Release 107) to compare the antigenic peptides, there are no homologous proteins, reported in NFT, aside from ADAM-1 and 2, above a threshold of

80% with a K-Tuple of 2. Importantly, these peptides show no homology to cytoskeletal

(including τ and neurofilaments) or other proteins (including cell cycle-related proteins such as CDK4 and p16). Further demonstrating the specificity of our findings, absorption of the primary antisera with peptide significantly reduced immunolabeling

(Figure 5.3A-D). Additionally, adjacent sections only treated with secondary antibodies or with a primary antibody against an irrelevant epitope showed background levels of staining in both AD and control cases (data not shown).

Immunoblot analysis of brain homogenates from control and disease cases revealed intense bands at the reported molecular weights of 60 kDa for ADAM-1 and 97 kDa for ADAM-2 and quantification of immunoblots showed significant differences (P<

0.05) for ADAM-1 and ADAM-2 between Alzheimer and control brain (Figure 5.5).

136 b. BRCA-1

BRCA-1 antisera (Ab-1) immunoreacts with neurofibrillary pathology in AD cases (Figure 5.6) while it does not react with tangles in control cases (Figure 5.8) that have no previous history of dementia or AD-type changes in the cortex upon postmortem.

BRCA-1 antibody (Ab-1) strongly labeled neurofibrillary lesions in the pyramidal neurons of AD brains (Figure 5.6).

We proceeded to compare BRCA-1 induction in AD and control patients using phosphorylated (Ser 1497) form of BRCA-1. We performed SDS-PAGE on brain homogenates from AD and age-matched control cases with the phospho-specific antibody to BRCA-1. Observation of the immunoblot above reveals a seeming increase in the phospho protein levels of BRCA-1 in AD versus controls. There is a significant increase in the amount of phosphorylated BRCA-1 in AD cases as compared to controls (Figure

5.7). However, at this point we are unsure whether this simply represents a change in post-translational modification of BRCA-1 in AD brain or an increase in protein level.

Perhaps the most interesting aspect of our preliminary studies was the fact that the

NFT and senile plaques found in control brains as a product of the normal aging process were not labeled by the BRCA-1 antibody (Figure 5.8 and Table 5.1). Testing of other neurodegenerative diseases revealed an inability of the Ab-1 antibody to immunolocalize any neuropathology in Down syndrome, Pick disease, tauopathy or normal aging. On the other hand, FAD cases caused by PS mutations did show reactivity.

137 c. Phosphoretinoblastoma (pRb)

The phospho-retinoblastoma (pRb) antisera made against different Rb phosphorylational sites, immunoreacts with the neurofibrillary in AD cases (Figure 5.9A and Table 5.2). The age matched control cases show no immunoreactivity. All of the pRb antisera immunoreacts differentially to localize the full spectrum of the AD pathology namely NFT, neuropil threads and also senile plaques (Figure 5.9A). This data is summarized in Table 5.2. The expression of pRb in AD in the cytoplasm of the neuron may be associated with the early history of AD.

5.5 Discussion

a. ADAM

To solidify the parallels to transformative processes, we looked at alterations in matrix proteins in AD. Indeed we did find that in AD there are alterations in ADAM 1 and 2 in Alzheimer disease (Gerst et al., 2000). Matrix proteins are altered in oncogenic poteins are know to become altered in oncogenic proteins (Kumar et al., 2005). Matrix metalloproteinases (MMPs) like ADAM 1 were initially recognized for their extracellular matrix (ECM)-degrading capability during tissue remodeling. Their importance was further highlighted by their role in metastasis. Clinical trials have since evaluated the potential of MMP inhibitors as anticancer therapeutics, but without success. These initial studies point to the complex, multifunctional capacity of MMPs in cancer as shown by their function, not only as strident mediators of advanced malignancies, but also as effectors of early stage tumorigenesis. Research now shows that MMPs, and their tissue inhibitors, affect tumour initiation and growth through loss of cell adhesion, evasion of

138 apoptosis, and deregulation of cell division. The extracellular nature of the metalloproteinase axis situates it as a master regulator of cell fate both of which show alterations during cancer and now AD.

We demonstrate the presence of ADAM (a disintegrin and metalloproteinase) proteins in the cytosolic compartment of vulnerable neurons and AD. Both ADAM-1 and

ADAM-2 proteins are found in neurons in AD tissue, but, in marked contrast, not in similar neuronal populations in control cases; with the exception of cases exhibiting age- related pathology. Immunoblot analyses revealed higher levels of ADAM in homogenates of AD compared to control brain, confirming the immunocytochemical findings that there is a selective induction of ADAM in specific neurons in the disease.

Many features of AD, including attempted cell cycle re-entry from quiescence (Vincent et al., 1996; McShea et al., 1997; Raina et al., 1999a), hyperphosphorylation (Grundke-

Iqbal et al., 1986) re-expression variously of a developmental phenotype, and especially dysregulated loss of cell-matrix integrity are in effect characteristics of a dedifferentiated state (Kumar et al., 2005). Therefore, it is reasonable to speculate roles for ADAM involvement in AD pathogenesis. Both ADAM-1 and -2 are developmentally regulated

(Wolfsberg et al., 1995; Huovila et al., 1996; Podbilewicz, 1996) and therefore, the increased ADAM in neurons in AD is consistent with the notion that neurons receive an activation signal from cell surface receptors that causes the post-mitotic cells to re-enter the cell cycle (McShea et al., 1999a). Additionally there also exists the possibility of

‘inside out’ signaling as indicated by the ADAM cytoplasmic tail (Wolfsberg et al.,

1995). The ADAM family of proteins is normally membrane bound and contains disintegrin as well as metalloprotease domains (Wolfsberg et al., 1995). In addition to

139 mediating intra- and extracellular events, ADAM proteins are also involved in integrin- mediated interactions. Notably, α-integrin is a necessary component of short-term memory (Grotewiel et al., 1998), likely acting by mediating synaptic plasticity via its ligand-binding and signaling properties. Therefore, perturbations in integrin homeostasis, possibly mediated by changes in ADAM, would ultimately affect short-term memory formation or the stability of the process. Indeed, we suspect that the induction and cytoplasmic, rather than membrane, localization of ADAM proteins reported here contributes toward the loss of synaptic connections in both numbers and types, and associated cognitive deficits as observed in AD (Masliah et al., 1989). Finally, of interest regarding the pathogenesis of AD, ADAM proteins have been shown to cleave alpha-2 macroglobulin (Loechel et al., 1998), a substrate that has been linked to variable genetic penetrance of the disease (Blacker et al., 1998).

The precursors of ADAM-1 and 2 mature via proteolytic cleavage of certain domains and disruption of this maturation pathway would lead to abnormal regulation and hence impair normal cell activity. Therefore, select induction of ADAM in vulnerable neurons, with the absence of normal ADAM activity at neuronal membranes, would lead to significant alterations in both extra- and intracellular domains, accelerating cell death.

It is intriguing to note that AD mimics the developmental environment in terms of attempted cycling, hyperphosphorylated τ and re-expression of proteins under developmental control. Taken together, these events can be seen as an attempted regression to a state of dedifferentiation wherein one of the most proximal events is the loss of cell-ECM/cell-cell contacts. Selective immunolocalization of evolutionarily

140 conserved MMPs ADAM-1 and 2 to a vulnerable neuronal population in the hippocampus further implicates cell matrix dysfunction, cell cycle reentry and attempted emergence of these neuronal populations out of quiescent senescence, as among the proximal pathogenic mechanisms in AD.

In conclusion, inappropriate mitogenic stimulation can lead to alterations in

ADAM proteins within specific neuronal population in AD and could provide a proximal link between the well established abnormalities in AD including alterations in the extracellular matrix, the cell membrane, the cytoskeleton, and, ultimately, neuronal loss and memory impairment.

b. BRCA-1

The subcellular location of BRCA-1 is abnormal in AD cases. Specifically,

BRCA-1 is normally a nuclear protein that is rarely seen at significant levels in the cytoplasmic compartment. There are two possible explanations that account for this phenomenon. First, there could be a deficiency in the nuclear shuttling of BRCA-1 and therefore it may be sequestered from its normal function. Alternatively, it is possible that in the disease state BRCA-1 takes on a novel function in the cytoplasm which has yet to be described.

BRCA-1 activation (i.e., phosphorylation) as we see here in the vulnerable neurons in AD is associated with multiple downstream effects in many different areas of cellular signal transduction. Our data indicates a novel association of BRCA-1 antisera

(Ab-1) to an epitope associated with NFT that is exclusive to the neurodegenerative

141 pathology in sporadic AD, PS mutants but not in age matched controls (without a history of dementia), Down’s syndrome, Pick’s disease and tauopathy.

It is intriguing that Ab-1 did not recognize any neurofibrillary pathology in the cases of Down’s syndrome, which serves as a model for AD. Additionally the inability of

Ab1 to recognize tangles in tauopathy and Pick’s disease further serves to indicate an etiopathological mechanism that is operative in these various conditions.

As of this date, there exist no markers that can distinguish NFT of diverse etiologies. This novel finding links AD neurofibrillary pathology to a unique mechanism that is not associated with other neurodegenerative conditions or normal aging, and hence has profound implications not only for screening and therapeutics but may have a predictive role in terms of selective susceptibility to AD.

There is emerging evidence of genomic instability and its sequel as a proximal feature in the pathogenesis of neurodegeneration in AD. Specifically, TUNNEL labeling indicative of fragmented DNA (Cotman and Anderson, 1995), alterations in mRNA resulting in abnormal protein synthesis along with oxidative modifications of DNA which together contribute to form the unstable genomic landscape. In environments of genetic instability such as breast cancer there is altered localization of BRCA-1 (Chen et al.,

1995), from its effectively nuclear localization (Wilson et al., 1999). Here we find that the same BRCA-1 antiserum as used by C. Wilson is altered in localization only in susceptible hippocampal populations in AD cases.

