NUCLEAR RECEPTOR ACTIVATION AND ALZHEIMER’S DISEASE PATHOGENESIS

By

PAIGE ELIZABETH CRAMER

Submitted in partial fulfillment of the requirements

For the degree of Doctor of Philosophy

Dissertation Advisor: Dr. Gary Landreth

Department of Neurosciences

CASE WESTERN RESERVE UNIVERSITY

May 2012

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Paige Elizabeth Cramer

Doctor of Philosophy Candidate for the ______degree*.

Evan Deneris (Signed )______(Chair of the committee)

Gary Landreth (signed)______

Stephen Maricich ______

Jerry Silver ______

Bruce Lamb ______

March 19,2012 (date) ______

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

Dedication Page

To Samantha Rodwicz, for forcing me to believe in scientific advancement and to hope

for a cure for spinal cord injury.

i

Table of Contents

Dedication Page ...... i

Table of Contents ...... ii

Table of Figures ...... v

Acknowledgements ...... vii

Abstract...... 1

Chapter 1:Introduction...... 3

Introduction ...... 4

History of Alzheimer’s disease ...... 4

Amyloid Precursor Protein and Aβ processing ...... 6

Genetics of AD and Late onset Alzheimer’s disease ...... 7

Amyloid hypothesis and Aβ homeostasis ...... 9

Clearance...... 11

Soluble versus Insoluble Aβ and Cognition ...... 14

Microglia and the undiseased brain ...... 16

Microglia in the AD brain ...... 18

Inflammation: ...... 22

Microglia and Aβ binding: ...... 23

NSAIDs ...... 24

Activation Status of Microglia ...... 26

Nuclear Receptors ...... 30 ii

Retionid X Receptors (RXRs)...... 33

Peroxisome Proliferator Activated Receptors (PPARs) ...... 34

Liver X Receptors (LXRs) ...... 35

ApolipoproteinE (ApoE) and Alzheimer’s Disease ...... 36

Nuclear Receptors and Alzheimer’s Disease mouse models ...... 38

Liver X Receptors ...... 39

Peroxisome Proliferator Activated Receptor ...... 40

Retinoid X Receptor ...... 42

Retinoic Acid Receptor ...... 44

Current Treatments for AD: ...... 45

Focus of the Thesis ...... 46

Chapter 2 :Ibuprofen attenuates oxidative damage through NOX2 inhibition in

Alzheimer’s Disease...... 58

Abstract ...... 59

Introduction ...... 60

Materials and Methods ...... 62

Results ...... 66

Discussion ...... 73

Chapter 3 : ApoE-directed Therapeutics Rapidly Clear β-amyloid and Reverse Deficits in

AD Mouse Models...... 93

Abstract: ...... 95

iii

Introduction: ...... 96

Materials and Methods: ...... 99

Results ...... 106

Discussion ...... 113

Chapter 4 : Discussion...... 138

NSAIDs as PPARγ agonists: ...... 141

Polarization of microglia through PPARγ response elements and other pathways: ...... 143

Infiltration of monocytes and macrophages ...... 150

Long term bexarotene treatment ...... 152

RXR and cytochrome P450 ...... 153

Aβ load and Cognitive Impairment Reversal ...... 156

SUMO ...... 157

RXR and ApoE isoforms ...... 159

RXR activation in astrocytes ...... 160

Nuclear Receptor activation and other diseases ...... 161

Clinical Trials for AD ...... 163

Conclusions ...... 164

Chapter 5: Works Cited...... 172

iv

Table of Figures

Figure 1-1: APP Processing ...... 48

Figure 1-2: Aβ aggregation and deposition ...... 50

Figure 1-3: Intracellular and Extracellular Clearance of sAβ ...... 52

Figure 1-4: Nuclear Receptor Activation and Repression ...... 54

Figure 1-5:LXR and PPAR are permissive binding partners with RXR ...... 56

Figure 2-1: Chronic ibuprofen treatment reduces AD-related plaque pathology in B6-

R1.40 mice ...... 79

Figure 2-2: Chronic ibuprofen treatment reduces microglial activation in B6-R1.40 mice

...... 81

Figure 2-3: Ibuprofen treatment reduces AD-related oxidative damage.: ...... 83

Figure 2-4: Chronic ibuprofen treatment reduced protein oxidation in aged B6-R1.40

mice: ...... 85

Figure 2-5: Fibrillar Aβ-stimulated Vav phosphorylation is inhibited by ibuprofen pretreatment...... 87

Figure 2-6: S-ibuprofen disrupts NOX2 complex assembly...... 89

Figure 2-7: S-ibuprofen inhibits the generation of NOX2-derived radicals in microglia stimulated with fAβ ...... 91

Figure 3-1: Bexarotene stimulates the expression of LXR target genes...... 116

Figure 3-2: Bexarotene stimulates the ApoE-dependent intracellular clearance of Aβ through the actions of LXR and PPARγ ...... 118

Figure 3-3:ISF levels of Aβ decrease after bexarotene treatment...... 120

Figure 3-4: Bexarotene stimulates the expression of LXR target genes in vivo...... 122

v

Figure 3-5: Aβ levels and plaque burden are reduced by bexarotene treatment...... 124

Figure 3-6: Short-term treatment of bexarotene in 11 month old APP/PS1 mice stimulates

clearance of Aβ ...... 126

Figure 3-7:Chronic bexarotene treatment reduces levels of soluble Aβ...... 128

Figure 3-8:Bexarotene treatment of an aggressive amyloidogenic mouse model stimulates

the clearance of Aβ ...... 130

Figure 3-9: Restoration of memory and cognition with bexarotene treatment...... 132

Figure 3-10: Rescue of cortical network activity with bexarotene...... 134

Figure 3-11: Improvement of odor habituation behavior in 12-14mo old Tg2576 mice treated with bexarotene for 3 days ...... 136

4-1 Time course of Reverse Cholesterol Transport proteins ...... 166

4-2 One month study: 14 consecutive/14 every other day Bexarotene treatment ...... 168

4-3 Bexarotene and ApoE isoform ...... 170

vi

Acknowledgements This project could not have been completed without the help from many different

people. First, I would like to thank my advisor, Dr. Gary Landreth. His enthusiasm for

science is infectious and his commitment to his graduate students and post doctoral

fellows is remarkable. Thank you, Gary, for helping me define myself as a scientist and

for the guidance and support along the way. I would also like to thank my committee

members, Drs. Evan Deneris, Jerry Silver, Bruce Lamb and Stephen Maricich. Their interest and support throughout my graduate career always made my committee meetings

enjoyable discussions. In addition to my advisor and committee members, many thanks

go to the Alzheimer’s Research Laboratory members, both past and present. Members have provided a wonderful environment for learning and collaborating scientifically, and have made for some treasured friendships in and outside of the lab. Specifically, many thanks go to Dr. Brandy Wilkinson. Brandy taught me how to be a scientist and how to perform many of the techniques used in my projects. Though brutally honest, her advice

both scientific and personal, I value deeply. I am very much indebted to her. She was

pivotal in both the ibuprofen project and the initial stages of the bexarotene project.

Similarly, Drs. Quingguang Jiang and Daniel C.Y. Lee provided many hours worth of

discussion. The bexarotene project would not have existed had QJ not determined

apoE’s involvement in Alzhiemer’s disease. I am forever grateful and thankful. Without

Donna Kirsch and Colleen Karlo, no scientific progress could have been made in the

laboratory. Colleen, you’ve become one of my most treasured friends in Cleveland.

Thank you for all of our conversations and advice. To the rest of the Alzheimer’s

laboratory, nowhere else have I known people to be so happy to be at work and willing to

and help others and share information.

vii

Ultimately, the support of my friends and family has made this possible. A special thank you goes to my parents, Harry and Gayle Cramer, without whom I never would have believed I had the ability to take such a big bite and finish. Thank you for believing in me and giving me support and encouragement any time and all the time. To my brother and sister in-law, Todd and Payal Cramer, thank you for your unwavering support, constant encouragement and willingness to accept phone calls at all hours. I owe you. To James Harris, thank you for our never ending discussions of science and helping me figure out the big picture. Thank you for teaching me to enjoy the act of relaxing and participating in things outside of the laboratory. I am forever appreciative of your support and encouragement through the highs and lows of graduate school.

viii

Nuclear Receptor Activation and Alzheimer’s Disease Pathogenesis

Abstract:

By

PAIGE ELIZABETH CRAMER

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder,

characterized by the presence of β-amyloid (Aβ) plaques and neurofibrillary tangles, and

is accompanied by a robust inflammatory response that clinically presents as the progressive loss of cognition and memory, ultimately leading to death. The disease is associated with a disruption of Aβ homeostasis and it’s deposition in the brain, which initiates a microglial-mediated immune response that produces pro-inflammatory cytokines and reactive nitrogen and oxygen species.

Using ibuprofen, a peroxisome proliferator activated receptor γ (PPARγ) agonist and non-steroidal anti-inflammatory (NSAID), we demonstrate that chronic ibuprofen treatment reduces Aβ burden by 90% in the parenchyma and prevents microglia, the primary producers of reactive oxygen and nitrogen species, from forming the NADPH oxidase. Ibuprofen prevents the phosphorylation of Vav and consequently RAC1 translocation to the plasma membrane, preventing NADPH oxidase assembly, thus inhibiting the release of superoxide radicals. Consequently, we report less oxidative damage in the brains of mice treated with ibuprofen.

We further extend our studies to assess the mechanism of Aβ removal from the

brain and the behavioral outcomes in mouse models of AD, by using a retinoid X

receptor (RXR) selective, FDA approved agonist, bexarotene. We show that bexarotene

utilizes its ability and obligation to heterodimerize with Liver X Receptor (LXR) to

1 induce the expression of LXR response element genes, apolipoprotein E, ABCA1 and

ABCG1, and elevating HDL levels in the brains of treated AD mice. We show that following 1 dose of bexarotene, soluble Aβ levels are reduced for up to 84 hours. We show that after just 72 hours of bexarotene treatment, the plaque burden is reduced by

50%. The consequence of reducing soluble Aβ levels is correlative to ameliorating the

AD-related behavioral deficits in three different AD mouse models and reversing a neural circuit deficit.

Together our data demonstrate the important roles for nuclear receptors in preventing the inflammation associated with AD and facilitating clearance of Aβ species.

These data may provide a novel therapeutic strategy for the treatment of the neurodegenerative disease and its prodromal states.

2

Chapter 1:Introduction

3

Introduction

Alzheimer’s disease (AD) is a complex neurodegenerative disorder. Aging, genetics and environmental factors, among others, have been shown to contribute to disease pathogenesis. The complexity of the disease presents substantial challenges to development of therapeutics. However, though the pathological hallmarks of the disease are known, the underlying cause of their pathology remains unclear and currently a cure,

an effective treatment, or a preventative therapy for AD does not exist.

History of Alzheimer’s disease

Alzheimer’s disease is the leading cause of dementia in the elderly, accounting for

60-80% of all cases of dementia (Schneider, Arvanitakis et al. 2009). In 2011, it was

estimated that 5.4 million Americans had the disease. Age is the greatest risk factor for

developing AD, though it’s not a normal part of aging. Medical advances and the

consequential increase in life expectancy have led to the increase in the number of cases

of AD. One in eight people over the age of 65 and about 50% of people over the age of

85 have AD. As a large segment of the American population—the baby boomers— is

approaching 65, it is expected that the number of cases of AD will continue to rise. By

2030, the number of people over 65 with the disease is estimated to be 7.7 million. By

2050, 11-16 million Americans may be affected by the disease (Hebert, Scherr et al.

2003). The cost of AD in 2011 was close to $200 billion dollars and is projected to rise

to $1.1 trillion dollars by 2050, barring any therapeutic advances (Alzheimer’s

Association.Changing the Trajectory of Alzheimer’s Disease: A National Imperative.

2010 www.alz.org).

4

Other risk factors for developing Alzheimer’s disease besides age include,

possession of the ApoE4 allele (Corder, Saunders et al. 1993; Katzman 1994; Bertram

and Tanzi 2008), gender (Seshadri, Wolf et al. 1997), head trauma (van Duijn, Tanja et

al. 1992), stroke (Gorelick 2004; Wolf 2012), education level (Evans, Hebert et al. 1997;

Kukull, Higdon et al. 2002; Evans, Bennett et al. 2003) metabolic disorders, especially,

metabolic syndrome (Farooqui, Farooqui et al. 2011), coronary artery disease (Olichney,

Sabbagh et al. 1997; Beach, Maarouf et al. 2011) and type II diabetes mellitus (Hoyer

2002; Arvanitakis, Wilson et al. 2004).

Once diagnosed with dementia of the Alzheimer’s type, the disease progresses

over an average of 8-12 years (Ganguli, Dodge et al. 2005; Helzner, Scarmeas et al.

2008) and is clinically characterized by progressive cognitive decline including memory

loss, loss of language and motor skills, deteriorating executive function, spatial

orientation and behavior, eventually culminating in death.

What is now known as Alzheimer’s disease was first reported in 1907 by Dr.

Alois Alzheimer. Dr. Alzheimer presented a patient, Augusta Deter who was presented

to him with the progressive loss of cognition. From his studies of her brain, he identified

what are now the hallmarks of the disease: intracellular tangles of proteins within

neurons, known as neurofibrillary tangles (NFTs) and dense extracellular plaques.

NFTs are comprised of the accumulation of hyper-phosphorylated forms of the microtubule associated protein, tau (Kosik, Joachim et al. 1986; Wood, Mirra et al.

1986). Microtubules are necessary for axonal transport and the structural stability of neuronal processes. Impairments in axonal transport are believed to contribute to the neuronal degeneration typified in the disease (Zempel and Mandelkow 2011).

5

The dense core plaques or amyloid plaques that Dr. Alzheimer’s described are

found throughout the cortex and hippocampus. The cerebellum and thalamus are typically

spared in the disease. In addition, these plaques can be seen in the cerebral vasculature,

known as cerebral amyloid angiopathy (CAA) that accompanies AD. Amyloid plaques

are extracellular deposits of amyloid beta (Aβ) peptides and are identified in autopsy

sections by staining using lipophilic dyes, by silver impregnation or more recently in alive humans by positron emission tomography (PET) Aβ tracers such as Pittsburgh

Compound B (Klunk, Engler et al. 2004).