The presence of a pro-oxidant environment in AD neurons along with the DNA modifications such as 8-oxoguanine that are associated with AD neurons brings into

142 focus the need for TCR of DNA that is associated with BRCA-1. Indeed BRCA-1 is associated with the TCR of the oxidative 8-oxoguanine lesions (Le Page et al., 2000).

BRCA-1 is at its lowest levels during G0 and early G1 with a rapid induction of expression by G1/S which is sustained throughout the G2/M phases (Chen et al., 1996). In a similar fashion, BRCA-1 is largely unphosphorylated in quiescent cells and throughout

G1 in cycling cells, it increasing becomes phosphorylated on specific serine residues as the cell cycle progresses (Pao et al., 2000). The detection of pBRCA-1 (Ser 1497) in these vulnerable neurons indeed points to the late G1 phosphorylation of BRCA-1 and confirms that the cell cycle has not only initiated but progressed to the late G1 stage. That

S1497 is phosphorylated by CDK2-cyclin A is consistent with progressive phosphorylation of BRCA-1 in late G1 / S stages of the cell cycle (Ruffner et al., 1999).

The consequences of this phosphorylation are not fully resolved yet. Additionally, the subcellular localization of BRCA-1 is tightly cell cycle regulated. While BRCA-1 is a largely a nuclear protein, during S phase of the cell cycle is translocates into distinct subnuclear bodies, or nuclear dots (Jin et al., 1997).

BRCA-1 activation secondary to oxidative DNA damage can lead to apoptosis through JNK recruitment and caspase activation given an appropriate background.

Previous studies demonstrate that nearly all of the players downstream of BRCA-1 in the apoptotic processes have been described as being in someway dysregulated in AD.

GADD 45 has been shown to be induced by the toxic Aβ25-35 peptide in the human NT-2 preneuronal cell line (Santiard-Baron et al., 2001) and is induced in susceptible neuronal populations in AD brain (Torp et al., 1998). Work done in our laboratory has demonstrated that the JNK kinase pathway is activated in AD susceptible neurons and is

143 associated with NFT pathology (Zhu et al., 2001). Given that in vitro JNK kinase is capable of phosphorylating τ protein in a pattern similar to that observed on PHF, this mechanism provides a link between oxidative DNA damage and formation of NFT.

Furthermore, the downstream effectors the JNK pathway described in BRCA-1-induced apoptosis have been described as associated with AD pathogenesis.

While the mechanisms responsible for the exclusive localization of BRCA-1 to

NFT in AD remain to be determined, one intriguing hypothesis is that BRCA-1 signifies a neurogenic/oncogenic stimulus that is only found in AD. In this regard, there are several examples showing cognitive improvements in dementia patients undergoing chemotherapy (Keimowitz, 1997). Therefore, it would be interesting to investigate the therapeutic efficacy of combination or simply agent antimitotic therapy with vincristine, carmustine, melphalan, cyclophosphamide, or prednisone for AD.

In summation, the novel association of BRCA-1 to susceptible hippocampal pyramidal neurons in AD cases further implicates oxidative stress and induction of repair mechanisms as well as mitotic dysregulation and its sequel in AD.

c. Phosphoretinoblastoma (pRb)

Our results showing increases in retinoblastoma protein in AD are consistent with the idea that the vulnerable neuron in AD has entered the cell division cycle. But what is more ominous is that there seems to be a usurping of the Rb controlled late G1 checkpoint.

pRb functionality is regulated by sequential phosphorylation on many sites as the cell progresses through the cell cycle. Thus pRb is largely hypophosphorylated in the G1

144 phase of the cell cycle and the maximal phosphorylation is achieved during G2.

Increasing phosphorylation of pRb, especially at the G1/S boundary, inactivates its tethering function of cellular proteins that regulate S phase entry. Hypophosphorylated pRb complexes with the E2F family members and prevents transcription by binding proteins that are associated with transcriptional repression. Upon increasing phosphorylation, Rb no longer sequesters of the E2F members. The E2F proteins, when associated with their DP (differentiation-regulated transcription factor) partner proteins, then activate the expression of a number of genes that promote S phase entry. These include DNA polymerase α, thymidylate synthase, dihydrofolate reductase, ribonucleotide reductase, and cyclin E. Conditional mouse knockouts showed this function of the E2F family members.

Indeed inactivation of Rb can result from a mutation of the Rb protein, but this is infrequent, and occurs primarily in retinoblastomas, osteosarcomas, and a minority of breast cancers (Pines and Hunter, 1994), (Sherr, 1996), (Weinberg, 1996). However, much more important is the functional inactivation of Rb by hyperphosphorylation. This is normally this is due to elevated cdk activity that is caused by overexpression of cyclins and also cyclin dependent kinases.

Indeed, many tumors show loss of Rb or increased expression of cyclin D1

(Bartek et al., 1993), (Pokrovskaja et al., 1996), (Bartkova et al., 1996). Thus either loss of the suppressor activity of Rb, or overexpression of cyclin D1 can over-ride this checkpoint (Sherr, 1996), (Weinberg, 1996). A fair amount of evidence has implicated D- type cyclins in neoplastic development. Additionally, the involvement of cdk4 in the neoplastic process was suggested by the fact that cdk4 amplification and/or

145 overexpression were detected in human glioblastomas, although in these tumors overexpression and/or amplification of D-type cyclins were not seen (Sonoda et al.,

1995), (Ichimura et al., 1996). In addition, cdk4 mutations are identified in patients with familial melanoma (Wolfel et al., 1995), and, recently, amplification and overexpression of cdk4 was also detected in sporadic breast carcinomas (An et al., 1999), ovarian carcinomas (Masciullo et al., 1997) and sarcomas (Kanoe et al., 1998). Taken together, all of these proteins that govern cell cycle control are exciting targets for therapy in AD.

It goes a long way to make sure that cell cycle dysfunction plays an important early role in AD.

5.6 Overall Conclusions

Nature uses what is at hand in order to respond to a diversity of stimuli. Indeed many different stressors recruit the same if not similar pathways. Survival is an attribute that is selected for in any given niche. The totality of the data seems to indicate that neurons in AD do try to survive. This makes sense from evolutionary perspectives in that there are a limited number of neurons and their survival is paramount. What is interesting is that this attempt at survival is akin (in terms of the machinery utilized) to a neoplastic lesion. Indeed the main feature that is common to cancr and AD is age. Increasing age is a risk factor for AD as well as cancer. Those people who are 65 and over have nearly a 12 fold increase in the incidence of a neoplastic lesion. This includes a vast spectrum of organ systems namely prostate, colon, pancreas, urinary bladder, stomach and lung cancer. AD prevalence and incidence goes up after 65 years also. The data indicates that

146 in AD as in neoplasia there is recruitment of the same mechanisms namely; oxidative stress, apoptotic avoidance, cell cycle dysregulation and DNA repair dysfunction.

In AD there is selective recruitment of NQO1 in neurons that survive. This is paralleled in neoplasia where NQO1 is also recruited in the neoplastic cells. We saw BH increased in the vulnerable neurons in AD and again BH is a feature of many types of cancers.

In order for a neoplasia to be successful there need to be mechanisms that avoid apoptosis. Indeed loss of apoptotic control is a hallmark of neoplasia. This cardinal feature is also paralleled in AD were neurons show apoptotic avoidance. Added to this picture is the dysregulation of cell phase control proteins in AD. Many of the cell cycle control proteins serve as protooncogenes. Their dysregulation is a hallmark of cancer.

The same phenomenon we see in AD.

This desynchrony in AD neurons in terms of the presence of markers of different phases of the cell cycle supports the idea that this is a dysmitotic process. It is difficult to know just which lesion/cell cycle marker is important in promoting the AD phenotype at any given time. The phenotypic heterogeneity in AD neurons with respect to cell cycle markers is also reflected in neoplasia and serves to indicate dysregulation of the cell cycle and loss of appropriate control. Inappropriate control of many of these cell cycle control proteins by themselves can lead to oncogenesis. There are a number of known CDK4 and cyclin D tumors.

Added to the above is the presence of pRb protein in AD neurons which is indicative of entry into S phase. Furthermore the presence of BRCA-1 in the cytoplasm raises even more questions as the state of DNA repair in AD especially that of

147 transcription coupled repair. Both BRCA-1 and Rb are known tumor suppressor proteins, each one by itself capable of inducing a mitotic lesion.

All of these parallels between AD and neoplasia are not coincidental but indicative of underlying similar mechanisms. Needless to say in AD there is not a clonal proliferation of neurons. AD neurons may resolve oncogenic stimulation in ways that does not lead to tumor. Better understanding of these mechanisms may lead to novel ways to deal with neoplasia. Finally, cytostatic chemotherapy may prove helpful in AD by stopping cell cycle re-entry or inducing stasis in one phase of the cell cycle. Cytostatic agents are used restrain tumor progression (rather than induce cytotoxic cytoreduction).

Aging theories include increased susceptability over time to insults, DNA repair dysfunction etc over time. All of these are also features that generate a neoplastic lesion.

Indeed all of these are also features of Alzheimers disease. AD could indeed represent a cancer in the postmitotic environment but more interestingly it could show how the postmitotic state behaves given oncogenic stimulation. Elucidation of mechanisms in handling oncogenic stimulation will have broad applications beyond AD itself.

5.7 Future Directions

a. Apoptotic Avoidance

If the apoptotic pathway is blocked in AD then there should be induction of proteins in these neurons that block apoptosis. Indeed IAP proteins are well known to block the progression of apoptosis at multiple levels. Indeed members of the IAP family have been seen in Down Syndrome (Seidl, Bajo et al. 1999). We predict that members of the IAP family will be recruited in these vulnerable neurons.

148

b. Functional Consequences of BRCA-1 Misexpression in Post-Mitotic Neurons

Since in AD the neurons that are affected are postmitotic, it would be of interest to investigate the consequences of BRCA-1 in postmitotic neurons (Figure 5.10). The studies on the role of BRCA-1 overexpression have used transformed dividing cell populations. Therefore, very little is known about the effect of BRCA-1 overexpression on post-mitotic neurons like those susceptible to AD pathogenesis. We speculate that: a) in dividing cell populations, the balance between apoptotic cell death and cell cycle arrest in response to BRCA-1 overexpression will be influenced by the balance of activation of the JNK and ERK signaling pathways, as described in the literature; b) in differentiated neuroblastoma cells and primary neurons overexpression of BRCA-1 will activate the

JNK signaling pathway and lead to apoptotic cell death. When JNK signaling is blocked the ERK pathway will be activated and may lead to abnormal cell cycle re-entry (Figure

5.11); and c) activation of the JNK and ERK signaling pathways will lead the hyperphosphorylation of τ protein on AD-relevant residues.

149 Table 5.1. BRCA-1 (Ab-1) immunoreactive AD cases and respective predispositional factors.

AD Genotype Ab1 Z1/AT8 % of AT8 Age PMI Relevant accessory path

A95-91M e3e3 181 256 70.7 93 22 AD, amyloid angio. A93-378 Na 116 422 27.4 84 11 AD A91-172 e3e4 169 307 55.0 83 6 AD, BW, microinfarct A90-261 Na 184 1162 15.8 83 9 AD, BW A98-46 Na 13 236 5.5 92 2.5 AD A98-15 Na 463 1263 36.6 AD A98-62 Na 1097 2213 49.5 77 3 AD A85-4 Na 73 393 18.5 AD A82-103 Na 270 785 34.3 76 17 AD, front meningioma A90-11 e3e3 128 329 38.9 85 6.5 AD, microinfarct A93-285 e3e4 0 0 0 80 5.5 AD, sup sag sinus thr A98-62 Na 1163 2160 53.8 77 3 AD, A91-228 e3e4 207 1121 18.4 79 4 AD, DLBD,CJD A95-145 e3e4 na na - 77 2 AD, amyl. Angiopathy A91-450 e3e4 326 na - 78 4 AD, A98-24 Na 627 967 64.8 77 7 AD Na 476 1458 32.6 A91-450 e3e4 971 1321 73.5 77 7 AD A90-11E e3e3 272 658 41.3 85 6.5 AD A83-442F Na 230 268 85.8 76 6 AD O88-24 Na 29 192 15.1 85 AD A84-11 Na 657 832 78.9 89 24 AD A87-18 Na 107 621 17.2 AD A88-22 Na 86 492 17.4 88 3.5 AD, Park

150 Table 5.2 Antisera to Retinoblastoma phosphorylational sites that immunoreacts with

AD neuropathology.

Rb phosphorylational sites Tangles Plaques Neuropil Threads

pSpT249/252 + + - pS612 ++ + + pS807 +++ ++ ++ pS811 +++ ++ ++ pT821 +++ +++ ++

151

Figure 5.1. The three main phenomena in AD are also seen in oncogenic lesions.

152

Figure 5.2. Serial sections of hippocampus from an Alzheimer disease case immunolabeled with antisera to ADAM-1 (A), ADAM-2 (B) and ubiquitin (C).

Essentially all of the NFT marked by ubiquitin (C) also contain ADAM proteins (A,B).

*Represents landmark blood vessel. Scale bar = 5-50 µm.

153

Figure 5.3. Serial sections immunolabeled with antisera to ADAM-1 (A) and ADAM-2

(C) absorbed with antigen (B) and (D), respectively. Arrowheads indicate Hirano bodies.

Insets in (A) and (C) show high magnification of neuronal ADAM-1 and 2 immunoreactivity, respectively. *Represent landmark blood vessel. Scale bar 5-50 µm; inset 10 µm.

154

Figure 5.4. Serial sections of hippocampus from a control showing age-related pathology as marked by τ (C) (arrowheads indicate NFT), immunostained with ADAM-1 (A) and

ADAM-2 (B), compared to a control exhibiting no pathology (F) immunostained with

ADAM-1 (D) and ADAM-2 (E). *Represents landmark vessel. Scale bar = 5-50 µm.

155

Figure 5.5. Immunoblot quantification of ADAM-1 and 2 in brain homogenates shows significant increases in AD vs. control brain, P < 0.05.

156

Figure 5.6. Paraffin embedded sections from AD stained with anti-BRCA-1 antibody. In

AD, NFT and senile plaques are strongly stained.

157

AD1 AD2 AD3 C 1 C 2 0.7

0.6

0.5

0.4

0.3

0.2

Ratio pBRCA1/actin of 0.1

0 AD Control

Figure 5.7. Western blot localizes pBRCA-1 (Ser 1497) immunoreactivity to 250 kDa.

158

A B

Figure 5.8 Adjacent sections from an aged control brain stained with anti-BRCA1 antibody (left) and AT8 (right), a marker for the phosphorylated tau in found in the NFT.

Note that BRCA1 does not label “control” NFT. This represents the first demonstration of a marker that is specific for NFTs lesions in AD brain.

159

A B

Figure 5.9. pRb (T821) immunolocalizes the whole spectrum of the AD neuropathology from intracellular tangles, to senile plaques and finally to neuropil threads (A). The age- matched control cases show no immunoreactivity (B).

160

Empty Vector or (Will be titrated between BRCA1-expressing 100-1000 plaque forming units) Adenovirus

Cultured cells

48, 72 and 96 hours

Assess Proliferative Assess Levels of Assess Activation of Assess Levels of Capacity Apoptosis Signaling Pathways Tau Phosphorylation Perform IHC and BrdU incorporation Perform TUNELstaining Western blot analysis Perform IHC and of cells with phospho-JNK and Western FACS sort cells to phospo-ERK1/2 antibodies blot analysis of samples determine the phase Measure DNA using phospj-specifc tau of any cell cycle arrest fragmentation Assess the upstream antibodies to AD relevant activators,GADD45 and Assess activation of caspase MEK4, and downstream as by IHC and Western blot c-jun, by IHC and Western analysis using cleavage analysis specific antibodies

Figure 5.10. Effects of BRCA1 overexpression are measured by various assays.

161

Empty Vector or (Will be titrated between BRCA1-expressing 100-1000 plaque forming units) Adenovirus

Cultured cells Add specific inhibitors for the JNK, SP600125. or the ERK1/2, PD98059, 48, 72 and 96 hours signaling pathaways

Assess Proliferative Assess Levels of Assess Activation of Assess Levels of Capacity Apoptosis Signaling Pathways Tau Phosphorylation Perform IHC and BrdU incorporation Perform TUNELstaining Western blot analysis Perform IHC and Western of cells with phospho-JNK and blot analysis of samples FACS sort cells to phospo-ERK1/2 antibodies using phospj-specifc tau determine the phase Measure DNA fragmentation antibodies to AD relevant of any cell cycle arrest Assess the upstream residues Assess activation of caspase activators,GADD45 and by IHC and Western blot MEK4, and downstream as analysis using cleavage c-jun, by IHC and Western specific antibodies analysis

Figure 5.11. Blocking JNK/ERK signaling will help delineate the role of JNK/ERK signal transduction.

162 5.8 Relevant Publications

1. Gerst, J.L., Raina, A.K., Pirim, I., McShea, A., Harris, P.L.R., Siedlak, S.L.,

Takeda, A., Petersen, R.B. and Smith, M.A. (2000) Altered cell-matrix associated

ADAM in Alzheimer disease. J. Neurosci. Res., 59, 680-684.

2. Zhu, X., Raina, A.K., Boux, H., Simmons, Z.L., Takeda, A. and Smith, M.A.

(2000) Activation of oncogenic pathways in degenerating neurons in Alzheimer

disease. Int. J. Devl. Neurosci., 18, 433-437.

3. Raina, A.K., Zhu, X., Rottkamp, C.A., Monteiro, M., Takeda, A. and Smith,

M.A. (2000) Cyclin’ toward dementia: cell cycle abnormalities and abortive

oncogenesis in Alzheimer disease. J. Neurosci. Res., 61, 128-133.

4. Zhu, X., Raina, A.K., Perry, G. and Smith, M.A. (2004) Alzheimer’s disease: the

two-hit hypothesis. Lancet Neurol., 3, 219-226.

163 Bibliography

Aksenov MY, Tucker HM, Nair P, Aksenova MV, Butterfield DA, Estus S, Markesbery

WR (1998) The expression of key oxidative stress-handling genes in different

brain regions in Alzheimer's disease. J Mol Neurosci 11:151-164.

Alberts B (2004) Essential cell biology, 2nd Edition. New York: Garland Science.

Allen JW, Eldadah BA, Huang X, Knoblach SM, Faden AI (2001) Multiple caspases are

involved in beta-amyloid-induced neuronal apoptosis. J Neurosci Res 65:45-53.

Alonso AC, Grundke-Iqbal I, Iqbal K (1996) Alzheimer's disease hyperphosphorylated

tau sequesters normal tau into tangles of filaments and disassembles microtubules.

Nat Med 2:783-787.

Alt JR, Gladden AB, Diehl JA (2002) p21(Cip1) Promotes cyclin D1 nuclear

accumulation via direct inhibition of nuclear export. J Biol Chem 277:8517-8523.

Alzheimer A, Stelzmann RA, Schnitzlein HN, Murtagh FR (1995) An English translation

of Alzheimer's 1907 paper, "Uber eine eigenartige Erkankung der Hirnrinde".

Clin Anat 8:429-431.

American Psychiatric Association., American Psychiatric Association. Task Force on

DSM-IV. (2000) Diagnostic and statistical manual of mental disorders : DSM-IV-

TR, 4th , text revision. Edition. Washington, DC: American Psychiatric

Association.

Ames BN GL (1991) Endogenous mutagens and the causes of aging and cancer. Mutat

Res 250:3-16.

164 An HX, Beckmann MW, Reifenberger G, Bender HG, Niederacher D (1999) Gene

amplification and overexpression of CDK4 in sporadic breast carcinomas is

associated with high tumor cell proliferation. Am J Pathol 154:113-118.

Anderson AJ, Su JH, Cotman CW (1996) DNA damage and apoptosis in Alzheimer's

disease: colocalization with c-Jun immunoreactivity, relationship to brain area,

and effect of postmortem delay. J Neurosci 16:1710-1719.