Amyloid Precursor Protein and Aβ processing

Alzheimer’s disease is typified by the progressive deposition of amyloid beta

(Aβ) in the brain. It took nearly 80 years after A. Alzheimer described the amyloid plaques to identify the 4 KiloDalton protein that comprises the pathological feature of

patients with AD and Down’s Syndrome (Glenner and Wong 1984; Glenner and Wong

1984). As all people with Downs’s Syndrome unfailingly become afflicted with

Alzheimer’s disease, Glenner had predicted the protein causing the plaques was found on

the triplicated chromosome 21 found in Down’s Syndrome. Three years later, it was

determined that Aβ was derived from a type 1 transmembrane protein, termed the amyloid precursor protein (APP), which mapped, as Glenner predicted, on chromosome

21(Goldgaber, Lerman et al. 1987; Goldgaber, Lerman et al. 1987; Kang, Lemaire et al.

1987; Podlisny, Lee et al. 1987; Tanzi, Haines et al. 1988).

Though known to be involved in synapse function and repair (Priller, Bauer et al.

2006), APP function largely remains unclear. We do know that APP can be processed to form the pathogenic species, Aβ or a non-pathogenic species. After sorting in the

6

endoplasmic reticulum and Golgi, APP is trafficked to the synaptic terminal. The non-

amyloidogenic pathway is initiated by α-secretase, which cleaves within the Aβ domain

and releases the large soluble N-terminal fragment (sAPPα) and the membrane-bound C- terminal fragment (α-CTF) (Lammich, Kojro et al. 1999). The C-terminal fragment is then cleaved by γ-secretase into a soluble N-terminal fragment (p3) and the APP intracellular domain (AICD). γ-secretase is composed of 4 different membrane associated proteins, including nicastrin, Aph-1, presenilins (PSEN1 and PSEN2) and

Pen-2 with the presenilins as the catalytically active subunits (Grimm, Tomic et al. 2002;

Sisodia and St George-Hyslop 2002). Alternatively, the amyloidogenic cleavages of

APP are initiated by β-secretase (BACE1), generating the soluble N-terminal β-secreted

APP (sAPPβ) and the membrane-bound fragment (β-CTF). BACE1 is a membrane bound aspartic protease. γ-secretase releasesthe amyloidogenic Aβ peptide from CTF- beta, which varies in length between 38 and 43 amino acids, and the AICD (Haass,

Schlossmacher et al. 1992; Haass, Hung et al. 1993; Vassar, Bennett et al. 1999; Selkoe

2001) (Figure 1). It is interesting to note that secretase activity is regulated by cellular lipid content, where β-secretase mainly cleavage occurs in the cholesterol-rich lipid rafts, while α-secretase activity favors non-raft membranes (Simons, Keller et al. 1998; Cordy,

Hussain et al. 2003; Ehehalt, Keller et al. 2003; Osenkowski, Ye et al. 2008).

Genetics of AD and Late onset Alzheimer’s disease

The familial, genetically inherited forms of the disease arise from mutations in three genes: APP, Presenilin1 (PSEN1) and Presenilin 2 (PSEN2). The mutations in the

APP gene are clustered near the cleavage sites of the β- and γ-secretase and mutations in the γ-secretase itself, both favoring the production of Aβ. The autosomally dominant

7 familial form of AD, otherwise known as early onset familial AD (EOAD), manifests itself in patients at a much younger age, between 30-60 years of old. To date, there are

32 mutations in APP, 185 mutations in PSEN1 and 13 mutations in PSEN2

(http://www.molgen.ua.ac.be/ADmutations/). Similarly, Down’s syndrome patients have an extra copy of chromosome 21, therefore an extra copy of the APP gene, and all develop AD between the ages of 40-50, further indicating APP’s involvement in disease etiology (Wisniewski, Dalton et al. 1985).

One interesting recent finding is that APP, in addition to carrying mutations that favor Aβ production, can be SUMOylated. Sumoylation is the covalent addition of a small ubiquitin-like modifier to lysine residues. Zhang and Sarge determined that lysines

587 and 595 in APP were targets of SUMO1 and that SUMOylation of APP led to decreased Aβ production. The locations of the lysines are 1 and 9 residues N-terminal to the β-secretase cleavage sites, respectively (Zhang and Sarge 2008). The Swedish mutation of APP, one of the familial, genetic dominant mutations, alters lysine 595 to asparagine, preventing SUMOylation and promoting β-secretase cleavage and the production of Aβ (Mullan, Crawford et al. 1992).

Despite only accounting for <5% of all cases of AD, many of these familial mutations have been crucial to our understanding of the disease. Many transgenic animals have been created using these human mutations to model the disease, allowing for further study.

The predominant form of Alzheimer’s disease is referred to as late onset,

‘sporadic’ AD (LOAD) and accounts for >95% of all cases of the disease. Genome-wide

8

linkage studies on LOAD have provided strong evidence for the association of many

genes with AD. Specifically,the apolipoprotein E4 (apoE4) allele is the single most

significant risk factor for AD (Strittmatter, Saunders et al. 1993; Jones, Holmans et al.

2010) and will be discussed in a later section.

Amyloid hypothesis and Aβ homeostasis

The most widely held hypothesis for AD pathogenesis is known as the “amyloid

cascade hypothesis.” It proposes that the deposition of Aβ is the causative agent of AD

pathology, and that the neurofibrillary tangles, cell loss, vascular damage, and dementia

are a consequence of the Aβ deposition (Hardy and Higgins 1992). Selkoe suggested that

the deposition of Aβ was the result of a gradual and chronic imbalance in the production

and/or clearance of Aβ, leading to an increase in its steady state levels within the brain

over the course of decades (Selkoe 2000). Thus, this theory posits that an excess of Aβ

leads to multimerization and fibrillarization, eventually forming the plaques that are

characteristic of AD.

The Aβ peptide is generated by the sequential cleavages of APP by β- and γ-

secretases, resulting in the production of peptides between 38-43 amino acids, typically

40 or 42 amino acids in length (Sisodia and Price 1995). Aβ is a self-associating protein

that can sequentially form dimers, trimers, a largely ill-defined, poorly characterized set of species of Aβ, including large oligomers, Aβ-derived diffusible ligands (ADDLs) and protofibrils, and eventually they form fibrils and plaques. The two hydrophobic amino acids that are present in Aβ42 and not Aβ40, lends itself to fibrillogensis (Standridge

2006). This fibrillogensis alters the C-terminus to form β-pleated sheets, which promotes

9

aggregation to form both diffuse and dense core plaques in the cortex and hippocampus

of the brain (Figure 2).

Amyloid plaque development and growth has been studied by Yan et al. and

others including Meyer-Luehmann et al., using serial in vivo multiphoton imaging. Yan

et al. were able to visualize the growth of individual plaques over a period of time in

different aged mice. They demonstrated that the plaque growth was more exponential in

younger transgenic mice when compared to older mice. Smaller plaques grew at a faster

rate in comparison to larger plaques. In addition, the authors used Compound E, a γ-

secretase inhibitor to shut down the production of Aβ. After Compound E exposure, the

growth rate of plaques was halted in pre-existing plaques and prevented new plaques

from forming. More importantly, the authors concluded that small changes in interstitial

fluid levels (ISF) of Aβ, had dramatic effects on plaque production and growth,

suggesting the importance of extracellular soluble Aβ pools (Yan, Bero et al. 2009).

Similarly, Hefendehl and her colleagues were able to monitor over 6 months the

appearance and development of Aβ plaques using multiphoton in vivo imaging. The

authors concluded that at 4-5 months of age, APPPS1-21 mice formed approximately 35

new plaques per cubic millimeter of neocortical volume per week. At a later age, in the

presence of plaques, the rate of appearance of newly formed plaques decreased. In

contrast to the previous study, Hefendehl et al., show that on average, both newly formed

and existing plaques grew at a similar rate per week (Hefendehl, Wegenast-Braun et al.

2011).

10

Clearance

Bateman and colleagues showed that in non-diseased humans, Aβ is generated and cleared at about equivalent rates of 7.6% and 8.3%, respectively (Bateman, Munsell et al. 2006). This suggested that even modest reductions in clearance of soluble Aβ could result in elevated levels of Aβ peptides and their aggregation into deposits. Up until the end of 2010, it was unknown whether LOAD was caused by an increase in production of

Aβ or a deficiency in clearing Aβ. Mawuenyega and colleagues determined that while the production of Aβ remained the same in the humans they tested, diseased or normal, the clearance of Aβ was hindered in the AD patients by 30% (Mawuenyega, Sigurdson et al. 2010). This discovery has prompted a renewed interest in developing pharmacological agents to enhance Aβ clearance.

There are two principle mechanisms for removal of Aβ from the brain: efflux of intact soluble Aβ (sAβ) to the peripheral circulation and the proteolytic degradation of both sAβ and fibrillar Aβ (fAβ).

Efflux of sAβ from the brain to the periphery can occur through a number of different mechanisms, including efflux across the blood brain barrier (BBB) mediated by the low density lipoprotein receptor-related protein-1 (LRP1), interstitial fluid (ISF)- cerebrospinal fluid (CSF) bulk flow into the lymphatic system and transport via the P- glycoprotein (PgP) efflux pump across the BBB. Other mechanism exist but are the beyond the scope of this thesis.

LRP1 is the most well-studied receptor for facilitating clearance of soluble Aβ.

LRP1 is an apoE receptor and is expressed at high levels in the CNS (Deane and

11

Zlokovic 2007). It has preferentially been shown to mediate Aβ efflux alone (Shibata,

Yamada et al. 2000) and in conjunction with ApoE (Deane and Zlokovic 2007) across

the brain capillary endothelia (Deane, Wu et al. 2004). Recently, it has been shown that

apoE-Aβ efflux efficiency is isoform dependent, such that apoE4;Aβ efflux is much slower than either apoE3:Aβ or apoE2:Aβ (Deane, Sagare et al. 2008). Additionally, it was shown that LRP1 expression on astrocytes mediates the efflux of Aβ-apoE into the perivascular space (Rolyan, Feike et al. 2011). Interestingly, studies in mice indicate that the expression of LRP-1 in microvessels decreases with age and that there is a corresponding decrease in amyloid clearance (Shibata, Yamada et al. 2000). Slow continuous drainage along perivascular basement membranes into the CSF and eventually into the bloodstream accounts for about 15% of Aβ clearance from the normal brain

(Shibata et al., 2000). PgP has also been implicated in facilitating transport of Aβ across the BBB. PgP is a member of the ATP-binding cassette subfamily B (ABCB) and serves as an efflux pump that is highly expressed on endothelial cells of the BBB (Deane and

Zlokovic 2007). Studies have shown that AD transgenic/PgP deficient mice clear Aβ at

half the rate of wild-type mice and have a corresponding increase in plaque burden

(Cirrito, Yamada et al. 2005).

Many cell types within the brain have been reported to take up Aβ species,

including microglia, astrocytes and vascular endothelial cells (Tanzi, Moir et al. 2004). It

has been shown by others and our lab that uptake of sAβ can occur through fluid phase

pinocytosis or by phagocytosis of fAβ (Paresce, Chung et al. 1997; Chung, Brazil et al.

1999; Koenigsknecht and Landreth 2004; Mandrekar, Jiang et al. 2009). Though

microglia have been the predominant cell type studied regarding this process, recent

12 studies have shown that astrocytes may play a larger role of Aβ removal from the brain than first thought (Funato, Yoshimura et al. 1998; Wyss-Coray, Loike et al. 2003;

Koistinaho, Lin et al. 2004). Whether Aβ is taken up by different cell types in the brain or it remains in the extracellular milieu, there are a number of proteases that have been reported to degrade it. The brain possesses intrinsic ability to degrade Aβ. Although there are many proteases that have been reported to degrade Aβ in vitro, neprilysin, insulin degrading enzyme (IDE), endothelin converting enzyme 1 (ECE1) and plasmin have been shown to have a role of degradation of Aβ in vivo (Shirotani, Tsubuki et al.

2001; Eckman, Watson et al. 2003; Farris, Mansourian et al. 2003; Liu, van Groen et al.

2011). The two principal degrading enzymes, however, are neprilysin and IDE.

Neprilysin is a cell membrane associated protease. It functions at the cell surface as well as in the membranes of intracellular compartments and mainly facilitates intracellular degradation of Aβ (Iwata, Mizukami et al. 2004). IDE is a neutral metalloprotease and is secreted into the extracellular milieu by microglia and neurons

(Qiu and Folstein 2006). IDE has been shown to mediate extracellular degradation of only soluble, but not oligomeric, species of Aβ (Qiu, Walsh et al. 1998). Inhibition or genetic inactivation of either of these genes results in elevated brain Aβ levels as well as increased plaque deposition (Farris, Mansourian et al. 2003; Miller, Eckman et al. 2003;

Dolev and Michaelson 2004; Iwata, Mizukami et al. 2004). Similarly, over-expression of

IDE or NEP results in the reduced accumulation of Aβ in the TgCRND8 mouse, harboring both the Swedish mutation and the Indiana mutation on APP (Leissring, Farris et al. 2003; Marr, Rockenstein et al. 2003; Iwata, Mizukami et al. 2004). Our lab has previously demonstrated that apolipoproteinE (apoE) and its lipidation status play an

13

integral role in both intracellular as well as extracellular degradation of soluble Aβ by

IDE and neprilysin (Jiang, Lee et al. 2008). (Figure 3)

Soluble versus Insoluble Aβ and Cognition

The question of whether amyloid beta plaques matter has become central to the

debate of disease pathogenesis. There is a very complex relationship between levels of

soluble Aβ and insoluble Aβ or plaque development and cognition. There is a disconnect

between cognitive function in animal models of the disease and the abundance of plaques

in the brains of murine models, which has provoked controversy. Murine data is

inconclusive. Though there are clear correlations of plaque load and behavioral deficits, there is also solid evidence of impaired memory, learning and synaptic deficits that

precede plaque pathology and is associated with soluble Aβ. In addition, there are a set

of studies utilizing LXR agonists, to be discussed in a later section, that have resulted in

behavioral improvements that were associated with significant (Riddell, Zhou et al. 2007;

Jiang, Lee et al. 2008; Fitz, Cronican et al. 2010), little (Donkin, Stukas et al. 2010) or no

(Vanmierlo, Rutten et al. 2011) change in plaque burden. More recently, behavioral

impairments in animal models has been linked to soluble levels Aβ (Cleary, Walsh et al.