Arendt T, Holzer M, Fruth R, Bruckner MK, Gartner U (1995) Paired helical filament-

like phosphorylation of tau, deposition of beta/A4-amyloid and memory

impairment in rat induced by chronic inhibition of phosphatase 1 and 2A.

Neuroscience 69:691-698.

Athar M (2002) Oxidative stress and experimental carcinogenesis. Indian J Exp Biol

40:656-667.

Balazs L, Leon M (1994) Evidence of an oxidative challenge in the Alzheimer's brain.

Neurochem Res 19:1131-1137.

Bartek J, Staskova Z, Draetta G, Lukas J (1993) Molecular pathology of the cell cycle in

human cancer cells. Stem Cells 11 Suppl 1:51-58.

Bartkova J, Zemanova M, Bartek J (1996) Abundance and subcellular localisation of

cyclin D3 in human tumours. Int J Cancer 65:323-327.

Baumann K, Mandelkow EM, Biernat J, Piwnica-Worms H, Mandelkow E (1993)

Abnormal Alzheimer-like phosphorylation of tau-protein by cyclin-dependent

kinases cdk2 and cdk5. FEBS Lett 336:417-424.

Beal MF (1995) Aging, energy, and oxidative stress in neurodegenerative diseases. Ann

Neurol 38:357-366.

165 Behl C, Davis J, Cole GM, Schubert D (1992) Vitamin E protects nerve cells from

amyloid beta protein toxicity. Biochem Biophys Res Commun 186:944-950.

Benitez-Bribiesca L, Sanchez-Suarez P (1999) Oxidative damage, bleomycin, and

gamma radiation induce different types of DNA strand breaks in normal

lymphocytes and thymocytes. A comet assay study. Ann N Y Acad Sci 887:133-

149.

Bertram MJ, Berube NG, Swanson XH, Pereira-Smith OM (1999) Assembly of a BAC

contig of the complementation group B cell senescence gene candidate region at

4q33-q34.1 and identification of expressed sequences. Genomics 56:353-354.

Beyer RE (1994) The relative essentiality of the antioxidative function of coenzyme Q--

the interactive role of DT-diaphorase. Mol Aspects Med 15 Suppl:s117-129.

Blacker D, Wilcox MA, Laird NM, Rodes L, Horvath SM, Go RC, Perry R, Watson B,

Jr., Bassett SS, McInnis MG, Albert MS, Hyman BT, Tanzi RE (1998) Alpha-2

macroglobulin is genetically associated with Alzheimer disease. Nat Genet

19:357-360.

Bowen RL, Isley JP, Atkinson RL (2000) An association of elevated serum gonadotropin

concentrations and Alzheimer disease? J Neuroendocrinol 12:351-354.

Bowen RL, Smith MA, Harris PL, Kubat Z, Martins RN, Castellani RJ, Perry G, Atwood

CS (2002) Elevated luteinizing hormone expression colocalizes with neurons

vulnerable to Alzheimer's disease pathology. J Neurosci Res 70:514-518.

Bowling AC, Beal MF (1995) Bioenergetic and oxidative stress in neurodegenerative

diseases. Life Sci 56:1151-1171.

166 Braak H, Braak E (1991) Neuropathological stageing of Alzheimer-related changes. Acta

Neuropathol (Berl) 82:239-259.

Breitner JC, Folstein MF (1984) Familial Alzheimer Dementia: a prevalent disorder with

specific clinical features. Psychol Med 14:63-80.

Brion JP, Couck AM, Passareiro E, Flament-Durand J (1985) Neurofibrillary tangles of

Alzheimer's disease: an immunohistochemical study. J Submicrosc Cytol 17:89-

96.

Bruni P, Minopoli G, Brancaccio T, Napolitano M, Faraonio R, Zambrano N, Hansen U,

Russo T (2002) Fe65, a ligand of the Alzheimer's beta-amyloid precursor protein,

blocks cell cycle progression by down-regulating thymidylate synthase

expression. J Biol Chem 277:35481-35488.

Cadenas E (1995) Antioxidant and prooxidant functions of DT-diaphorase in quinone

metabolism. Biochem Pharmacol 49:127-140.

Castellani RJ, Perry G, Harris PL, Cohen ML, Sayre LM, Salomon RG, Smith MA

(1998) Advanced lipid peroxidation end-products in Alexander's disease. Brain

Res 787:15-18.

Castellani RJ, Harris PL, Sayre LM, Fujii J, Taniguchi N, Vitek MP, Founds H, Atwood

CS, Perry G, Smith MA (2001) Active glycation in neurofibrillary pathology of

Alzheimer disease: N(epsilon)-(carboxymethyl) lysine and hexitol-lysine. Free

Radic Biol Med 31:175-180.

Chalmers K, Wilcock GK, Love S (2003) APOE epsilon 4 influences the pathological

phenotype of Alzheimer's disease by favouring cerebrovascular over parenchymal

accumulation of A beta protein. Neuropathol Appl Neurobiol 29:231-238.

167 Chen Y, Farmer AA, Chen CF, Jones DC, Chen PL, Lee WH (1996) BRCA1 is a 220-

kDa nuclear phosphoprotein that is expressed and phosphorylated in a cell cycle-

dependent manner. Cancer Res 56:3168-3172.

Chen Y, Chen CF, Riley DJ, Allred DC, Chen PL, Von Hoff D, Osborne CK, Lee WH

(1995) Aberrant subcellular localization of BRCA1 in breast cancer. Science

270:789-791.

Cohen GM (1997) Caspases: the executioners of apoptosis. Biochem J 326 ( Pt 1):1-16.

Colton CA, Gilbert DL (1987) Production of superoxide anions by a CNS macrophage,

the microglia. FEBS Lett 223:284-288.

Cotman CW, Anderson AJ (1995) A potential role for apoptosis in neurodegeneration

and Alzheimer's disease. Mol Neurobiol 10:19-45.

Cotman CW, Su JH (1996) Mechanisms of neuronal death in Alzheimer's disease. Brain

Pathol 6:493-506.

Cras P, Smith MA, Richey PL, Siedlak SL, Mulvihill P, Perry G (1995) Extracellular

neurofibrillary tangles reflect neuronal loss and provide further evidence of

extensive protein cross-linking in Alzheimer disease. Acta Neuropathol (Berl)

89:291-295.

Cross CE, Halliwell B, Borish ET, Pryor WA, Ames BN, Saul RL, McCord JM, Harman

D (1987) Oxygen radicals and human disease. Ann Intern Med 107:526-545.

Cross JV, Deak JC, Rich EA, Qian Y, Lewis M, Parrott LA, Mochida K, Gustafson D,

Vande Pol S, Templeton DJ (1999) Quinone reductase inhibitors block

SAPK/JNK and NFkappaB pathways and potentiate apoptosis. J Biol Chem

274:31150-31154.

168 Crutcher KA, Scott SA, Liang S, Everson WV, Weingartner J (1993) Detection of NGF-

like activity in human brain tissue: increased levels in Alzheimer's disease. J

Neurosci 13:2540-2550.

Desjardins P, Ledoux S (1998) Expression of ced-3 and ced-9 homologs in Alzheimer's

disease cerebral cortex. Neurosci Lett 244:69-72.

Eisen A, Utley RT, Nourani A, Allard S, Schmidt P, Lane WS, Lucchesi JC, Cote J

(2001) The yeast NuA4 and Drosophila MSL complexes contain homologous

subunits important for transcription regulation. J Biol Chem 276:3484-3491.

Elledge SJ (1996) Cell cycle checkpoints: preventing an identity crisis. Science

274:1664-1672.

Engidawork E, Gulesserian T, Seidl R, Cairns N, Lubec G (2001) Expression of

apoptosis related proteins in brains of patients with Alzheimer's disease. Neurosci

Lett 303:79-82.

Eymin B, Haugg M, Droin N, Sordet O, Dimanche-Boitrel MT, Solary E (1999) p27Kip1

induces drug resistance by preventing apoptosis upstream of cytochrome c release

and procaspase-3 activation in leukemic cells. Oncogene 18:1411-1418.

Ferrando AA, Velasco G, Campo E, Lopez-Otin C (1996) Cloning and expression

analysis of human bleomycin hydrolase, a cysteine proteinase involved in

chemotherapy resistance. Cancer Res 56:1746-1750.

Fratiglioni L, Grut M, Forsell Y, Viitanen M, Grafstrom M, Holmen K, Ericsson K,

Backman L, Ahlbom A, Winblad B (1991) Prevalence of Alzheimer's disease and

other dementias in an elderly urban population: relationship with age, sex, and

education. Neurology 41:1886-1892.

169 Gamblin TC, King ME, Kuret J, Berry RW, Binder LI (2000) Oxidative regulation of

fatty acid-induced tau polymerization. Biochemistry 39:14203-14210.

Gerst JL, Raina AK, Pirim I, McShea A, Harris PL, Siedlak SL, Takeda A, Petersen RB,

Smith MA (2000) Altered cell-matrix associated ADAM proteins in Alzheimer

disease. J Neurosci Res 59:680-684.

Gibson L, Holmgreen SP, Huang DC, Bernard O, Copeland NG, Jenkins NA, Sutherland

GR, Baker E, Adams JM, Cory S (1996) bcl-w, a novel member of the bcl-2

family, promotes cell survival. Oncogene 13:665-675.

Gilmore PM, Quinn JE, Mullan PB, Andrews HN, McCabe N, Carty M, Kennedy RD,

Harkin DP (2003) Role played by BRCA1 in regulating the cellular response to

stress. Biochem Soc Trans 31:257-262.

Goedert M, Jakes R, Crowther RA, Six J, Lubke U, Vandermeeren M, Cras P,

Trojanowski JQ, Lee VM (1993) The abnormal phosphorylation of tau protein at

Ser-202 in Alzheimer disease recapitulates phosphorylation during development.

Proc Natl Acad Sci U S A 90:5066-5070.