2005; Cramer, Cirrito et al. 2012), preceding plaque deposition (Jiang, Lee et al. 2008;

Selkoe 2008; Zhang, Bai et al. 2012). This controversy extends to humans, as recent

reports of immunotherapy in humans have shown plaque removal but no improvement in

cognition (Kurz and Perneczky 2011). The advent of imaging techniques has shown

amyloid deposition is invariably found in AD patients, but surprisingly, it is also

observed in cognitively normal individuals. It is not known whether the plaque

deposition is an indicator of future conversion to AD (Sperling, Aisen et al. 2011).

14

Additionally, it is unclear if the amyloid plaques act as sinks, sequestering Aβ in a

tombstone like manner, or alternatively if they act as reservoirs of potentially toxic Aβ.

To determine whether insoluble Aβ plaques release smaller fragments of Aβ, Shankar

and colleagues isolated human fibrillar Aβ, and homogenizing the TBS-insoluble pellets

of human Aβ-rich Alzheimer’s disease cortex in 2% SDS. After immunoprecipitation

and Western blotting the supernatant, the authors found no additional Aβ fragments that

were liberated from the dense core plaques by SDS or TBS, suggesting there is no release

of Aβ from the amyloid plaque (Shankar, Li et al. 2008). Conversely, Cirrito and

collaborators showed that by rapidly inhibiting Aβ production by Compound E a

selective γ-secretase inhibitor, the interstitial fluid Aβ half life of young pre-depositing

PDAPP mouse models of AD was close to 2 hours, however the half life of ISF Aβ in older, plaque-containing mice was twice as long as in young mice as measured by in vivo microdialysis. The increase in half life of Aβ indicates that there is a pool of Aβ that is released to maintain equilibrium, thus elongating the half life of Aβ in the older mice

(Cirrito, May et al. 2003). This data was recapitulated by Hong and colleagues, using in vivo microdialysis. After supplying the rapid inhibitor of Aβ production, the γ-secretase inhibitor, Compound E, Aβ40, Aβ38 and Aβ42 levels of 3 month old pre-depositing J20 mice all fell rapidly, whereas after Compound E injection of 24 month old J20 mice,

Aβ38 and Aβ40 levels fell rapidly, but Aβ42 levels did not. The half lives of Aβ38 and

Aβ40 were similar, but the half life of Aβ42 was significantly longer in the old mice compared to the young mice. Aβ42 is more rapidly generated in the PDAPP mouse and is more fibrillogenic, lending itself to plaque formation. This suggests that there is a pool of less soluble Aβ42 that may be released from plaques, accounting for the lack of

15

decrease in Aβ42 half-life (Hong, Quintero-Monzon et al. 2011). Data from both studies

by Cirrito and Hong suggest that Aβ plaques act as a reservoir for toxic Aβ species.

Recent studies suggest there are separate, independent pools of Aβ. Recent studies

have provided compelling evidence that Aβ peptides are produced as a consequence of

endogenous neuronal activity (Cirrito, Yamada et al. 2005) and released into the interstitial space (Cirrito, May et al. 2003). Indeed, the level of synaptic activity is correlated with interstitial fluid (ISF) Aβ levels. There is general agreement that small oligomeric Aβ species interfere with synaptic transmission and this likely underlies the memory and cognitive deficits that typify the disease (Mucke, Masliah et al. 2000;

Shankar, Li et al. 2008; Fukumoto, Takahashi et al. 2010; Selkoe 2011). ISF Aβ levels are high in young animals and fall in parallel with the appearance of amyloid deposits.

Interestingly, soluble Aβ levels rise with age, and were very recently found to be largely in complexes >500kDa, and as expected, insoluble levels of Aβ rise with age (Hong,

Quintero-Monzon et al. 2011).

Microglia and the undiseased brain

Microglia are the principal immune effector cells in the brain. It is widely accepted that microglia originated from peripherally-derived mesodermal/mesenchymal progenitors. Recently, it was further determined that murine microglia are derived embryonically from yolk sac haematopoietic macrophages at E8.5 and populate the brain by E10.5 (Ginhoux, Greter et al. 2010; Saijo and Glass 2011). Microglia maintain

normal tissue homeostasis, they take up and dispose of cellular debris in the brain

parenchyma and elicit the first line of defense in the brain by mounting an inflammatory

response upon injury or pathogen.

16

Typically the turnover rate of resident microglia is low, through local self

renewal, where they comprise a stable population of cells. In the mature CNS, they

account for 5% of the total glial population in the cerebral cortex (Lawson, Perry et al.

1990; Lawson, Perry et al. 1992). Microglia typically have a small soma and are highly

ramified. It was thought that microglia in a normal healthy brain were “quiescent” and

“resting,” however, with the development of 2-photon microscopy, recent studies have

demonstrated the dynamic nature of microglia in immune surveillance of the brain

(Davalos, Grutzendler et al. 2005; Nimmerjahn, Kirchhoff et al. 2005) Using the

microglial-specific Cx3cr1 locus to drive enhanced green fluorescent protein (EGFP)

expression, 2-photon microscopy revealed that microglia are evenly spaced throughout

the brain at approximately 6 cells per cubic millimeter. The processes of microglia are

highly mobile, continuously extending and retracting, resulting in the monitoring of the

entire brain every few hours. During the surveillance process, it has been shown that

microglia make brief contact with astrocytes, neurons and vasculature (Davalos,

Grutzendler et al. 2005). The dynamic and finely tuned surveillance enables stationary

microglia to scan their environment without disrupting the neuronal network of the brain

parenchyma (Hanisch and Kettenmann 2007).

In the presence of a pathogen or injury, studies have shown that microglia recognize and readily migrate to the source of damage. Davalos et al. showed that microglia are capable of extending their processes within one minute of noxious stimuli.

Following the stimuli, laser-induced damage, surrounding microglial cells migrated to the site within 30 minutes and extended processes towards the stimuli. Studies concluded that

ATP release from astrocytes and P2Y12 receptors were necessary for the chemotaxis

17

(Davalos, Grutzendler et al. 2005; Haynes, Hollopeter et al. 2006). More recently, it was

shown that not only ATP released from astrocytes, but ATP-induced ATP release from the lysosomes of microglia are necessary for migration towards and process extension to the site of injury (Dou, Wu et al. 2012).

Microglia in the AD brain

Reactive microglia in the AD brain are intimately associated with dense-core Aβ plaques. They rapidly extend their processes and migrate towards the plaques (Bolmont,

Haiss et al. 2008; Meyer-Luehmann, Spires-Jones et al. 2008), and have been suggested to play a role in plaque dynamics. In addition to surrounding the dense core plaques, they have also been found to associate with more diffuse plaques (Itagaki, McGeer et al. 1989;

Akiyama and McGeer 1990; Mackenzie, Hao et al. 1995; Akiyama, Mori et al. 1999;

Stalder, Phinney et al. 1999). Proliferation of microgia in the vicinity of plaques contributes to their accumulation at the plaque periphery (Stalder, Phinney et al. 1999;

Bornemann, Wiederhold et al. 2001). Microglial number and size increases relative to the plaque size (Wegiel, Wang et al. 2001; Sasaki, Shoji et al. 2002; Wegiel, Imaki et al.

2003). The association of microglia with amyloid plaques in the brain is consistent in both transgenic mouse models as well as in human cases of AD (Itagaki, McGeer et al.

1989; Frautschy, Yang et al. 1998).

A more recent study by Meyer-Luehmann et al., using 2-photon microscopy,

identified newly formed plaques and the subsequent microglial chemotaxis to the plaque

after 1-2 days of plaque formation. Despite the recruitment of microglia to the plaque, the

clearance of the deposit was not observed. Additionally, the authors reported microglial

recruitment towards dystrophic neurites that surrounded the newly formed plaque, but did

18

not comment on the microglial contribution to the neuritic dystrophy (Meyer-Luehmann,

Spires-Jones et al. 2008). Using a similar technique, Bolmont et al. demonstrated that microglia migrate to plaques while maintaining a highly dynamic interface between plaques and glial cells. Importantly, Bolmont et al. were able to observe the internalization and delivery of Aβ fragments to the endocytic pathway of microglial cells, something that was not observed by Meyer-Leuhmann et al. (Bolmont, Haiss et al. 2008).

Interestingly, the elimination of microglia by treating animals with ganciclovir, from the brains of AD mice did not seem to change plaque pathology, despite the rise in soluble levels of Aβ, stimulating re-assessment of how microglia contribute to disease progression and calling into question whether endogenous plaque-associated microglia clear plaques at all (Grathwohl, Kalin et al. 2009) However, the interpretation of these experiments is confounded by the morbidity and mortality that accompanied the ganciclovir treatment, the loss of other cell types besides microglia, most notably pericytes and the integrity of the BBB.

There are conflicting studies that have either shown microglia to be competent phagocytes of fAβ or not. Recently, it has been proposed that peripherally-derived monocytes and macrophages can traffic into and populate the brain, and subsequently perform microglial-related functions (Simard and Rivest 2004; Malm, Koistinaho et al.

2005; Stalder, Ermini et al. 2005). However, in the initial studies, the experimental design of whole body irradiation was called into question as Ajami and Mildner suggested the monocytic infiltration was an artifact of the experimental procedure of whole body irradiation (Ajami, Bennett et al. 2007; Mildner, Schmidt et al. 2007;

Ransohoff 2007). El Khoury et al. provided more compelling evidence of infiltration.

19

The authors reported that the knocking out the chemokine, Ccl2/MCP-1 receptor, CCR2, in a mouse model of AD led to many fewer microglia surrounding plaques in the brain and increased Aβ levels. Myeloid lineage cells expressing CCR2 migrate in response to the chemokine, Ccl2 that is expressed at the site of injury, or in the case of AD, by plaque associated microglia and astrocytes. The authors proposed that this signaling system may play a role in stimulating peripheral monocytes or macrophage infiltration or perivascular macrophages into the brain. However, the loss of Ccr2 in the AD mouse model was associated with a similar number of microglia to that of nontransgenic mice.

The authors argue that the number of monocytes/microglia in the brains of the AD mouse

model was due to infiltration of peripherally derived cells, solely based on the abundance of cells expressing high levels of CD45, a defining characteristic of these cells, as endogenous microglia express low levels of CD45. CD45 is a protein tyrosine kinase that is thought to be a regulator of microglial activation (Tan, Town et al. 2000; Tan, Town et al. 2000; El Khoury, Toft et al. 2007). Conflicting data by Mildner et al., revealed no change in Aβ burden in CCR2-/-Tg2576 mice. These authors further showed no difference in the number of microglia or macrophages in the brains the CCR2-/-Tg2576 and the CCR2+/+ Tg2576 using the well-established, Iba1 marker (Mildner, Schlevogt et al. 2011). Town and colleagues have provided additional support for the notion that bone marrow-derived monocytes/macrophages translocate into the parenchyma of the brain of animal models of AD and facililate Aβ clearance. These authors expressed a dominant negative form of the TGFβ receptor under a CD11c promoter expressed in myeloid lineage cells. They demonstrated that inhibiting the TGFβ signaling pathway, using the dominant negative form of the receptor, allows for the infiltration of peripheral

20

monocytes/macrophages into the brain. Plaque pathology was dramatically less severe

and behavior was improved in aged APP-expressing Tg2576 mice when TGFβ signaling

was suppressed. Significantly, these mice exhibited increased numbers of plaque-

associated, CD45-hi expressing cells and other macrophage markers. The authors argue that the increase in CD45-hi cells reflects the infiltration of the brain by blood-derived macrophages. However, they did not detect significant expression of the dominant negative TGFβ receptor transgene in the brain, a potential confound in the interpretation of the data, given the magnitude of the increase in CD45-hi-expressing cells (Town,

Laouar et al. 2008). The interpretation of these studies and others is reliant upon the reliability of the markers employed. While the conclusions are reasonable if CD45-hi and

CD11c expression is restricted to infiltrating cells and is unaltered for significant periods of time once the cells are resident in the brain. Unfortunately, a subset of endogenous microglia has recently been reported to express CD11c, a canonical marker of dendritic cells (Bulloch, Miller et al. 2008). Additionally, there is a lack of specific markers to distinguish between CNS-endogenous microglia and peripheral monocytes. Though a number of current studies utilize CD45 expression to distinguish between peripheral monocytes versus endogenous microglia, there have been no definitive studies to show that CD45-hi microglia in the brain are derived from peripheral monocytes. It is still unknown whether in the normal, healthy brain, there is little need or impetus for peripheral macrophages and monocytes to infiltrate the brain, or they are physically inhibited from doing so by the BBB, but in a diseased state there may be entry signals or

BBB breakdown that permits their entry.

21

Inflammation:

In addition to the pathological hallmarks of the disease, amyloid deposits and

neurofibrillary tangles, AD is also characterized by a robust inflammatory response.

There is an extensive literature that documents increases in the levels of inflammatory

cytokines, chemokines, immune cell surface proteins, acute phase proteins, complement

proteins, and oxidative damage in the brains and CSF of AD patients. Oxidative stress is

manifested by protein oxidation, lipid peroxidation and 3-nitrotyrosine formation among

others in the AD brain. The levels of inflammatory molecules suggest a central role for

chronic inflammation in AD pathogenesis, supporting an inflammatory hypothesis of the

disease. The CNS possesses intrinsic mechanisms to suppress microglial activation and

thus cytokine release through cell-cell interactions between microglia and neurons as well as astrocytes and other glial cell types. One such example is the fractalkine receptor

(CX3CR1) that is expressed on the cell surface of microglia. The ligand for this receptor, fractalkine (CX3CL1) is expressed on neurons and astrocytes. While microglia survey the brain parenchyma, the interaction between ligand and receptor prevents microglial activation. It is postulated that during neurodegeneration, this ligand-receptor interaction is lost, due to loss of neurons and contributes to the proinflammatory cytokine release by microglia (Ransohoff 2007). Duan et al. showed that in the Tg2576 mouse model of AD, there were half the number of fractalkine positive cells in the regions of deposition as compared to the nontransgenic controls (Duan, Yang et al. 2008). Similarly, Lee et al., showed that in the CX3XR1 null APPPS1-21 or the R1.40 mouse, there was dramatic reductions in plaque pathology and corresponding decreases in markers of inflammation including CD68 and Tnfα expression (Lee, Han et al. 2010). Neuroinflammation reflects

22

the ongoing disease processes and accompanies Aβ deposition, it remains unclear

whether these responses are detrimental, beneficial or both in the pathogensis of

AD(Wyss-Coray 2006).

Microglia and Aβ binding:

Our lab has previously shown that microglia can bind to amyloid fibrils through a

complex of cell surface receptors which serve to activate tyrosine kinase-based intracellular signaling cascades (Bamberger, Harris et al. 2003; Reed-Geaghan, Savage et al. 2009) that promote the acquisition of a reactive, phagocytic, proinflammatory profile, producing proinflammatory cytokines, chemokines as well as reactive oxygen species

(Wilkinson, Koenigsknecht-Talboo et al. 2006; Reed-Geaghan, Savage et al. 2009). Our lab has recently reported that the expression of the Toll like receptors, TLRs 2 and 4 along with their coreceptor, CD14, are part of the complex on microglia to recognize fAβ as necessary to mediate and activate the intracellular signal cascade, leading to cytokines,

IL-6, TNFα and reactive nitrogen species (RNS) to be released (Fassbender, Walter et al.

2004; Reed-Geaghan, Savage et al. 2009; Reed-Geaghan, Reed et al. 2010). In addition to the secretion of cytokines, fAβ activation of microglia leads to the assembly of the microglia phagocytic NADPH oxidase (NOX2) complex resulting in superoxide production (Wilkinson, Cramer et al. 2012). It has been hypothesized that microglia are the major source of reactive oxygen and nitrogen species (ROS and RNS) in AD

(McDonald, Brunden et al. 1997; Bianca, Dusi et al. 1999). In addition to promoting an

inflammatory milieu, the microglial-mediated NOX2 induction has been shown to induce

Aβ production (Tamagno, Bardini et al. 2002), further suggesting that inflammation plays

a role in AD pathogenesis. Many in vitro studies have supported the idea that an

23

inflammatory environment negatively affects the capacity of microglia to engage in

phagocytosis and clear fAβ deposits. The TLRs along with their coreceptor, CD14 have

also been implicated in the uptake of Aβ (Liu, Walter et al. 2005). After cytokine treatment, it has been shown that microglia lose their capacity to phagocytose the Aβ peptide (Koenigsknecht-Talboo and Landreth 2005) , which results in the accumulation

of Aβ intracellularly (Yamamoto, Kiyota et al. 2008). Importantly, it was recently

reported that loss of CD14 in the APP/PS1 animal model of AD is associated with the

enhanced expression of proinflammatory genes endcoding the cytokines Tnfα and Ifnγ,

decreased levels of the microglial/macrophage alternative activation markers Fizz1 and

Ym1, and increased expression of the anti-inflammatory gene Il-10. Knocking out CD14

resulted in a significant change in the inflammatory environment of the brain. This likely

reflects a more heterogeneous population of microglia within the brain than initially

anticipated (Reed-Geaghan, Reed et al. 2010). The activation of microglia has been

expanded to a broad spectrum of phenotypes described in a later section.

NSAIDs

A subset of retrospective epidemiological studies have shown chronic, long-term

use of some non-steroidal anti-inflammatory drugs (NSAIDs) reduced the risk, delayed

the onset of AD and decreased the severity of the cognitive symptoms (in t' Veld,

Ruitenberg et al. 2001; McGeer and McGeer 2001; Szekely and Zandi 2010).

Therapeutic long term NSAID treatments have been shown to suppress inflammation, reduce amyloid deposition, alter amyloid precursor protein (APP) processing, and

improve cognitive performance in murine models of AD (Lim, Chu et al. 2001; Jantzen,

Connor et al. 2002; Yan, Zhang et al. 2003; Kotilinek, Westerman et al. 2008; McKee,

24

Carreras et al. 2008) . The best characterized action of NSAIDs is to suppress

inflammation, primarily through their ability to inhibit the COX enzyme COX-1/2,

leading to the reduced biosynthesis of pro-inflammatory prostaglandins (Botting 2010).

COX-1 is constitutively expressed in most tissues and mainly has ‘housekeeping’

functions. COX- 2 is elevated in inflammatory responses. The epidemiological and

murine studies led to clinical trials of NSAIDs in AD that failed to demonstrate any

benefit to patients (Aisen, Schafer et al. 2003; Group, Lyketsos et al. 2007; Arvanitakis,

Grodstein et al. 2008; Breitner, Haneuse et al. 2009). However, it should be noted that

53 of the patients that enrolled in the ADAPT Research Group (Lyketsos, Breitner et al.

2007) were enrolled erroneously, and when those patients, already diagnosed with MCI

or prodromal AD, were removed from the study the effect of celecoxib or naproxen were

no longer deleterious and revealed no significant change in the hazard ratio of developing

AD. Furthermore, reports by the authors of the ADAPT study have suggested that further

follow up of the study participants have shown a least a trend to a beneficial effect of the

treatments. Thus, there has been a renewed interest in this class of drugs for AD

treatment stemming from findings suggesting that some, but not all, NSAIDs can act

independently from their classic anti-inflammatory mechanisms, which may play a role in their disease-modifying actions (Lehmann, Lenhard et al. 1997; Combs, Bates et al. 2001;

Eriksen, Sagi et al. 2003), (Weggen, Eriksen et al. 2001; Zhou, Su et al. 2003; Lleo,

Berezovska et al. 2004). Similarly, further study is needed as it is difficult to make a rational risk/benefit comparison when the only published data with respect to NSAID use in AD is derived from a handful of patients who converted to AD in a single disbanded

25 study (Group, Lyketsos et al. 2007; Lyketsos, Breitner et al. 2007; Szekely and Zandi

2010) .

Activation Status of Microglia

Traditionally, microglia were classified into two primary phenotypes: a

‘quiescent’ or ‘activated’ state. The transition from the ‘quiescent’ state to the ‘activated’ state was associated with inflammation and disease. However, now it is clear that microglial activation status is determined by their immediate environment (Gordon

2003). Microglial cells display more phenotypic and functional heterogeneity in the CNS than was previously appreciated. Slowly, it is being recognized that microglia are highly plastic and play diverse roles in the brain. Several studies have elucidated not just two, but a continuum of different activation states. The appreciation of the phenotypic diversity of microglia has been attributed to studies of peripheral macrophage biology.

Gordon and colleagues have proposed a classification system to describes the ‘classical’ proinflammatory activation states as M1 and ‘alternative’ activation states as M2(Gordon

2003; Mantovani, Sica et al. 2004; Martinez, Helming et al. 2009). More recent attempts have been made to define the diverse macrophage phenotypes to reflect their plasticity and heterogeneity. It is now widely accepted and appreciated that there is a spectrum of activation states that prevents an easy identification, but that is influenced by the local environment of the cells and the capacity of the cells to initiate and resolve the tissue response to pathogens or injury (Mosser and Edwards 2008). However, it is not clear that the same classifications are appropriate for describing the endogenous brain microglial cells (Morgan, Gordon et al. 2005; Martinez, Helming et al. 2009).

26

M1 activation, otherwise known as classical activation, is characterized by the

upregulation of a variety of cell surface receptors, production of Th1-proinflammatory

cytokines, nitric oxide expression (Mantovani, Sica et al. 2004). In contrast, microglia

exposed to the anti-inflammatory cytokines interleukin 4 (IL-4), IL-10, IL-13, and

transforming growth factor β (TGF-β), reflective of a Th2 type response, become

alternatively activated (M2) and suppress the expression of Th-1 proinflammatory

cytokines. This alternate activation state demonstrates a greater capacity for phagocytosis

and does not produce nitric oxide. M2 microglia have been thought to play a role in

tissue repair and express gene such as YM1, YM2, FIZZ1, Mannose Receptor, and

Arginase 1 (Colton, Mott et al. 2006).

It is important to note, however, that M1 and M2 activation states only represent

only 2 different phenotypes in a continuum of macrophage activation which may also

encompass a combination of microglial phenotypes as well (Aisen, Schafer et al. 2003;

Mosser and Edwards 2008). The M2 alternative activation state has been further

subclassified into three distinct categories: M2a, M2b and M2c. M2a macrophages are

induced by IL-4 and IL-13 and exhibit an anti-inflammatory phenotype. M2b

macrophages are a true mixed phenotype as they highly express a subset of pro-

inflammatory cytokines, characteristic of M1 activation, but also express a high level of

the anti-inflammatory cytokine IL-10. The M2b phenotype is induced by the exposure to immune complexes, agonsits for IL-1R and TLRs (Mosser 2003). M2c macrophages are considered to be in an ‘acquired deactivation’ state, induced by IL-10, TGFβ, glucocorticoids or by contact with apoptotic cells and result in the suppression of the innate immune response. Cells in the M2c state play a role in the phagocytic removal of

27

cellular debris without the induction of the classical immune response, allowing for

normal tissue maintenance and repair (Mantovani, Sica et al. 2004).

Recent studies have shown that AD mouse models display a switch in microglial

activation status in response to disease progression. In two mouse models of the disease

(Tg2576 and the APP-SwDI) Colton and colleagues showed an increase in mRNA levels

of alternative activation makers, Arg1, mannose receptor and YM1 in comparison to

nontransgenic controls. Similarly, brain samples from AD patients have shown a similar

increase in M2 markers (Colton, Mott et al. 2006). However, a study by Jimenez et al.,

describes an age-dependent switch in microglial phenotype in the hippocampus of

APP/PS1 AD mice. Early in disease progression, microglia exhibit an M2 phenotype

early in amyloidosis that switches to an M1 phenotype in the older AD mice. The

authors found in young APP/PS1 mice, microglia surrounding Aβ plaques displayed

markers of M2 activation, YM1 and IL-4. At later ages, following more extensive plaque deposition, microglia exhibited the classical M1 phenotype, with an increased production of TNFα and a suppression of YM1 and IL-4 (Jimenez, Baglietto-Vargas et al. 2008).

Importantly, the authors showed that the soluble brain extracts containing small oligomeric Aβ from 18 month APP/PS1 mice stimulated the production of TNFα by glial

cells, which suggests that recognition of oligomeric Aβ by microglial cells is a key component to program these cells to an M1 phenotype. Both sets of data from Colton et

al. and Jimenez et al., suggest that microglial phenotype and activation status depends on

the microenvironment, such that as Aβ deposition and the proportion of microglia

surrounding plaques increases, the proportion of microglia in a given activation state

changes, influencing the global brain milieu and therefore subsequent Aβ deposition.

28

Consistently, using an adeno-associated viral vector to induce the expression of the M2 cytokine, IL-4, in the CNS of APP+PS1 mice, Kiyota et al. showed reduced amyloid pathology, decreased astro- and microgliosis and improvements in spatial learning in the radial arm water maze. These data indicate that polarizing the AD brain to an M2 state may exert beneficial effects on the pathogenesis of the disease (Kiyota, Okuyama et al.

2010). Additionally, it has recently been recognized that the interferon regulatory factor,

IRF5 determines the commitment of macrophages to the M1 phenotype, such that by knocking down IRF5, M1-specific cytokine production in peritoneal macrophages of

IRF5 -/- mice was prevented Similarly, IRF5 was shown to inhibit M2 activation markers, suggesting its role in the commitment of macrophages to an M1 phenotype

(Krausgruber, Blazek et al. 2011). Unpublished data in our lab suggests that microglia associated with Aβ plaques express IRF5, but that microglia distant from plaques are not

IRF5 positive, supporting the idea of a heterogenous population of microglia within the brain of AD mouse models. Further investigation is required to determine whether the

IRF5 positive microglia remain in an M1 status or are able to be polarized to an M2 phenotype. The idea that microglial phenotype is dependent upon age and pathology is important. The significance of this activation status change in ameliorating disease pathology requires further investigation.

A recent study by Choi et al., suggests that inhibiting the NADPH oxidase formation in microglia, drives microglial phenotype from a classical to an alternative activation state. After intracerebral injection of LPS into knockout mice of the regulatory subunit, p47 or the flavocytochrome b558 member, gp91, markers of alternative activation, including YM1, FIZZ1 and MARCO, were significantly upregulated as

29

compared to their wildtype counterparts. Concurrently, there was a decrease in the

expression of TNFα in the gp91-/- mice treated with LPS compared to wild type controls

(Choi, Aid et al. 2012). These data suggest that the formation of the NADPH oxidase plays a critical role in modulating the microglial phenotype towards a pro-inflammatory, classically activated state and that inhibition of NADPH oxidase represents a promising neuroprotective approach to reduce oxidative stress and modulate microglial phenotype towards an alternative state.

Recently, it has been appreciated that activation of a subclass of nuclear receptors

(to be discussed further in the next section), the peroxisome proliferator activated receptors (PPARs), controls the activation status of macrophages, promoting the acquisition of an M2 activation state. Activating either PPARγ or PPARδ in the periphery results in the expression of Arg1 and IL-4, M2 markers (Fuentes, Roszer et al. 2010).

Recent data from our lab has shown that an acute treatment of aged APP/PS1 mice with

PPARγ agonist, pioglitazone can induce the expression of M2 activation markers in the brains of the AD mice (Mandrekar-Colucci, submitted). Because microglia play such an integral role in AD pathophysiology, elucidating the role of microglial activation states and their subsequent effect of AD pathology is important and warrants further research, not only to understand the disease mechanisms, but also for developing new strategies for

AD therapy.