Gomes WA, Kessler JA (2001) Msx-2 and p21 mediate the pro-apoptotic but not the anti-

proliferative effects of BMP4 on cultured sympathetic neuroblasts. Dev Biol

237:212-221.

Good PF, Perl DP, Bierer LM, Schmeidler J (1992) Selective accumulation of aluminum

and iron in the neurofibrillary tangles of Alzheimer's disease: a laser microprobe

(LAMMA) study. Ann Neurol 31:286-292.

Good PF, Werner P, Hsu A, Olanow CW, Perl DP (1996) Evidence of neuronal oxidative

damage in Alzheimer's disease. Am J Pathol 149:21-28.

170 Graeber MB (1999) No man alone: the rediscovery of Alois Alzheimer's original cases.

Brain Pathol 9:237-240.

Grotewiel MS, Beck CD, Wu KH, Zhu XR, Davis RL (1998) Integrin-mediated short-

term memory in Drosophila. Nature 391:455-460.

Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M, Wisniewski HM, Binder LI (1986)

Abnormal phosphorylation of the microtubule-associated protein tau (tau) in

Alzheimer cytoskeletal pathology. Proc Natl Acad Sci U S A 83:4913-4917.

Gupta S, Rifici V, Crowley S, Brownlee M, Shan Z, Schlondorff D (1992) Interactions of

LDL and modified LDL with mesangial cells and matrix. Kidney Int 41:1161-

1169.

Hampton MB, Fadeel B, Orrenius S (1998) Redox regulation of the caspases during

apoptosis. Ann N Y Acad Sci 854:328-335.

Hensley K, Carney JM, Mattson MP, Aksenova M, Harris M, Wu JF, Floyd RA,

Butterfield DA (1994) A model for beta-amyloid aggregation and neurotoxicity

based on free radical generation by the peptide: relevance to Alzheimer disease.

Proc Natl Acad Sci U S A 91:3270-3274.

Heslop KE, Goss-Sampson MA, Muller DP, Curzon G (1996) Serotonin metabolism and

release in frontal cortex of rats on a vitamin E-deficient diet. J Neurochem

66:860-864.

Hileman EO, Liu J, Albitar M, Keating MJ, Huang P (2004) Intrinsic oxidative stress in

cancer cells: a biochemical basis for therapeutic selectivity. Cancer Chemother

Pharmacol 53:209-219.

171 Hofseth LJ, Hussain SP, Harris CC (2004) p53: 25 years after its discovery. Trends

Pharmacol Sci 25:177-181.

Huovila AP, Almeida EA, White JM (1996) ADAMs and cell fusion. Curr Opin Cell Biol

8:692-699.

Ichimura K, Schmidt EE, Goike HM, Collins VP (1996) Human glioblastomas with no

alterations of the CDKN2A (p16INK4A, MTS1) and CDK4 genes have frequent

mutations of the retinoblastoma gene. Oncogene 13:1065-1072.

Iqbal K, Zaidi T, Thompson CH, Merz PA, Wisniewski HM (1984) Alzheimer paired

helical filaments: bulk isolation, solubility, and protein composition. Acta

Neuropathol (Berl) 62:167-177.

Iqbal K, Grundke-Iqbal I, Zaidi T, Merz PA, Wen GY, Shaikh SS, Wisniewski HM,

Alafuzoff I, Winblad B (1986) Defective brain microtubule assembly in

Alzheimer's disease. Lancet 2:421-426.

Israsena N, Kessler JA (2002) Msx2 and p21(CIP1/WAF1) mediate the proapoptotic

effects of bone morphogenetic protein-4 on ventricular zone progenitor cells. J

Neurosci Res 69:803-809.

James TW, Bohman R (1981) Proliferation of mitochondria during the cell cycle of the

human cell line (HL-60). J Cell Biol 89:256-260.

Janicki SM, Monteiro MJ (1999) Presenilin overexpression arrests cells in the G1 phase

of the cell cycle. Arrest potentiated by the Alzheimer's disease

PS2(N141I)mutant. Am J Pathol 155:135-144.

Janicki SM, Stabler SM, Monteiro MJ (2000) Familial Alzheimer's disease presenilin-1

mutants potentiate cell cycle arrest. Neurobiol Aging 21:829-836.

172 Jin Y, Xu XL, Yang MC, Wei F, Ayi TC, Bowcock AM, Baer R (1997) Cell cycle-

dependent colocalization of BARD1 and BRCA1 proteins in discrete nuclear

domains. Proc Natl Acad Sci U S A 94:12075-12080.

Johnson GV, Bailey CD (2002) Tau, where are we now? J Alzheimers Dis 4:375-398.

Kanemaru K, Takio K, Miura R, Titani K, Ihara Y (1992) Fetal-type phosphorylation of

the tau in paired helical filaments. J Neurochem 58:1667-1675.

Kanoe H, Nakayama T, Murakami H, Hosaka T, Yamamoto H, Nakashima Y,

Tsuboyama T, Nakamura T, Sasaki MS, Toguchida J (1998) Amplification of the

CDK4 gene in sarcomas: tumor specificity and relationship with the RB gene

mutation. Anticancer Res 18:2317-2321.

Kasper DL, Harrison TR (2005) Harrison's principles of internal medicine, 16th / Edition.

New York: McGraw-Hill Medical Pub. Division.

Kawada M, Yamagoe S, Murakami Y, Suzuki K, Mizuno S, Uehara Y (1997) Induction

of p27Kip1 degradation and anchorage independence by Ras through the MAP

kinase signaling pathway. Oncogene 15:629-637.

Keimowitz RM (1997) Dementia improvement with cytotoxic chemotherapy. A case of

Alzheimer disease and multiple myeloma. Arch Neurol 54:485-488.

Khachaturian ZS (1985) Diagnosis of Alzheimer's disease. Arch Neurol 42:1097-1105.

Kirkland RA, Windelborn JA, Kasprzak JM, Franklin JL (2002) A Bax-induced pro-

oxidant state is critical for cytochrome c release during programmed neuronal

death. J Neurosci 22:6480-6490.

173 Kirschner DA, Abraham C, Selkoe DJ (1986) X-ray diffraction from intraneuronal paired

helical filaments and extraneuronal amyloid fibers in Alzheimer disease indicates

cross-beta conformation. Proc Natl Acad Sci U S A 83:503-507.

Koldamova RP, Lefterov IM, DiSabella MT, Lazo JS (1998) An evolutionarily conserved

cysteine protease, human bleomycin hydrolase, binds to the human homologue of

ubiquitin-conjugating enzyme 9. Mol Pharmacol 54:954-961.

Kumar V, Abbas AK, Fausto N, Robbins SL, Cotran RS (2005) Robbins and Cotran

pathologic basis of disease, 7th / Edition. Philadelphia: Elsevier/Saunders.

Lazo JS, Pitt BR (1995) Metallothioneins and cell death by anticancer drugs. Annu Rev

Pharmacol Toxicol 35:635-653.

Le Page F, Randrianarison V, Marot D, Cabannes J, Perricaudet M, Feunteun J, Sarasin

A (2000) BRCA1 and BRCA2 are necessary for the transcription-coupled repair

of the oxidative 8-oxoguanine lesion in human cells. Cancer Res 60:5548-5552.

LeBlanc A, Liu H, Goodyer C, Bergeron C, Hammond J (1999) Caspase-6 role in

apoptosis of human neurons, amyloidogenesis, and Alzheimer's disease. J Biol

Chem 274:23426-23436.

Ledesma MD, Bonay P, Colaco C, Avila J (1994) Analysis of microtubule-associated

protein tau glycation in paired helical filaments. J Biol Chem 269:21614-21619.

Lenaz G (1998) Role of mitochondria in oxidative stress and ageing. Biochim Biophys

Acta 1366:53-67.

Lenferink AE, Busse D, Flanagan WM, Yakes FM, Arteaga CL (2001) ErbB2/neu kinase

modulates cellular p27(Kip1) and cyclin D1 through multiple signaling pathways.

Cancer Res 61:6583-6591.

174 Levine AJ (1997) p53, the cellular gatekeeper for growth and division. Cell 88:323-331.

Lim GP, Yang F, Chu T, Gahtan E, Ubeda O, Beech W, Overmier JB, Hsiao-Ashec K,

Frautschy SA, Cole GM (2001) Ibuprofen effects on Alzheimer pathology and

open field activity in APPsw transgenic mice. Neurobiol Aging 22:983-991.

Lindwall G, Cole RD (1984) Phosphorylation affects the ability of tau protein to promote

microtubule assembly. J Biol Chem 259:5301-5305.

Lloyd DR, Hanawalt PC (2002) p53 controls global nucleotide excision repair of low

levels of structurally diverse benzo(g)chrysene-DNA adducts in human

fibroblasts. Cancer Res 62:5288-5294.

Lodish HF (2004) Molecular cell biology, 5th Edition. New York: W.H. Freeman and

Company.

Loechel F, Gilpin BJ, Engvall E, Albrechtsen R, Wewer UM (1998) Human ADAM 12

(meltrin alpha) is an active metalloprotease. J Biol Chem 273:16993-16997.

Lopez Salon M, Morelli L, Castano EM, Soto EF, Pasquini JM (2000) Defective

ubiquitination of cerebral proteins in Alzheimer's disease. J Neurosci Res 62:302-

310.

Lovell MA, Ehmann WD, Butler SM, Markesbery WR (1995) Elevated thiobarbituric

acid-reactive substances and antioxidant enzyme activity in the brain in

Alzheimer's disease. Neurology 45:1594-1601.

Lu KP, Liou YC, Vincent I (2003) Proline-directed phosphorylation and isomerization in

mitotic regulation and in Alzheimer's Disease. Bioessays 25:174-181.

Lucassen PJ, Chung WC, Vermeulen JP, Van Lookeren Campagne M, Van Dierendonck

JH, Swaab DF (1995) Microwave-enhanced in situ end-labeling of fragmented

175 DNA: parametric studies in relation to postmortem delay and fixation of rat and

human brain. J Histochem Cytochem 43:1163-1171.