Nuclear Receptors

Nuclear receptors are ligand-activated transcription factors (Aranda and Pascual

2001; Castrillo and Tontonoz 2004). Nuclear receptors serve as master regulators of metabolism and have, thus, been targeted for drug development for a number of diseases.

30

Nuclear receptors induce gene expression though transcriptional transactivation. The

receptor exists in an inactive state associated with a co-repressor complex, consisting of

nuclear co-repressors (NCoR or SMRT) and associated histone deacetylases (HDAC).

Upon receptor ligation the co-repressor complex is replaced by co-activators, such as

CBP/p300, and is able to activate sequence-specific response elements on the promoters of target genes and regulate their transcription. (Glass and Rosenfeld 2000; Aranda and

Pascual 2001; McKenna and O'Malley 2002; Castrillo and Tontonoz 2004) (Figure 4).

The co-activator proteins act to drive gene transcription due to their intrinsic histone

acetyltransferase activity. Histone acetylation allows for chromatin to decondense,

facilitating targeted gene expression.

All members of the nuclear receptor family are structurally conserved and are

comprised of three main domains, the amino (N)-terminal activation domain, known as the activation function 1 domain (AF1), which is necessary for coactivator recruitment, the carboxy (C)-terminal ligand binding domain (LBD) and the DNA binding domain

(DBD). The DBD is highly conserved and mediates binding to specific response elements on the promoters of specific target genes (Glass and Saijo 2010).

In addition to transcriptional transactivation of their target genes, the nuclear

receptors LXR and PPARγ act to silence expression of a variety inflammatory genes by a

mechanism known as transcriptional transrepression. Transrepression of NFκB or

STAT1 inflammatory gene activation by PPARγ and LXRs is effected by the small

ubiquitin-like modifier known as SUMO. The SUMO protein family consists of SUMO1,

SUMO2 and SUMO3. SUMO 2 and SUMO3 share about 95% sequence homology.

SUMO conjugation (SUMOylation) is achieved through a process that resembles protein

31

ubiquitination, and it is thought to regulate inflammation. SUMO can be reversibly

covalently added to lysines. The SUMO consensus sequence is generally €-K-x- E/D, where € is a hydrophobic amino acid, x is any amino acid, D/E is aspartic/glutamic acid, and K (lysine) is the specific SUMOylation target. Sumoylation is mediated by an enzymatic cascade that involves three classes of enzymes. The activation step of SUMO is mediated by the heterodimeric E1 activating enzyme SAE1/2, the conjugating step of

SUMO is mediated by a single E2-conjucating enzyme, Ubc9 and the ligation step is

mediated by a variety of E3 ligases, including PIAS and HDAC4 (Treuter and Venteclef

2011). It has been previously shown that PIAS1 (protein inhibitor of activated STAT1)

promotes the conjugation of SUMO1 to PPARγ. The SUMOylated PPARγ then binds the

nuclear corepressor complex of NFkB to prevent transcription of the inflammatory genes

(Pascual, Fong et al. 2005). LXRs have been shown to be SUMOylated with SUMO2/3

by HDAC4. Importantly, it appears that LXRα and LXRβ are SUMOylated by different

ligases. LXRα utilizes PIAS1 to prevent STAT1 from actively binding to DNA to initiate

transcription while LXRβ is conjugated by HDAC4 with SUMO2/3 to prevent STAT1

inflammatory gene transcription by maintaining the nuclear corepressor complex (Figure

5) (Ghisletti, Huang et al. 2007).

Type II nuclear receptors form obligate heterodimers with the retinoid X receptor

(RXRs) to form a functional transcription factor. While the roles of these receptors have

been studied in depth in the periphery, little is known about their function the brain.

The LXRs and PPARs act coordinately to regulate lipid metabolism, allowing the

organism to rapidly adjust whole body metabolism to dietary intake. RXRs are

considered “promiscuous” due to their ability to form heterodimers with a variety of

32

other type II nuclear receptors. RXR binding partners have further been classified as

either permissive or nonpermissive. Both LXR and PPARs are permissive binding

partners to RXR, such that the activation of the receptor complex is achieved by ligation

of either RXR or the other binding partner (Figure 5). Heterodimers formed between

RXR and nonpermissive binding partners, such as thyroid hormone receptor (TR),

retinoic acid receptor (RAR) or the vitamin D receptor(VDR), can only be activated by

the ligands specific to the nonpermissive binding partner (Desvergne, Michalik et al.

2006; Szeles, Poliska et al. 2010).

Retionid X Receptors (RXRs)

RXRs are the “master regulators” of the type II nuclear receptor family. RXRs are

involved in a diverse biological processes including cell proliferation, differentiation, and

in glucose, fatty acid and cholesterol metabolism(Ahuja, Szanto et al. 2003). There are

three isoforms of RXRs; RXRα, RXRβ, RXRγ encoded on three separate chromosomes,

9, 6 and 1 in the human and 2, 17 and 1 in the mouse, respectively. RXRα and β are most

prevalent in the neocortex and hippocampus (Moreno, Farioli-Vecchioli et al. 2004).

RXRs form transcriptionally active heterodimers with multiple nuclear receptor family

members, in the brain, most prominently PPARs and LXRs (Hegele 2005). RXRs can

function as homodimers, where they have previously been shown to bind to and activate

PPAR response element genes independent of PPAR ligands (Lehmann, Lenhard et al.

1997), although this is a quantitatively minor role for these receptors.

Initial genetic analysis of RXR function pointed to a key role in embryonic

development, which is believed to be mediated primarily by RXR–RAR dimers (Dolle

2009) and RAR activation by all-trans retinoic acid (ATRA) (Mascrez, Ghyselinck et al.

33

2009). Whereas the biological relevance of ATRA is no longer disputed as a ligand for

RAR, the 9-cis retinoic acid isomer, initially identified as a bona fide ligand for RXR in

vitro had not been identified as the endogenous in vivo ligand until very recently, when it

was found in the pancreas (Kane 2012). Retinoic acids are metabolites of vitamin A

(retinol). Retinol comes from the diet as retinyl esters, mostly from animal products such

as liver, eggs, and milk, or as carotenoid precursors in plant products, particularly green

leafy vegetables. All trans retinoic acid and 9-cis retinoic acid among other retinoic acids are considered teratogens. This teratogenic effect is caused by the interference of the

exogenous retinoic acid with endogenous retinoic acid signaling, which plays a role in

patterning the developing embryo. All trans- and 9-cis retinoic acid have been employed

for a variety of uses, including dermatological disorders and leukemia. RXR selective

agonists are known as rexinoids and similarly are considered teratogens.

Peroxisome Proliferator Activated Receptors (PPARs)

PPARs play essential roles in energy metabolism, adipocyte differentiation and

insulin sensitization and tumor suppression. Their endogenous ligands are dietary lipids

and their metabolites (Forman, Chen et al. 1997). PPARs are the dominant regulator of fatty acid metabolism by their ability to transactivate genes endcoding enzymes of lipid metabolism, providing a key linkage between diet and the genome. There are three isoforms, PPARα, PPARβ/δ and PPARγ. PPARα is an activator of mitochondrial and

peroxisomal fatty acid β- oxidation in liver and is not highly expressed in the CNS.

PPARβ/δ is the most abundant isoform and plays a role in fatty acid oxidation in muscle.

PPARγ is involved in lipid storage, insulin sensitivity and energy metabolism that has

been shown to promote adipocyte differentiation (Varga, Czimmerer et al. 2011). The

34

natural ligands of PPARs include long chain fatty acids, eicosanoids, oxidized

lipoproteins and lipids(Lehrke and Lazar 2005), which corresponds well to its function in

regulating the metabolic response to dietary lipid intake. Therefore, agonists of PPARγ,

specifically, have been developed for treatment of type II diabetes. Two

thiazolidinediones (TZDs) PPARγ agonists have been approved by the FDA

(pioglitazone, ActosTM ; rosiglitazone AvandiaTM) and are widely prescribed for this indication PPARγ has been most widely studied in the CNS and has been shown to exert anti-inflammatory actions (Chawla, Boisvert et al. 2001; Landreth, Jiang et al. 2008).

Liver X Receptors (LXRs)

The LXRs are ligand-activated nuclear receptors whose principal function is to regulate whole body cholesterol homeostasis (Fitzgerald, Moore et al. 2002; Joseph,

Castrillo et al. 2003; Tontonoz and Mangelsdorf 2003). LXRs are activated by oxysterols, hydroxylated forms of cholesterol, and have been shown to exert potent anti- inflammatory actions. Of the two isoforms of LXRs, LXRα is found predominantly in cells specialized in cholesterol metabolism (Aranda and Pascual 2001; Joseph, Castrillo et al. 2003; Kalaany and Mangelsdorf 2006), whereas LXRβ is ubiquitously expressed.

Both isoforms are expressed in the brain. Due to their ability to regulate cholesterol metabolism, LXR agonists have emerged as possible therapeutic targets for atherosclerosis. Unfortunately, their therapeutic utility is compromised due to their induction of hepatic steatosis. LXRs have also been shown to play an important role in the CNS. LXRβ knockout mice display adult onset motor degeneration, accompanied by axonal dystrophy, astrogliosis and lipid accumulation (Andersson, Gustafsson et al.

2005). Animals lacking both LXRα and LXRβ have a variety of brain abnormalities,

35 including the accumulation of lipid droplets, loss of neurons, astrocytic proliferation, gross brain defects and vasculature defects (Wang, Schuster et al. 2002)

ApolipoproteinE (ApoE) and Alzheimer’s Disease

The ApoE gene is the principal genetic risk factor for sporadic AD. The principal isoforms in humans are apoε2, apoε3, and apoε4. One copy of the apoε4 allele confers a

3-4 fold increased risk and two copies, one from each parent, confers a 12 fold risk for developing AD. Apoε2 alleles are thought to be protective against developing AD. The difference among the three isoforms is two amino acids at positions 112 and 158. Apoε2 has cysteines at both locations, apoε3 has cysteine and arginine, respectively and apoε4 has both arginines.

ApoE is a main constituent of high-density lipoproteins (HDLs), which is the obligatory cholesterol carrier in the CNS. It acts as a cholesterol acceptor to mediate cholesterol efflux. In the brain, ApoE is primarily synthesized and secreted by astrocytes although it is also expressed at lower levels by microglia. The lipidation of apoE is primarily carried out by the ATP binding cassette protein A1 (ABCA1). ABCA1 transfers phospholipid and cholesterol from plasma membranes to apoE. This process is known as the reverse cholesterol transport (RCT) pathway. Consistent with its role in apoE lipidation and metabolism, the level of ABCA1 is correlated with HDL size and levels of apoE in the murine brain (Hirsch-Reinshagen, Zhou et al. 2004; Riddell, Zhou et al. 2008). Importantly, ApoE avidly binds to Aβ and has been found to codeposit with

Aβ in AD brains. The Aβ-apoE interaction is isoform dependent and can be further influenced by its lipidation status (LaDu, Pederson et al. 1995; Jiang, Lee et al. 2008).

36

Our recent work demonstrated that microglia deficient in Abca1, which have poorly lipidated apoE-HDLs, were unable to efficiently degrade Aβ (Jiang, Lee et al. 2008).

Three independent studies have demonstrated that Abca1 deficiency results in increased Aβ levels, amyloid deposition and cerebral amyloid angiopathy (CAA) in 4 different mouse models of AD (Wahrle, Jiang et al. 2004; Hirsch-Reinshagen, Maia et al.

2005; Koldamova, Staufenbiel et al. 2005). While decreased levels of Aβ and plaque burden correlated with overexpressing Abca1 in the PDAPP mice (Wahrle, Jiang et al.

2008).

Murine apoE shares 70% homology in amino acid sequence with human apoE. It has been demonstrated that there are functional differences between murine apoE and human apoE isoforms on cholesterol distribution in synaptic plasma membranes and apoE secretion in vivo (Holtzman, Bales et al. 1999) . Therefore, it has been suspected that human apoE isoforms have different effect on Aβ compared to murine apoE. The isoform-specific actions of apoE on Aβ deposition have been studied in APP mice expressing the human apoE isoforms. Mice expressing human apoε4 exhibit higher levels of Aβ deposition and cytokine production in comparison to Apoε3- or Apoε2- expressing mice (Fagan, Watson et al. 2002; Castellano, Kim et al. 2011; Zhu, Nwabuisi-

Heath et al. 2012). Interestingly, APP mice lacking murine apoe fail to develop compact plaques suggesting a role for apoE in the deposition of Aβ in the brain. These mice lacking apoe were still able to form diffuse plaques and had equivalent or even higher levels of Aβ within the brains, arguing a role of apoE in Aβ clearance (DeMattos 2004).

37

Numerous studies have shown that both PPARγ and LXRα induce the expression

of apoE and ABCA1. We postulate that it is through the expression of these proteins that

the nuclear receptors exert their effects on Aβ pathology.

Nuclear Receptors and Alzheimer’s Disease mouse models

In the past decade, drugs targeting certain nuclear receptors, mainly Liver X

Receptor (LXR) and Peroxisome Proliferator-Activated Receptor gamma (PPARγ) have

been shown to ameliorate AD pathology in animal models of the disease. Genome-wide

association studies have linked dyslipidemia and high cholesterol levels to AD risk.

Genes associated with cholesterol regulation, including apoE, LRP1, ABCA1, LXRβ,

among others, have also been shown to share a linkage with AD. However, the exact

mechanism through which ApoE, especially Apoε4, confers such a risk for developing

AD is just recently beginning to be understood. Our lab was the first to show that

lipidated apoE acts to promote the proteolytic degradation of Aβ, providing the

mechanistic linkage between the major genetic risk factor for AD and the normal

clearance of Aβ from the brain. Additionally, it was shown that the lipidation of apoE

enhanced the degradation of soluble species of Aβ both extracellularly by insulin

degrading enzyme (IDE) and intracellularly, in the endolytic pathway, by neprilysin.