Lukiw WJ, Rogaev EI, Wong L, Vaula G, McLachlan DR, St George Hyslop P (1994)

Protein-DNA interactions in the promoter region of the amyloid precursor protein

(APP) gene in human neocortex. Brain Res Mol Brain Res 22:121-131.

Marlatt M, Lee HG, Perry G, Smith MA, Zhu X (2004) Sources and mechanisms of

cytoplasmic oxidative damage in Alzheimer's disease. Acta Neurobiol Exp (Wars)

64:81-87.

Martins LM, Kottke TJ, Kaufmann SH, Earnshaw WC (1998) Phosphorylated forms of

activated caspases are present in cytosol from HL-60 cells during etoposide-

induced apoptosis. Blood 92:3042-3049.

Martins RN, Harper CG, Stokes GB, Masters CL (1986) Increased cerebral glucose-6-

phosphate dehydrogenase activity in Alzheimer's disease may reflect oxidative

stress. J Neurochem 46:1042-1045.

Masciullo V, Scambia G, Marone M, Giannitelli C, Ferrandina G, Bellacosa A, Benedetti

Panici P, Mancuso S (1997) Altered expression of cyclin D1 and CDK4 genes in

ovarian carcinomas. Int J Cancer 74:390-395.

Masliah E, Terry R, Buzsaki G (1989) Thalamic nuclei in Alzheimer disease: evidence

against the cholinergic hypothesis of plaque formation. Brain Res 493:240-246.

Master E, Chan SL, Ali-Khan Z (1997) Ubiquitin (Ub) interacts non-covalently with

Alzheimer amyloid precursor protein (betaPP): isolation of Ub-betaPP conjugates

from brain extracts. Neuroreport 8:2781-2786.

176 Mattson MP, Fu W, Waeg G, Uchida K (1997) 4-Hydroxynonenal, a product of lipid

peroxidation, inhibits dephosphorylation of the microtubule-associated protein

tau. Neuroreport 8:2275-2281.

Maurer K, Volk S, Gerbaldo H (1997) Auguste D and Alzheimer's disease. Lancet

349:1546-1549.

McShea A, Wahl AF, Smith MA (1999a) Re-entry into the cell cycle: a mechanism for

neurodegeneration in Alzheimer disease. Med Hypotheses 52:525-527.

McShea A, Zelasko DA, Gerst JL, Smith MA (1999b) Signal transduction abnormalities

in Alzheimer's disease: evidence of a pathogenic stimuli. Brain Res 815:237-242.

McShea A, Harris PL, Webster KR, Wahl AF, Smith MA (1997) Abnormal expression of

the cell cycle regulators P16 and CDK4 in Alzheimer's disease. Am J Pathol

150:1933-1939.

McShea A, Samuel T, Eppel JT, Galloway DA, Funk JO (2000) Identification of CIP-1-

associated regulator of cyclin B (CARB), a novel p21-binding protein acting in

the G2 phase of the cell cycle. J Biol Chem 275:23181-23186.

Middleton G, Wyatt S, Ninkina N, Davies AM (2001) Reciprocal developmental changes

in the roles of Bcl-w and Bcl-x(L) in regulating sensory neuron survival.

Development 128:447-457.

Minami M, Jin KL, Li W, Nagayama T, Henshall DC, Simon RP (2000) Bcl-w

expression is increased in brain regions affected by focal cerebral ischemia in the

rat. Neurosci Lett 279:193-195.

Mirra SS, Heyman A, McKeel D, Sumi SM, Crain BJ, Brownlee LM, Vogel FS, Hughes

JP, van Belle G, Berg L (1991) The Consortium to Establish a Registry for

177 Alzheimer's Disease (CERAD). Part II. Standardization of the neuropathologic

assessment of Alzheimer's disease. Neurology 41:479-486.

Moller HJ, Graeber MB (1998) The case described by Alois Alzheimer in 1911.

Historical and conceptual perspectives based on the clinical record and

neurohistological sections. Eur Arch Psychiatry Clin Neurosci 248:111-122.

Montine TJ, Huang DY, Valentine WM, Amarnath V, Saunders A, Weisgraber KH,

Graham DG, Strittmatter WJ (1996) Crosslinking of apolipoprotein E by products

of lipid peroxidation. J Neuropathol Exp Neurol 55:202-210.

Morishima-Kawashima M, Hasegawa M, Takio K, Suzuki M, Yoshida H, Watanabe A,

Titani K, Ihara Y (1995) Hyperphosphorylation of tau in PHF. Neurobiol Aging

16:365-371; discussion 371-380.

Morsch R, Simon W, Coleman PD (1999) Neurons may live for decades with

neurofibrillary tangles. J Neuropathol Exp Neurol 58:188-197.

Munch G, Kuhla B, Luth HJ, Arendt T, Robinson SR (2003) Anti-AGEing defences

against Alzheimer's disease. Biochem Soc Trans 31:1397-1399.

Nagy Z, Esiri MM, Cato AM, Smith AD (1997) Cell cycle markers in the hippocampus

in Alzheimer's disease. Acta Neuropathol (Berl) 94:6-15.

Nakagawa T, Zhu H, Morishima N, Li E, Xu J, Yankner BA, Yuan J (2000) Caspase-12

mediates endoplasmic-reticulum-specific apoptosis and cytotoxicity by amyloid-

beta. Nature 403:98-103.

Neal KC, Pannuti A, Smith ER, Lucchesi JC (2000) A new human member of the MYST

family of histone acetyl with high sequence similarity to Drosophila

MOF. Biochim Biophys Acta 1490:170-174.

178 Neaves WB, Johnson L, Porter JC, Parker CR, Jr., Petty CS (1984) Leydig cell numbers,

daily sperm production, and serum gonadotropin levels in aging men. J Clin

Endocrinol Metab 59:756-763.

Nixon RA, Saito KI, Grynspan F, Griffin WR, Katayama S, Honda T, Mohan PS, Shea

TB, Beermann M (1994) Calcium-activated neutral proteinase (calpain) system in

aging and Alzheimer's disease. Ann N Y Acad Sci 747:77-91.

Nouspikel T, Hanawalt PC (2003) When parsimony backfires: neglecting DNA repair

may doom neurons in Alzheimer's disease. Bioessays 25:168-173.

Nunomura A, Perry G, Hirai K, Aliev G, Takeda A, Chiba S, Smith MA (1999a)

Neuronal RNA oxidation in Alzheimer's disease and Down's syndrome. Ann N Y

Acad Sci 893:362-364.

Nunomura A, Perry G, Pappolla MA, Wade R, Hirai K, Chiba S, Smith MA (1999b)

RNA oxidation is a prominent feature of vulnerable neurons in Alzheimer's

disease. J Neurosci 19:1959-1964.

Nunomura A, Perry G, Pappolla MA, Friedland RP, Hirai K, Chiba S, Smith MA (2000)

Neuronal oxidative stress precedes amyloid-beta deposition in Down syndrome. J

Neuropathol Exp Neurol 59:1011-1017.

Nunomura A, Perry G, Aliev G, Hirai K, Takeda A, Balraj EK, Jones PK, Ghanbari H,

Wataya T, Shimohama S, Chiba S, Atwood CS, Petersen RB, Smith MA (2001)

Oxidative damage is the earliest event in Alzheimer disease. J Neuropathol Exp

Neurol 60:759-767.

Obrenovich ME, Joseph JA, Atwood CS, Perry G, Smith MA (2002) Amyloid-beta: a

(life) preserver for the brain. Neurobiol Aging 23:1097-1099.

179 Olashaw N, Pledger WJ (2002) Paradigms of growth control: relation to Cdk activation.

Sci STKE 2002:RE7.

Ott M, Robertson JD, Gogvadze V, Zhivotovsky B, Orrenius S (2002) Cytochrome c

release from mitochondria proceeds by a two-step process. Proc Natl Acad Sci U

S A 99:1259-1263.

Pao GM, Janknecht R, Ruffner H, Hunter T, Verma IM (2000) CBP/p300 interact with

and function as transcriptional coactivators of BRCA1. Proc Natl Acad Sci U S A

97:1020-1025.

Park DS, Obeidat A, Giovanni A, Greene LA (2000) Cell cycle regulators in neuronal

death evoked by excitotoxic stress: implications for neurodegeneration and its

treatment. Neurobiol Aging 21:771-781.

Perez M, Hernandez F, Gomez-Ramos A, Smith M, Perry G, Avila J (2002) Formation of

aberrant phosphotau fibrillar polymers in neural cultured cells. Eur J Biochem

269:1484-1489.

Perry G, Smith MA (1998) Is oxidative damage central to the pathogenesis of Alzheimer

disease? Acta Neurol Belg 98:175-179.

Perry G, Friedman R, Shaw G, Chau V (1987) Ubiquitin is detected in neurofibrillary

tangles and senile plaque neurites of Alzheimer disease brains. Proc Natl Acad

Sci U S A 84:3033-3036.

Perry G, Nunomura A, Lucassen P, Lassmann H, Smith MA (1998) Apoptosis and

Alzheimer's disease. Science 282:1268-1269.

180 Perry G, Lipphardt S, Mulvihill P, Kancherla M, Mijares M, Gambetti P, Sharma S,

Maggiora L, Cornette J, Lobl T, et al. (1988) Amyloid precursor protein in senile

plaques of Alzheimer disease. Lancet 2:746.

Philipp-Staheli J, Payne SR, Kemp CJ (2001) p27(Kip1): regulation and function of a

haploinsufficient tumor suppressor and its misregulation in cancer. Exp Cell Res

264:148-168.

Pike CJ (1999) Estrogen modulates neuronal Bcl-xL expression and beta-amyloid-

induced apoptosis: relevance to Alzheimer's disease. J Neurochem 72:1552-1563.

Pines J, Hunter T (1994) The differential localization of human cyclins A and B is due to

a cytoplasmic retention signal in cyclin B. Embo J 13:3772-3781.

Podbilewicz B (1996) ADM-1, a protein with metalloprotease- and disintegrin-like

domains, is expressed in syncytial organs, sperm, and sheath cells of sensory

organs in Caenorhabditis elegans. Mol Biol Cell 7:1877-1893.