Interestingly, in vitro experiments showed that loss of either abca1 or apoE prevented the

degradation of soluble Aβ, suggesting their utility in the degradation process (Jiang, Lee

et al. 2008). Most importantly, this study utilized an LXR agonist, GW3965, as apoE,

ABC transporters, (ABCA1 and ABCG1), and HDL modifying enzymes (CETP and

PLTP) are direct targets of LXRs.

38

Liver X Receptors

LXR activation has been shown to have a positive effect on Aβ clearance and

behavioral improvement. A study by Riddell et al. has shown that treatment of 3 month

old Tg2576 animals with the LXR agonist T0901317 improved disease-related behavioral impairments as well as reductions in brain Aβ40 levels, a finding consistent with a report

by Koldamova et al (Koldamova, Staufenbiel et al. 2005). We have shown a reduction in

plaque burden in aged Tg2576 animals after four months of LXR agonist treatment and

memory improvements in young Tg2576 animals after an accute treatment with the LXR

agonist, GW3965 (Jiang, Lee et al. 2008). Vanmierlo et al. showed that despite there

being no change in Aβ load in a long term LXR agonist treatment using T0901317 on

aged APP/PS1 mice, the mice had improved memory performance and cholesterol

turnover (Vanmierlo, Rutten et al. 2011). In addition, LXR activation has also been

shown to exhibit potent anti-inflammatory actions (Joseph, Castrillo et al. 2003; Valledor

2005; Zelcer and Tontonoz 2006). Consistent with its involvement in AD pathogensis,

Zelcer et al., showed that the genetic inactivation of either LXRα or LXRβ decreased protein levels of both apoE and ABCA1, resulted in an exacerbation of Aβ pathology in the APP/PS1 mice (Zelcer, Khanlou et al. 2007).

Recent studies have shown the vital role for ABCA1 in LXR-mediated Aβ

clearance. Donkin et al. showed that in APP/PS1 mice that either expressed or lacked

abca1, when treated with the LXR agonist, GW3965, at either a high or low dose of the

drug upregulated brain levels of ABCA1 and apoE, had decreased plaque burden and had

improvements in memory behaviors. APP/PS1 mice lacking abca1 displayed little

change in brain levels of apoE, no change in plaque burden or behavior (Donkin, Stukas

39

et al. 2010). In other studies, Koldamova et al., Hirsch-Reinshagen et al., and Wahrle et

al., have shown that ABCA1-deficient mice crossed with various mouse models of AD

have increased or trending towards increased levels of Aβ in the absence of ABCA1,

suggesting that poorly lipidated apoE promotes amyloidogenesis (Hirsch-Reinshagen,

Maia et al. 2005; Koldamova, Staufenbiel et al. 2005; Wahrle, Jiang et al. 2005) All of

the studies suggest that apoE and ABCA1 are necessary for the beneficial effects of LXR

agonsits and also suggests that LXRs are excellent therapeutic targets for AD.

Peroxisome Proliferator Activated Receptor

PPARγ has been shown to act in concert with the LXRs to induce the expression of genes involved in reverse cholesterol transport in the periphery, including ApoE,

ABCA1 and ABCG1 (Fitzgerald, Moore et al. 2002; Joseph, Castrillo et al. 2003;

Tontonoz and Mangelsdorf 2003). LXRα induction reciprocally activates PPARγ expression, driving a positive feedback loop (Seo, Moon et al. 2004).

Aside from its primary function as a master regulator of lipid homeostasis (Berger and Wagner 2002; Picard and Auwerx 2002; Castrillo and Tontonoz 2004), recent work from our own laboratory and others have implicated PPARγ activity in ameliorating AD related pathology and cognitive impairment (Jiang, Lee et al. 2008). We have tested the effects of oral administration (20mg/kg/day) of pioglitazone for 4 months in one year old

Tg2576 mice and observed modest effects on total Aβ levels due to poor BBB permeability (Yan, Zhang et al. 2003). In a subsequent study, 10 month old V717IhAPP mice were treated orally with double the dosage (40mg/kg/day) with pioglitazone for only 7 days. This short treatment resulted in reductions in sAβ42 levels, inflammatory markers and plaque burden (Heneka, Sastre et al. 2005) and was considered the first

40

study that provided conclusive evidence for the utility of PPARγ agonists in animal

models of AD. PPARγ activation has also been shown to improve disease-related

cognitive impairments in AD mouse models. Pederson examined the effects of another

agonist to PPARγ, rosaglitazone, and found that activation of PPARγ ameliorated

behavioral deficits in the Tg2576 AD mouse model, reduced brain Aβ42 levels, but did

not change the plaque pathology, owing to rosaglitazone’s poor BBB penetrability

(Pedersen, McMillan et al. 2006) . In a more recent study, the long term effects of a low

dose of rosaglitazone were analyzed. APP/PS1 animals were treated with 3 mg/kg/day

for 12 weeks. These mice showed a 50% decrease in plaque area, a decrease in Aβ

oligomers and behavioral improvements as assessed by the Morris water maze (Toledo

and Inestrosa 2010). In a study by Escribano et al., 9 month J20 AD mice were treated

with 5 mg/kg/day rosiglitazone for 4 months. The animals displayed behavioral

improvements and after 4 months showed a 50% decrease in Aβ40 and Aβ42 levels,

consistent with the previous report. The authors did observe a modest increase in

ABCA1 and argued that the enhanced Aβ clearance could be attributed to the lipidation

of apoE by ABCA1 (Escribano, Simon et al. 2010). Despite very promising results in

mouse studies and promising initial results in humans (Watson, Cholerton et al. 2005), rosiglitazone treatment failed to show any cognitive improvements in AD patients in

Phase III clinical trials (Gold, Alderton et al. 2010; Harrington, Sawchak et al. 2011).

One explanation is the low penetrance across the BBB and efficient efflux from the brain

(Festuccia, Oztezcan et al. 2008) that compromised its clinical utility (Pedersen and

Flynn 2004; Pedersen, McMillan et al. 2006; Strum, Shehee et al. 2007).

41

Most recent work from our lab has provided the first evidence for a metabolic link

between PPAR and LXR pathways in the brain. An acute treatment of pioglitazone, at 80

mg/kg/day of APP/PS1 mice resulted in the rapid reduction of Aβ peptides and plaque

levels in the brain and was associated with improvements in contextual fear conditioning

in 12 month old mice. In addition, Mandrekar-Colucci et al., showed that PPARγ

activation not only enhances the expression of apoE, ABCA1 and the lipidation of apoE,

but also changes the inflammatory milieu of the brain converting the brain from the

“classical” inflammatory (M1) state, that is associated with AD, to an alternative (M2)

phenotype, associated with robust phagocytosis and tissue repair state. These data

provide support for the therapeutic use of PPARγ agonists for the treatment of AD

(Mandrekar-Colucci et al. submitted). Additionally, there has been an pilot clinical trial

using pioglitazone for AD, which further solidifies it’s potential as a therapeutic for the

disease (Geldmacher, Fritsch et al. 2011).

Retinoid X Receptor

As LXR and PPARγ both heterodimerize with RXR, logic would dictate that

RXR agonist treatment would also enhance the clearance of Aβ from the brains of AD mouse models through the reverse cholesterol transport pathway. However, very little is known about the action of RXR agonists in the CNS. Rexinoids are RXR selective agonists. RXR activation can be achieved by 9-cis RA, though at high enough concentrations, it will activate RAR. 9-cis RA induces the expression of several LXR target genes, including apoE, in a human astrocytoma cell line (Chen, Costa et al. 2011).

More importantly, combined treatment with the LXR agonist, TO901317 synergistically

increases the expression of apoE (Liang, Lin et al. 2004). 9-cis RA has also been

42

reported to induce the expression of ABCA1 in primary neurons, astrocytes, microglia

(Koldamova, Lefterov et al. 2003) and macrophages and is able to increase apolipoprotein lipidation status (Nishimaki-Mogami, Tamehiro et al. 2008).

Another rexinoid, honokiol, has recently been shown to induce both the mRNA

and protein expression of ABCA1 in RAW264.7 cells and in human glioma cell line and

induce cholesterol efflux in peripheral macrophages (Jung, Horike et al. 2010; Kotani,

Tanabe et al. 2010). Similarly, honokiol significantly decreased cell death of PC12 cells in the presence of Aβ (Hoi, Ho et al. 2010). More recently, honokiol has been shown to

act in a synergistic manner with the LXR agonist, T0901317 or with rexinoid,

bexarotene, to induce the expression of ABCA1, ABCG1 and apoE (Kotani, Tanabe et al.

2012). Honokiol is a naturally occurring RXR agonist found in Magnolia bark.

Regardless of their numerous beneficial effects, rexinoids often raise triglyceride and

cholesterol levels and dyregulate the thyroid hormone axis. For that reason, only one

known rexinoid has been FDA approved, bexarotene.

Bexarotene is a highly selective synthetic RXR agonist, known as LG10269,

(Bexarotene, TargretinTM), and is FDA approved for the treatment of cutaneous T-cell

lymphoma. It is being investigated for in the treatment of psoriasis and breast cancer

(Yan, Zhang et al. 2003; Farol and Hymes 2004). Bexarotene has been shown to induce

the expression of ABCA1 and ABCG1, both proteins are involved in apoE lipidation

(Lalloyer, Fievet et al. 2006), indicating that the RXR agonist is sufficient to drive the

activity of LXRs (Lalloyer, Fievet et al. 2006). Though there have been many studies on bexarotene’s effects in the periphery regarding its effect on triglyercides, inflammation

43 and cancer (Mukherjee, Strasser et al. 1998; Lalloyer, Fievet et al. 2006; Nunez, Alameda et al. 2010; Zhang, Pan et al. 2011), very little is known about its action in the brain.

Retinoic Acid Receptor

Recent studies have shown that activating one of the nonpermissive binding partners to RXR, RAR using all trans retinoic acid (ATRA) induces the expression of

ADAM10 (α-secretase), preventing the production of Aβ both in vitro assays in neuronal primary cultures and in vivo in the Tg2576 AD mouse model (Jarvis,

Goncalves et al. 2010). Though the authors did not look at RCT genes, apoE or ABCA1,

Ding et al. reported that a systemic 8 week treatment of 20 mg/kg every three days of

ATRA effectively reduced Aβ accumulation and tau hyperphosphorylation, decreased glial activation and rescued the spatial learning and memory deficits in 5 month old

APP/PS1 mice (Ding, Qiao et al. 2008).

Retinoid signaling and amyloid formation has recently been tested. Ono et al., demonstrated that Vitamin A has anti-amyloidogenic and de-stabilizing effects on amyloid fibril formation in vitro (Ono, Yoshiike et al. 2004). This data was corroborated by Corcoran et al, who showed that one year old rats deficient of Vitamin A since weaning age, develop amyloid beta plaques in their cerebral cortcies (Corcoran, So et al.

2004). In addition, vitamin A deficient rats exhibit inflammation, indicating a role in regulation of inflammatory cytokines by retinoid signaling (Reifen 2002). Similarly, these animals display deficits in lipid efflux, a key step in the reverse cholesterol pathway, further implicating retinoid signaling in Alzheimer’s disease (Yang, Chen et al.

2010).

44

Most recently, evidence suggests that the RXR/RAR signaling pathway is

involved with long term potentiation (LTP) and hippocampal memory. Developing a

dominant negative RAR mouse, Nomoto et al., showed mice have deficits in recognition

by social interaction assessment, impaired spatial memory as assessed by the Morris

Water maze, and have also shown impaired maintenance of CA1-LTP. These data suggests that learning and memory are directly linked through signaling by RXR heterodimer partners (Nomoto, Takeda et al. 2012)

Current Treatments for AD:

The earliest hypothesis as to what led to AD was that there was reduced production of the neurotransmitter acetylcholine. Early AD therapies were designed to preserve acetylcholine by acetylecholinesterase inhibitors (Giacobini 1994; Stahl 2000;

Birks 2006) . Five drugs for treatment of AD were approved by the FDA—four were cholinergic inhibitors, donepezil (Aricept), galantamine (Razadyne), rivastigmine

(Exelon patch) and tacrine (Cognex) and one neuropeptide modifying agent, memantine

(Namenda) (Trinh, Hoblyn et al. 2003; Qaseem, Snow et al. 2008). These therapies, however, only relieve symptoms of the disease for only a portion of AD patients, for at most 6 months, after which the neurons releasing the acetylcholine die and the acetylcholinesterase inhibitors have little to inhibit. Unfortunately, to date, all clinical trials of supposed ‘disease-modifying’ compounds in individuals with AD have failed to show clinical benefit—memory improvements or delay or slowing of memory impairments. There are several reasons for the failures of these trials including the inability to detect the drug benefit and the key consideration, that therapeutic intervention may be most successful prior to the onset of dementia, in the pre-symptomatic stage of

45

AD, before substantial neuronal loss has occurred. Developing ‘disease-modifying’

therapies that combat the underlying disease pathogenesis are of upmost importance to quell the upcoming socioeconomic burden that Alzeimer’s disease will cause the United

States and the world.

Focus of the Thesis The focus of this thesis is to analyze the pathways of two different ‘disease modifying’ compounds and how they exert their beneficial effects on amyloid pathogenesis in different animals models of AD.

In Chapter 2, we demonstrate the role of ibuprofen in decreasing amyloid deposition and preventing oxidative damage in the R1.40 mouse model of AD. We further show that ibuprofen of microglia or monocytes with racemic or S-ibuprofen inhibits the Aβ-stimulated Vav tyrosine phosphorylation, NADPH oxidase assembly and superoxide production. Interestingly, we show that Aβ-stimulated Vav phosphorylation

was not inhibited by COX inhibitors, suggesting a non-classical pathway of ibuprofen

action.