Pokrovskaja K, Ehlin-Henriksson B, Bartkova J, Bartek J, Scuderi R, Szekely L, Wiman

KG, Klein G (1996) Phenotype-related differences in the expression of D-type

cyclins in human B cell-derived lines. Cell Growth Differ 7:1723-1732.

Pope WB, Lambert MP, Leypold B, Seupaul R, Sletten L, Krafft G, Klein WL (1994)

Microtubule-associated protein tau is hyperphosphorylated during mitosis in the

human neuroblastoma cell line SH-SY5Y. Exp Neurol 126:185-194.

Poulin P, Zakzanis KK (2002) In vivo neuroanatomy of Alzheimer's disease: evidence

from structural and functional brain imaging. Brain Cogn 49:220-225.

181 Prakash SK, Van den Veyver IB, Franco B, Volta M, Ballabio A, Zoghbi HY (1999)

Characterization of a novel chromo domain gene in xp22.3 with homology to

Drosophila msl-3. Genomics 59:77-84.

Praprotnik D, Smith MA, Richey PL, Vinters HV, Perry G (1996) Plasma membrane

fragility in dystrophic neurites in senile plaques of Alzheimer's disease: an index

of oxidative stress. Acta Neuropathol (Berl) 91:1-5.

Pratico D, Uryu K, Leight S, Trojanoswki JQ, Lee VM (2001) Increased lipid

peroxidation precedes amyloid plaque formation in an animal model of Alzheimer

amyloidosis. J Neurosci 21:4183-4187.

Pratico D, Clark CM, Liun F, Rokach J, Lee VY, Trojanowski JQ (2002) Increase of

brain oxidative stress in mild cognitive impairment: a possible predictor of

Alzheimer disease. Arch Neurol 59:972-976.

Premkumar DR, Smith MA, Richey PL, Petersen RB, Castellani R, Kutty RK, Wiggert

B, Perry G, Kalaria RN (1995) Induction of heme oxygenase-1 mRNA and

protein in neocortex and cerebral vessels in Alzheimer's disease. J Neurochem

65:1399-1402.

Preuss U, Doring F, Illenberger S, Mandelkow EM (1995) Cell cycle-dependent

phosphorylation and microtubule binding of tau protein stably transfected into

Chinese hamster ovary cells. Mol Biol Cell 6:1397-1410.

Print CG, Loveland KL, Gibson L, Meehan T, Stylianou A, Wreford N, de Kretser D,

Metcalf D, Kontgen F, Adams JM, Cory S (1998) Apoptosis regulator bcl-w is

essential for spermatogenesis but appears otherwise redundant. Proc Natl Acad

Sci U S A 95:12424-12431.

182 Raina AK, Monteiro MJ, McShea A, Smith MA (1999a) The role of cell cycle-mediated

events in Alzheimer's disease. Int J Exp Pathol 80:71-76.

Raina AK, Takeda A, Nunomura A, Perry G, Smith MA (1999b) Genetic evidence for

oxidative stress in Alzheimer's disease. Neuroreport 10:1355-1357.

Raina AK, Zhu X, Rottkamp CA, Monteiro M, Takeda A, Smith MA (2000) Cyclin'

toward dementia: cell cycle abnormalities and abortive oncogenesis in Alzheimer

disease. J Neurosci Res 61:128-133.

Raina AK, Pardo P, Rottkamp CA, Zhu X, Pereira-Smith OM, Smith MA (2001a)

Neurons in Alzheimer disease emerge from senescence. Mech Ageing Dev 123:3-

9.

Raina AK, Hochman A, Ickes H, Zhu X, Ogawa O, Cash AD, Shimohama S, Perry G,

Smith MA (2003) Apoptotic promoters and inhibitors in Alzheimer's disease:

Who wins out? Prog Neuropsychopharmacol Biol Psychiatry 27:251-254.

Raina AK, Hochman A, Zhu X, Rottkamp CA, Nunomura A, Siedlak SL, Boux H,

Castellani RJ, Perry G, Smith MA (2001b) Abortive apoptosis in Alzheimer's

disease. Acta Neuropathol (Berl) 101:305-310.

Rall SC, Jr., Weisgraber KH, Mahley RW (1982) Human apolipoprotein E. The complete

amino acid sequence. J Biol Chem 257:4171-4178.

Ratel D, Boisseau S, Nasser V, Berger F, Wion D (2001) Programmed cell death or cell

death programme? That is the question. J Theor Biol 208:385-386.

Reynisdottir I, Massague J (1997) The subcellular locations of p15(Ink4b) and p27(Kip1)

coordinate their inhibitory interactions with cdk4 and cdk2. Genes Dev 11:492-

503.

183 Ricci JE, Gottlieb RA, Green DR (2003) Caspase-mediated loss of mitochondrial

function and generation of reactive oxygen species during apoptosis. J Cell Biol

160:65-75.

Riley RJ, Workman P (1992) DT-diaphorase and cancer chemotherapy. Biochem

Pharmacol 43:1657-1669.

Roth KA, Kuan C, Haydar TF, D'Sa-Eipper C, Shindler KS, Zheng TS, Kuida K, Flavell

RA, Rakic P (2000) Epistatic and independent functions of caspase-3 and Bcl-

X(L) in developmental programmed cell death. Proc Natl Acad Sci U S A 97:466-

471.

Ruffner H, Jiang W, Craig AG, Hunter T, Verma IM (1999) BRCA1 is phosphorylated at

serine 1497 in vivo at a cyclin-dependent kinase 2 phosphorylation site. Mol Cell

Biol 19:4843-4854.

Russell RL, Siedlak SL, Raina AK, Bautista JM, Smith MA, Perry G (1999) Increased

neuronal glucose-6-phosphate dehydrogenase and sulfhydryl levels indicate

reductive compensation to oxidative stress in Alzheimer disease. Arch Biochem

Biophys 370:236-239.

Salvesen GS, Dixit VM (1999) Caspase activation: the induced-proximity model. Proc

Natl Acad Sci U S A 96:10964-10967.

Santiard-Baron D, Lacoste A, Ellouk-Achard S, Soulie C, Nicole A, Sarasin A, Ceballos-

Picot I (2001) The amyloid peptide induces early genotoxic damage in human

preneuron NT2. Mutat Res 479:113-120.

184 Sayre LM, Zelasko DA, Harris PL, Perry G, Salomon RG, Smith MA (1997) 4-

Hydroxynonenal-derived advanced lipid peroxidation end products are increased

in Alzheimer's disease. J Neurochem 68:2092-2097.

Schreck R, Baeuerle PA (1991) A role for oxygen radicals as second messengers. Trends

Cell Biol 1:39-42.

Scully R, Chen J, Ochs RL, Keegan K, Hoekstra M, Feunteun J, Livingston DM (1997)

Dynamic changes of BRCA1 subnuclear location and phosphorylation state are

initiated by DNA damage. Cell 90:425-435.

Seidl, R., M. Bajo, et al. (1999). "Neuronal apoptosis inhibitory protein (NAIP)-like

immunoreactivity in brains of adult patients with Down syndrome." J Neural

Transm Suppl 57: 283-91.

Selznick LA, Holtzman DM, Han BH, Gokden M, Srinivasan AN, Johnson EM, Jr., Roth

KA (1999) In situ immunodetection of neuronal caspase-3 activation in

Alzheimer disease. J Neuropathol Exp Neurol 58:1020-1026.

Seril DN, Liao J, Yang GY, Yang CS (2003) Oxidative stress and ulcerative colitis-

associated carcinogenesis: studies in humans and animal models. Carcinogenesis

24:353-362.

Seril DN, Liao J, Ho KL, Warsi A, Yang CS, Yang GY (2002) Dietary iron

supplementation enhances DSS-induced colitis and associated colorectal

carcinoma development in mice. Dig Dis Sci 47:1266-1278.

Sheng JG, Zhou XQ, Mrak RE, Griffin WS (1998) Progressive neuronal injury associated

with amyloid plaque formation in Alzheimer disease. J Neuropathol Exp Neurol

57:714-717.

185 Sherr CJ (1996) Cancer cell cycles. Science 274:1672-1677.

Slooter AJ, van Duijn CM (1997) Genetic epidemiology of Alzheimer disease. Epidemiol

Rev 19:107-119.

Smith CD, Carney JM, Starke-Reed PE, Oliver CN, Stadtman ER, Floyd RA,

Markesbery WR (1991) Excess brain protein oxidation and enzyme dysfunction

in normal aging and in Alzheimer disease. Proc Natl Acad Sci U S A 88:10540-

10543.

Smith MA (1998) Alzheimer disease. Int Rev Neurobiol 42:1-54.

Smith MA, Perry G (1996) Alzheimer disease: protein-protein interaction and oxidative

stress. Bol Estud Med Biol 44:5-10.

Smith MA, Perry G (1997) The pathogenesis of Alzheimer disease: an alternative to the

amyloid hypothesis. J Neuropathol Exp Neurol 56:217.

Smith MA, Monnier VM, Sayre LM, Perry G (1995a) Amyloidosis, advanced glycation

end products and Alzheimer disease. Neuroreport 6:1595-1596.

Smith MA, Richey Harris PL, Sayre LM, Beckman JS, Perry G (1997) Widespread

peroxynitrite-mediated damage in Alzheimer's disease. J Neurosci 17:2653-2657.

Smith MA, Rudnicka-Nawrot M, Richey PL, Praprotnik D, Mulvihill P, Miller CA, Sayre

LM, Perry G (1995b) Carbonyl-related posttranslational modification of

neurofilament protein in the neurofibrillary pathology of Alzheimer's disease. J

Neurochem 64:2660-2666.

Smith MA, Kutty RK, Richey PL, Yan SD, Stern D, Chader GJ, Wiggert B, Petersen RB,

Perry G (1994a) Heme oxygenase-1 is associated with the neurofibrillary

pathology of Alzheimer's disease. Am J Pathol 145:42-47.