In chapter 3, we demonstrate the mechanisms of action of the FDA approved

drug, Bexarotene in the brains of AD mouse models. Using the FDA approved dose, we

show that the 1 dose of the drug can lower ISF Aβ levels just 6 hours after administration

and can suppress levels for up to 84 hours. Additionally, bexarotene is capable of

reversing behavioral impairments in three different mouse models of AD with the

concurrent upregulation in apoE, ABCA1 and the lipidated forms of apoE. These data

suggests bexarotene could be a potential therapeutic, by means of the reverse cholesterol

transport and apoE lipidation, for the treatment of AD and its antecedent phases.

46

47

Figure 1-1: APP Processing

48

49

Figure 1-2: Aβ aggregation and deposition

50

51

Figure 1-3: Intracellular and Extracellular Clearance of sAβ

The brain possesses efficient intrinsic Aβ clearance mechanisms. Aβ peptides are produced and cleared in the brain at approximately the same rate. The proteolytic degradation of Aβ can occur both intracellularly and extracellularly by Neprilysin (NEP) and Insulin Degrading Enzyme (IDE), respectively. Additionally, it has been previously shown that increasing the lipidation of ApoE can enhance the degradation both intracellularly and extracellularly.

52

53

Figure 1-4: Nuclear Receptor Activation and Repression

A: In the absence of agonist, type II nuclear receptors heterodimerize with RXR, bind to response elements located on the promoters of target genes in association with a corepressor complex, preventing the expression of target genes. In the presence of ligand, the receptor heterodimer undergoes a conformational change, exchanging the corepressor complex for a coactivator complex, promoting the transcription of the target genes. B: Upon ligand binding, the nuclear receptor become sumoylated and recruited to the NFkB-corepressor complex. This interaction prevents the removal of the corepressor complex from the promoter of the proinflammatory genes and maintains these genes in a repressed state

54

A

B

55

Figure 1-5:LXR and PPAR are permissive binding partners with RXR

Activation of the nuclear receptors LXR and PPARγ induce the expression of genes necessary for lipid metabolism and insulin sensitization. Interestingly, the activation of

PPARγ induces the expression of LXRs. Similarly, LXRα activation induces the expression of PPARγ, perpetutating a positive feedback loop of activation. In addition, activation of the heterodimers can be achieved by either ligands for LXR and PPARγ, promoting the expression of the individual heterodimer pair target genes or by ligands to

RXR, promoting the expression of both PPARγ and LXR heterodimer pair target genes, simultaneously.

56

57

Chapter 2 :Ibuprofen attenuates oxidative damage through NOX2 inhibition in

Alzheimer’s Disease

Brandy L. Wilkinsona,1, Paige E. Cramera,1, Nicholas H. Varvela,b, Erin Reed-Geaghana,

Qingguang Jianga, Alison Szaboa, Karl Herrupc, Bruce T. Lambb,d, Gary E. Landretha*

a Alzheimer Research Laboratory, Department of Neurosciences, Case Western Reserve University School of Medicine, Cleveland, OH, USA b Department of Neurosciences, Lerner Research Institute, The Cleveland Clinic Foundation, Cleveland, OH, USA c Department of Cell Biology and Neuroscience, Rutgers, The State University of New Jersey, Piscataway, NJ, USA d Department of Genetics, Case Western Reserve University, Cleveland, OH, USA 1 These authors contributed equally to this manuscript.

Received 6 January 2010; received in revised form 10 June 2010; accepted 21 June 2010; published 2012

Acknowledgements: This work was supported by grants from the National Institutes of

Health (AG16740, G.E.L.; AG023012, B.T.L.; AG024494, K.H.), the Blanchette Hooker

Rockefeller Foundation, and the American Health Assistance Foundation (G.E.L.).

B.L.W. and E.R.G. were supported in part through Ruth L. Kirschstein National Research

Service Awards from the National Institutes of Health (F32 AG24031 and F31-

NS057867). Thanks to Natalie Cherovsky for technical help and Colleen Karlo for review of the manuscript.

*Reprinted with authorship rights from Neurobiology of Aging 58

Abstract

Considerable evidence points to important roles for inflammation in Alzheimer’s disease

(AD) pathophysiology. Epidemiological studies have suggested that long-term nonsteroidal anti-inflammatory drug (NSAID) therapy reduces the risk for Alzheimer’s disease; however, the mechanism remains unknown. We report that a 9-month treatment of aged R1.40 mice resulted in 90% decrease in plaque burden and a similar reduction in microglial activation. Ibuprofen treatment reduced levels of lipid peroxidation, tyrosine nitration, and protein oxidation, demonstrating a dramatic effect on oxidative damage in vivo. Fibrillar β-amyloid (Aβ) stimulation has previously been demonstrated to induce the assembly and activation of the microglial nicotinamide adenine dinucleotide phosphate (NADPH) oxidase leading to superoxide production through a tyrosine kinase-

based signaling cascade. Ibuprofen treatment of microglia or monocytes with racemic or

S-ibuprofen inhibited Aβ-stimulated Vav tyrosine phosphorylation, NADPH oxidase assembly, and superoxide production. Interestingly, Aβ-stimulated

Vav phosphorylation was not inhibited by COX inhibitors. These findings suggest that ibuprofen acts independently of cyclooxygenase COX inhibition to disrupt signaling cascades leading to microglial NADPH oxidase (NOX2) activation, preventing oxidative damage and enhancing plaque clearance in the brain.

59

Introduction

Alzheimer’s disease (AD) is characterized by the formation of focal, compact β-

amyloid (Aβ) deposits within the brain. These deposits are surrounded by phenotypically

activated microglia, which are responsible for a locally-induced chronic inflammatory

response and affect Aβ homeostasis. It has been proposed that inflammation plays an

important role in AD pathogenesis as the AD brain exhibits elevated levels of

inflammatory molecules or other immune mediators (Akiyama, Barger et al. 2000;

Bamberger and Landreth 2002). Chronically activated microglia also generate reactive

oxygen (ROS) and nitrogen species. Several markers of oxidative damage including lipid

peroxidation (Mark, Lovell et al. 1997; Sayre, Zelasko et al. 1997), nucleic acid oxidation

(Nunomura, Perry et al. 1999), and protein oxidation (Smith, Richey Harris et al. 1997)

are increased in the AD brain (Sonnen, Larson et al. 2009). There is compelling evidence

that much of the oxidative damage observed in the AD brain is due to free radical

production by microglia and precedes Aβ deposition (Wilkinson, Koenigsknecht-Talboo

et al. 2006; Pratico 2008). The etiological events leading to AD remain unknown;

however, our findings suggest inflammation and oxidative damage play critical roles in

AD pathogenesis.

Microglia, the brain’s principal immune effector cells are a potential source of

oxidative stress (Banati, Gehrmann et al. 1993; Akiyama, Barger et al. 2000). We have

previously demonstrated that microglia employ a multi-receptor cell surface complex,

comprised of CD36, α6β1 integrin, CD47, and the class A scavenger receptor

(Bamberger, Harris et al. 2003), TLR2/4 and CD14 (Landreth and Reed-Geaghan 2009;

Reed-Geaghan, Savage et al. 2009) to detect and respond to Aβ fibrils. Fibrillar Aβ

60

engagement of this receptor complex initiates a tyrosine kinase-based intracellular signaling cascade. Tyrosine phosphorylation of Vav faciliates Rac activation, nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX2) assembly and superoxide production (Wilkinson, Koenigsknecht-Talboo et al. 2006). The sustained microglial proinflammatory response results in overproduction of ROS, which can ultimately be neurotoxic.

A number of epidemiological studies have reported that chronic nonsteroidal anti- inflammatory drug (NSAID) therapy was associated with a dramatically reduced incidence of AD (McGeer, et al., 1996), confering a 60-80% reduction in risk (in t' Veld, et al., 2001,Stewart, et al.,1997, Vlad, et al., 2008). Long-term ibuprofen treatment also suppresses inflammation, reduces amyloid deposition, alters APP processing, and improves cognitive performance in murine models of AD (Lim, Yang et al. 2000; Lim,

Chu et al. 2001; Jantzen, Connor et al. 2002; Yan, Zhang et al. 2003; Kotilinek,

Westerman et al. 2008; McKee, Carreras et al. 2008). Together, these findings led to clinical trials of NSAIDs in AD that failed to demonstrate any benefit to patients(Aisen,

Schafer et al. 2003; Reines, Block et al. 2004; Group, Lyketsos et al. 2007; Arvanitakis,

Grodstein et al. 2008; Breitner, Haneuse et al. 2009). A recent renewed interest in this class of drugs for AD treatment stems from findings suggesting that some, but not all,

NSAIDs can act independently from their classic anti-inflammatory mechanisms, which may play a role in their disease-modifying actions (Lehmann, Lenhard et al. 1997;

Combs, Johnson et al. 2000; Weggen, Eriksen et al. 2001; Eriksen, Sagi et al. 2003;

Zhou, Su et al. 2003; Lleo, Berezovska et al. 2004). These findings raise the question of how NSAIDs might influence other pathogenic features of AD such as oxidative damage.

61

We have investigated whether chronic ibuprofen treatment could alter AD-related oxidative damage in a mouse model, and how ibuprofen might inhibit intracellular signaling cascades responsible for NOX2 assembly and release of ROS.

Materials and Methods

Materials- Ibuprofen was obtained from Sigma (St. Louis, MO). This compound was formulated into standard, color-coded animal chow by Research Diets (New Brunswick,

NJ) at a final concentration of 375 ppm ibuprofen.

The Aβ peptide corresponding to the human Aβ amino acids 25-35 was purchased

from American Peptide Co. (Sunnyvale, CA). The method used to fibrillarize Aβ

peptides has been well characterized (Burdick, Soreghan et al. 1992; Lorenzo and

Yankner 1994).

Transgenic mice and ibuprofen treatment - The B6-R1.40 transgenic mouse line contains hAPP with the Swedish (K670M/N671L) FAD mutation as previously described

(Lamb, Bardel et al. 1999; Lehman, Kulnane et al. 2003) . Fifteen-month-old male and female B6-R1.40 mice were fed drug-supplemented or control chow ad libitum for 9 months. The amount of animal chow consumed was approximately 5 g/day/animal, resulting in a final dosage of 62.5 mg/kg/day as previously described (Yan, Zhang et al.

2003). Mice were observed on a weekly basis and exhibited no overt signs of distress.

Mice were sacrificed at 24 months of age. All animal studies were approved by the Case

Western Reserve University School of Medicine Institutional Animal Care and Use

Committee.

62

Histology and Immunohistochemistry –Mice were anesthetized with Avertin (0.02

cc/mg body weight) and perfused transcardially with 0.1 M sodium phosphate buffer

followed by 4% paraformaldehyde. Brains were dissected, post-fixed, cryoprotected and

sagittally sectioned (10 µm). Tissue sections were incubated overnight at 4°C with either

6E10 (Signet Laboratories, USA; 1:1000), CD45 (Serotec, USA; 1:300), or Iba1 (Wako,

Japan; 1:500) antibodies. Sections treated with anti-Aβ (6E10) and anti-CD45 antibodies were then incubated with the appropriate biotinylated secondary antibodies, and detected through the avidin-biotin-peroxidase complex (Vector, USA). Peroxidase activity was visualized by diaminobenzidine (Vector, USA). For immunofluorescent staining, Iba1 was detected with an Alexa Fluor 488 antibody and 6E10 was detected with Alexa Fluor

546 antibody (Molecular Probes, USA; 1:1000).

Thioflavin-S staining was performed to visualize dense core plaques. Sections were rehydrated and then stained with 1% Thioflavin S (Sigma). Nuclei were visualized with a propidium iodide (0.15 μM) counterstain.

Tissue homogenization and Western blotting –Animals were sacrificed by cervical dislocation and brain tissue was immediately removed. Brains were bisected sagittally along the midline. Hemibrains, excluding the cerebellum, were homogenized in ice-cold

tris-buffered saline with protease inhibitors (0.5 mM PMSF, 0.2 mM Na3VO4, protease

inhibitor cocktail (Sigma, 1:100), 1 mM EDTA) using a glass-on-glass homogenizer.

The homogenate was centrifuged at 5,000 rcf for 10 min at 4˚C. Protein concentration

was determined by the Bradford method (Bradford 1976).

63

Lysates from brain homogenates were resolved by SDS PAGE on a 4-12% Bis

Tris gel (Invitrogen, USA) and transferred to polyvinylidene difluoride (PVDF)

membranes. Blots were incubated overnight with either anti-3-nitrotyrosine (Alpha

Diagnostics, USA; 1:1000), anti-4-HNE (Chemicon, USA; 1:2000), or anti- dinitrophenylhydrazine (DNPH) (Chemicon, USA; 1:150) antibodies at 4˚C. Proteins were detected by chemiluminescence (Pierce, USA). Blots were stripped and reprobed with anti-GAPDH (Trevigen, USA; 1:5000) as a protein loading control. Band intensities were quantified using NIH Image 1.62 software (Bethesda, MD).

Tissue Culture -- Human THP-1 monocytes (American Type Culture Collection, USA) were grown in RPMI 1640 medium (Whittaker Bioproducts, USA) containing 10% heat-

inactivated fetal bovine serum (Hyclone, USA), 5 × 10 5 M 2-mercaptoethanol, 5 mM

HEPES, and 15 µg/ml gentamycin in 5% CO2. THP-1 monocytes are used in these assays as they do not attach to the tissue culture substrate through integrin-based adhesive

mechanisms, allowing dissection of Aβ fibril-dependent signaling mechanisms in the

absence of high basal levels of tyrosine kinase-based integrin signaling. THP-1

monocytes response to fAβ peptides faithfully replicates the response observed in

primary microglia (Combs, Johnson et al. 1999; Combs, Johnson et al. 2000; Combs,

Bates et al. 2001; Bamberger, Harris et al. 2003; Koenigsknecht and Landreth 2004).

Primary microgliawere obtained from postnatal day 1-3 mouse brains as described

previously (Combs, Johnson et al. 1999; Combs, Johnson et al. 2000; Combs, Bates et al.

2001).