186 Smith MA, Taneda S, Richey PL, Miyata S, Yan SD, Stern D, Sayre LM, Monnier VM,

Perry G (1994b) Advanced Maillard reaction end products are associated with

Alzheimer disease pathology. Proc Natl Acad Sci U S A 91:5710-5714.

Smith TW, Lippa CF (1995) Ki-67 immunoreactivity in Alzheimer's disease and other

neurodegenerative disorders. J Neuropathol Exp Neurol 54:297-303.

Sonoda Y, Yoshimoto T, Sekiya T (1995) Homozygous deletion of the MTS1/p16 and

MTS2/p15 genes and amplification of the CDK4 gene in glioma. Oncogene

11:2145-2149.

Sperandio S, de Belle I, Bredesen DE (2000) An alternative, nonapoptotic form of

programmed cell death. Proc Natl Acad Sci U S A 97:14376-14381.

Stadelmann C, Deckwerth TL, Srinivasan A, Bancher C, Bruck W, Jellinger K,

Lassmann H (1999) Activation of caspase-3 in single neurons and autophagic

granules of granulovacuolar degeneration in Alzheimer's disease. Evidence for

apoptotic cell death. Am J Pathol 155:1459-1466.

Stennicke HR, Jurgensmeier JM, Shin H, Deveraux Q, Wolf BB, Yang X, Zhou Q,

Ellerby HM, Ellerby LM, Bredesen D, Green DR, Reed JC, Froelich CJ, Salvesen

GS (1998) Pro-caspase-3 is a major physiologic target of caspase-8. J Biol Chem

273:27084-27090.

Steusloff K, Rocken C, Saeger W (1998) Basement membrane proteins, apolipoprotein E

and glycosaminoglycans in pituitary adenomas and their correlation to amyloid.

Virchows Arch 433:29-34.

Stewart WF, Kawas C, Corrada M, Metter EJ (1997) Risk of Alzheimer's disease and

duration of NSAID use. Neurology 48:626-632.

187 Strittmatter WJ, Weisgraber KH, Huang DY, Dong LM, Salvesen GS, Pericak-Vance M,

Schmechel D, Saunders AM, Goldgaber D, Roses AD (1993) Binding of human

apolipoprotein E to synthetic amyloid beta peptide: isoform-specific effects and

implications for late-onset Alzheimer disease. Proc Natl Acad Sci U S A 90:8098-

8102.

Subbarao KV, Richardson JS, Ang LC (1990) Autopsy samples of Alzheimer's cortex

show increased peroxidation in vitro. J Neurochem 55:342-345.

Takasugi N, Tomita T, Hayashi I, Tsuruoka M, Niimura M, Takahashi Y, Thinakaran G,

Iwatsubo T (2003) The role of presenilin cofactors in the gamma-secretase

complex. Nature 422:438-441.

Takeda A, Smith MA, Avila J, Nunomura A, Siedlak SL, Zhu X, Perry G, Sayre LM

(2000) In Alzheimer's disease, heme oxygenase is coincident with Alz50, an

epitope of tau induced by 4-hydroxy-2-nonenal modification. J Neurochem

75:1234-1241.

Terry RD, Gonatas NK, Weiss M (1964) Ultrastructural Studies in Alzheimer's Presenile

Dementia. Am J Pathol 44:269-297.

Thinakaran G, Harris CL, Ratovitski T, Davenport F, Slunt HH, Price DL, Borchelt DR,

Sisodia SS (1997) Evidence that levels of presenilins (PS1 and PS2) are

coordinately regulated by competition for limiting cellular factors. J Biol Chem

272:28415-28422.

Thomas JE, Smith M, Tonkinson JL, Rubinfeld B, Polakis P (1997) Induction of

phosphorylation on BRCA1 during the cell cycle and after DNA damage. Cell

Growth Differ 8:801-809.

188 Torp R, Su JH, Deng G, Cotman CW (1998) GADD45 is induced in Alzheimer's disease,

and protects against apoptosis in vitro. Neurobiol Dis 5:245-252.

Trucco C, Oliver FJ, de Murcia G, Menissier-de Murcia J (1998) DNA repair defect in

poly(ADP-ribose) polymerase-deficient cell lines. Nucleic Acids Res 26:2644-

2649.

Tsang SY, Tam SC, Bremner I, Burkitt MJ (1996) Research communication copper-1,10-

phenanthroline induces internucleosomal DNA fragmentation in HepG2 cells,

resulting from direct oxidation by the hydroxyl radical. Biochem J 317 ( Pt 1):13-

16.

Tseng HC, Zhou Y, Shen Y, Tsai LH (2002) A survey of Cdk5 activator p35 and p25

levels in Alzheimer's disease brains. FEBS Lett 523:58-62.

Venugopal R, Jaiswal AK (1996) Nrf1 and Nrf2 positively and c-Fos and Fra1 negatively

regulate the human antioxidant response element-mediated expression of

NAD(P)H:quinone oxidoreductase1 gene. Proc Natl Acad Sci U S A 93:14960-

14965.

Verhagen AM, Vaux DL (2002) Cell death regulation by the mammalian IAP antagonist

Diablo/Smac. Apoptosis 7:163-166.

Vincent I, Rosado M, Davies P (1996) Mitotic mechanisms in Alzheimer's disease? J Cell

Biol 132:413-425.

Vincent I, Jicha G, Rosado M, Dickson DW (1997) Aberrant expression of mitotic

cdc2/cyclin B1 kinase in degenerating neurons of Alzheimer's disease brain. J

Neurosci 17:3588-3598.

189 Waters SI, White JM (1997) Biochemical and molecular characterization of bovine

fertilin alpha and beta (ADAM 1 and ADAM 2): a candidate sperm-egg

binding/fusion complex. Biol Reprod 56:1245-1254.

Weinberg RA (1996) The molecular basis of carcinogenesis: understanding the cell cycle

clock. Cytokines Mol Ther 2:105-110.

Williamson R, Scales T, Clark BR, Gibb G, Reynolds CH, Kellie S, Bird IN, Varndell

IM, Sheppard PW, Everall I, Anderton BH (2002) Rapid tyrosine phosphorylation

of neuronal proteins including tau and focal adhesion kinase in response to

amyloid-beta peptide exposure: involvement of Src family protein kinases. J

Neurosci 22:10-20.

Wilson CA, Ramos L, Villasenor MR, Anders KH, Press MF, Clarke K, Karlan B, Chen

JJ, Scully R, Livingston D, Zuch RH, Kanter MH, Cohen S, Calzone FJ, Slamon

DJ (1999) Localization of human BRCA1 and its loss in high-grade, non-inherited

breast carcinomas. Nat Genet 21:236-240.

Wilson DM, Binder LI (1997) Free fatty acids stimulate the polymerization of tau and

amyloid beta peptides. In vitro evidence for a common effector of pathogenesis in

Alzheimer's disease. Am J Pathol 150:2181-2195.

Wisniewski HM, Narang HK, Terry RD (1976) Neurofibrillary tangles of paired helical

filaments. J Neurol Sci 27:173-181.

Wolfel T, Hauer M, Schneider J, Serrano M, Wolfel C, Klehmann-Hieb E, De Plaen E,

Hankeln T, Meyer zum Buschenfelde KH, Beach D (1995) A p16INK4a-

insensitive CDK4 mutant targeted by cytolytic T lymphocytes in a human

melanoma. Science 269:1281-1284.

190 Wolfsberg TG, Primakoff P, Myles DG, White JM (1995) ADAM, a novel family of

membrane proteins containing A Disintegrin And Metalloprotease domain:

multipotential functions in cell-cell and cell-matrix interactions. J Cell Biol

131:275-278.

Yan SD, Chen X, Schmidt AM, Brett J, Godman G, Zou YS, Scott CW, Caputo C,

Frappier T, Smith MA (1994) Glycated tau protein in Alzheimer disease: a

mechanism for induction of oxidant stress. Proc Natl Acad Sci U S A 91:7787-

7791.

Yan SD, Yan SF, Chen X, Fu J, Chen M, Kuppusamy P, Smith MA, Perry G, Godman

GC, Nawroth P, et al. (1995) Non-enzymatically glycated tau in Alzheimer's

disease induces neuronal oxidant stress resulting in cytokine gene expression and

release of amyloid beta-peptide. Nat Med 1:693-699.

Yan SD, Chen X, Fu J, Chen M, Zhu H, Roher A, Slattery T, Zhao L, Nagashima M,

Morser J, Migheli A, Nawroth P, Stern D, Schmidt AM (1996) RAGE and

amyloid-beta peptide neurotoxicity in Alzheimer's disease. Nature 382:685-691.

Zheng W, Johnston SA (1998) The nucleic acid binding activity of bleomycin hydrolase

is involved in bleomycin detoxification. Mol Cell Biol 18:3580-3585.

Zhu X, Ogawa O, Wang Y, Perry G, Smith MA (2003) JKK1, an upstream activator of

JNK/SAPK, is activated in Alzheimer's disease. J Neurochem 85:87-93.

Zhu X, Rottkamp CA, Boux H, Takeda A, Perry G, Smith MA (2000) Activation of p38

kinase links tau phosphorylation, oxidative stress, and cell cycle-related events in

Alzheimer disease. J Neuropathol Exp Neurol 59:880-888.

191 Zhu X, Raina AK, Lee HG, Casadesus G, Smith MA, Perry G (2004a) Oxidative stress

signalling in Alzheimer's disease. Brain Res 1000:32-39.

Zhu X, Castellani RJ, Takeda A, Nunomura A, Atwood CS, Perry G, Smith MA (2001)

Differential activation of neuronal ERK, JNK/SAPK and p38 in Alzheimer

disease: the 'two hit' hypothesis. Mech Ageing Dev 123:39-46.

Zhu X, Wang Y, Ogawa O, Lee HG, Raina AK, Siedlak SL, Harris PL, Fujioka H,

Shimohama S, Tabaton M, Atwood CS, Petersen RB, Perry G, Smith MA (2004b)

Neuroprotective properties of Bcl-w in Alzheimer disease. J Neurochem 89:1233-

1240.

192