Cell Stimulation and Immunoprecipitations THP-1 monocytes were collected and resuspended in Hank’s Balanced Salt Solution (HBSS) and preincubated with racemic

64

ibuprofen, the S- or-R-enantiomers of ibuprofen or cycloxygenase inhibitors for 1 h at

37°C. Vav immunoprecipitations were performed as previously described (Wilkinson,

Koenigsknecht-Talboo et al. 2006). For western blotting, samples were resolved on 9 or

12% SDS-PAGE gels and transferred as mentioned above. Blots were probed with either

anti-phospho-Tyr (4G10; Upstate, USA; 1:1000), -Vav (Santa Cruz, USA; 1:1000) -

phospho-p38 (Cell Signaling; USA 1:1000), or -p38 (Santa Cruz, USA; 1:1000)

antibodies overnight at 4°C. The proteins were detected by as mentioned above and

reprobed with primary antibody for load controls.

Cellular Fractionation THP-1 cells were fractionated as previously described

(Wilkinson, et al. 2006). THP-1 cells (6 × 106 cells) were collected and resuspended in

HBSS for 30 min at 37°C followed by preincubation with S-ibuprofen for 1 hr. Cells

were then stimulated for 10 min with fAβ25-35 (60 μM) and lysed in relaxation buffer (100

mM KCl, 3 mM NaCl, 3.5 mM MgCl2, 1.25 mM EGTA, and 10 mM PIPES, pH 7.3) on

ice for 15 min followed by 10 s of sonication. Cells were cleared by centrifugation at

5000 x g for 5 min at 4°C. The supernatant was then centrifuged for 1 h at 110,000 x g at

4 °C in a Beckman Coulter SW50.1 rotor. The resulting supernatant was removed and

saved as the "cytosolic" fraction, and the membrane pellet was resuspended in relaxation

buffer with 1% Igepal (NP40). Lysates were resolved on a 4-12% Bis-Tris gel,

transferred and blocked. The blots were probed overnight with an anti-Rac antibody

(1:1000) to determine the relative amount of Rac in each fraction and an anti-flotillin

(1:1000) antibody as a membrane marker to assess the efficacy of the fractionation procedure.

65

Measurement of superoxide production Intracellular superoxide radical generation was

assayed by nitroblue tetrazolium (NBT, Roche, Basel, Switzerland) reduction as

previously described (Wilkinson, Koenigsknecht-Talboo et al. 2006) . For these

experiments, primary microglia from C57BL/6 mice were plated overnight in serum-free

DMEM-F12. The microglia were pretreated for 1hr at 37ºC with S-ibuprofen. NBT

(1mg/ml) with or without fAβ25-35 (60 μM) in serum-free DMEM-F12 was then added to the wells for 30 min. Phorbol 12-myristate 13-acetate (PMA; 390 nM)) was used as a positive control (McDonald, Brunden et al. 1997; Bianca, Dusi et al. 1999; Bamberger,

Harris et al. 2003). Three random fields of cells (>100 cells) were counted on an inverted

microscope.

Statistical Analysis— All experiments were performed a minimum of three times. Data

from each experiment are expressed as mean ±standard deviation. Two-tailed Student’s

t-test was performed between +IBU and –IBU samples. Values statistically different

from controls were calculated using a one-way ANOVA, and the Tukey-Kramer multiple-comparisons test was used to determine p-values. Significance was considered at

a probability (p) value equal of less than 0.05.

Results Chronic ibuprofen treatment significantly reduces amyloid deposition in aged B6-

R1.40 mice- The R1.40 mouse develops extracellular Aβ deposits and associated

neuropathology that closely resembles alterations observed in human AD (Lamb, Bardel

et al. 1999; Kulnane and Lamb 2001; Lehman, Kulnane et al. 2003). Few extracellular

Aβ plaques are evident in the B6-R1.40 mouse between 14-15 months of age. We began

chronic ibuprofen treatment at 15 months of age and sacrificed the animals at 24 months

66

of age. We observed an approximate 90% reduction in A  plaque depo

plaques/section) in the parenchyma of aged, 24-month-old ibuprofen-treated B6-R1.40

animals (n=5) when compared to age-matched non-treated B6-R1.40 animals (n=5) as measured by quantitative 6E10 immunohistochemistry (p< 0.001; Figure 1A-B). We also observed a corresponding reduction in Thioflavin-S positive, compact Aβ plaques in the parenchyma of the ibuprofen-treated animals (Figure 1C-D). Notably, Thioflavin-S positive amyloid deposition is present in the cerebral vessel walls of both the ibuprofen- treated and the control animals. These results in the B6-R1.40 mice are similar to effects observed in other ibuprofen-treated AD animal models (Lim, Yang et al. 2000; Lim, Chu et al. 2001; Yan, Zhang et al. 2003), but differ from those reported by Jantzen et al. who found a reduction in overall plaque burden with no changes in Congo red stained plaques

(Jantzen, Connor et al. 2002). Ibuprofen treatment has previously been shown to suppress the phenotypic activation of microglia in AD animal models (Lim, Yang et al.

2000; Yan, Zhang et al. 2003). We observed microglia with an amoeboid, “activated” morphology clustered around amyloid plaques in the control B6-R1.40 mice. However, microglia had a more ramified or “resting” morphology in the ibuprofen-treated B6-

R1.40 mice (Figure 2A). Chronic ibuprofen treatment in the B6-R1.40 mice resulted in little, if any, amyloid deposition, and we found very few microglia adjacent to the 6E10- positive amyloid plaques remaining in these mice.

Microglia activation was also assessed by CD45-immunoreactivity. CD45 is a

tyrosine phosphatase that is important for immune cell signaling, and this molecule has

been shown to be elevated in activated microglia in both the human AD brain (Masliah,

Mallory et al. 1991) and AD animal models (Wilcock, Gordon et al. 2001). Following

67

chronic ibuprofen treatment in the B6-R1.40 mouse, we observed a profound reduction in

CD45-positive microglia staining compared to control B6-1.40 mice (Figure 2B),

consistent with previous reports (Lim, Yang et al. 2000; Yan, Zhang et al. 2003).

AD-related oxidative damage is attenuated by chronic ibuprofen treatment in aged B6-

R1.40 mice- Activated of microglia are a significant source of reactive oxygen species

(ROS) production and oxidative damage in a variety of neurodegenerative diseases

including AD (Block, Zecca et al. 2007). The reduction in microglial activation

following chronic ibuprofen treatment in the B6-R1.40 mice led us to examine potential

alterations in oxidative damage. We first examined whether the B6-R1.40 mice had

increased basal levels of oxidative stress, a phenomena observed in other APP mutant

mice (Mohmmad Abdul, Sultana et al. 2006). We measured the levels of 4-

hydroxynonenal (4HNE) protein adducts in brain lysates from 24 month-old age-matched

C57BL/6 (wildtype) mice and B6-R1.40 mice. 4HNE is a product of lipid peroxidation and has been shown to exert a host of adverse biological side-effects (Uchida and

Stadtman 1992). A robust elevation of 4HNE is evident in the brains of the B6-R1.40 mice (Figure 3A). We next examined whether chronic ibuprofen treatment could ameliorate the substantial oxidative damage found in aged B6-R1.40 mice. Indeed, ibuprofen-treated B6-R1.40 mice show a 70% reduction in the accumulation of 4HNE protein adducts when compared to control animals (p < 0.01; Figure 3B) indicating lipid peroxidation is attenuated by ibuprofen treatment.

The increased production of nitrotyrosine has also been described in AD, and has shown a high correlation with disease state. The nitration of tyrosine residues is the result of the highly reactive peroxynitrite radical, which is produced by a reaction of

68 nitric oxide with the superoxide anion. The nitration of tyrosine residues in proteins compromises their action in cellular signaling and alters protein structure. We examined the presence of nitrotyrosine-containing proteins in brain lysates from ibuprofen-treated and control B6-R1.40 mice. The presence of nitrotyrosine was 60% lower in the ibuprofen-treated animals than in the control animals (p < 0.05; Figure 3C).

We also evaluated the addition of carbonyl groups to protein side chains; a sensitive method for the detection and quantification of protein oxidation (Stadtman and

Levine 2000). The carbonylation of brain proteins was 4-fold greater in the control 24 month-old B6-R1.40 mice than in the ibuprofen-treated mice (p < 0.05; Figure 4). Taken together, these data indicate that chronic ibuprofen treatment suppresses oxidative damage in the B6-R1.40 AD mouse model.

Ibuprofen treatment inhibits fAβ-stimulated ROS production in primary microglia-

Several potential sources of ROS exist within microglia including the NADPH oxidase, mitochondria respiratory chain, xanthine oxidase, microsomal enzymes, cycloxygenase and lipoxygenase. In response to fAβ; however, it has been postulated that the primary source of ROS and the source of widespread oxidative damage found in the AD brain is the microglial NADPH oxidase (NOX2) (Markesbery 1997; McDonald, Brunden et al.

1997; Bianca, Dusi et al. 1999; Shimohama, Tanino et al. 2000; Qin, Cartier et al. 2005;

Wilkinson, Koenigsknecht-Talboo et al. 2006) .The reduction in microglia activation and oxidative damage in the ibuprofen-treated R1.40 mice led us to examine whether ibuprofen could inhibit NOX2-derived ROS production in fAβ-stimulated microglia. To examine the effect of ibuprofen treatment on the fAβ-stimulated respiratory burst, we utilized primary microglia obtained from C57BL/6 mice. Analysis of intracellular

69

superoxide production was monitored by the reduction of NBT, and PMA was used as a

positive control. We observed that pretreatment with racemic ibuprofen attenuated fAβ-

stimulated superoxide production in primary microglia by 30% when compared to non-

treated controls (Figure 5A). This data suggests ibuprofen acts to inhibit a NOX2-

derived respiratory burst in primary microglia.

Ibuprofen inhibits fAβ-stimulated Vav phosphorylation in THP-1 cells- We reported a

mechanistic link between the microglia fAβ cell surface receptor complex (Bamberger,

Harris et al. 2003) and downstream signaling events leading to NOX2 complex-derived reactive oxygen production (Wilkinson, Koenigsknecht-Talboo et al. 2006; Reed-

Geaghan, Savage et al. 2009). Assembly and function of the NOX2 enzyme complex is dependent on the fAβ-stimulated phosphorylation of Vav, a guanine nucleotide exchange factor (GEF) for the Rac1 GTPase (Wilkinson, Koenigsknecht-Talboo et al. 2006). We examined whether ibuprofen could inhibit fAβ-stimulated Vav phosphorylation. THP-1 cells were pretreated with racemic ibuprofen for 1 h followed by exposure to fAβ Indeed, ibuprofen pretreatment suppressed the Tyr-phosphorylation of Vav (p < 0.05; Figure

5B).

We evaluated the ability of the individual enantiomers of ibuprofen to inhibit fAβ- stimulated Vav phosphorylation. The S-enantiomer of ibuprofen is the active enantiomer with respect to COX inhibition. Pretreatment with the S-enantiomer was able to inhibit

Vav phosphorylation in a dose-dependent manner; however, pretreatment with the R- enantiomer had no effect (p < 0.05; Figure 5C). These data argue that the inhibition of

COX activity may play a role in suppressing fAβ-induced reactive oxygen production.

We next wanted to determine whether COX-1 or COX-2 was responsible for inhibiting

70

Vav phosphorylation. THP-1 cells were pretreated for 1 h with sc-560, a COX-1 specific inhibitor, or CAY10404, a highly specific COX-2 inhibitor, followed by stimulation with fAβ. Neither inhibitors was able to prevent an increase in Vav Tyr-phosphorylation in response to the fAβ peptide (Figure 5D). These data suggest that the action of the S- enantiomer of ibuprofen on reducing Vav phosphorylation is mediated through a COX- independent mechanism.

(S)-ibuprofen disrupts downstream signaling to NOX2 complex assembly- We next established whether additional signaling events leading to NOX2 complex assembly were disrupted by ibuprofen treatment. The downstream target of Vav GEF activity is the small GTPase, Rac1, which is an integral component of the NOX2 enzyme complex.

Vav facilitates Rac GDP/GTP exchange converting Rac into its active conformation. Rac then translocates to the plasma membrane and interacts with other NOX2 enzyme components to assemble the catalytically active oxidase. Previously, we have demonstrated that fAβ stimulation increased Rac GTP-loading and promoted redistribution of Rac from the cytosol to the plasma membrane (Wilkinson,

Koenigsknecht-Talboo et al. 2006). Here, we report that S-ibuprofen pretreatment of

THP-1 cells leads to a dose-dependent inhibition of fAβ-mediated Rac translocation to the membrane (Figure 6A). These data demonstrate that treatment with S-ibuprofen impairs NOX2 complex assembly.

Since S-ibuprofen treatment inhibits signaling cascades leading to a defect in oxidase assembly, we next determined whether parallel signaling pathways responsible for activation of p47phox were also impaired. p47phox, a cytosolic component of the NOX2

enzyme complex, must be phosphorylated on serine residues to initiate translocation to

71 the membrane where it interacts with the membrane-bound cytochrome b558. It is well established that upstream p38 MAPK activity is critical for both superoxide production and p47phox phosphorylation in phagocytes (Detmers, Zhou et al. 1998; Yamamori,

Inanami et al. 2000). Important for our studies, we have previously demonstrated that p38 phosphorylation is upregulated following exposure to Aβ fibrils in THP-1 cells

(McDonald, Bamberger et al. 1998). To determine the effect of S-ibuprofen on p38 activity, THP-1 cells were pretreated with S-ibuprofen followed by exposure to fAβ; cellular lysates were then analyzed for levels of p38 phosphorylation. We report that S- ibuprofen pretreatment diminished Aβ-induced p38 activity and R-ibuprofen had no effect (p < 0.001; Figure 6B). These findings suggest that S-ibuprofen acts to disrupt parallel signaling cascades leading to improper assembly of the NOX2 enzyme complex.

(S)-ibuprofen inhibits fAβ-stimulated ROS production in primary microglia- In light of our previous findings that several critical signaling molecules are regulated by ibuprofen treatment, we hypothesized that S-ibuprofen might inhibit the generation of ROS in primary murine microglia stimulated with fAβ peptides. Following pretreatment with S- ibuprofen, Aβ-stimulated intracellular superoxide production was monitored in primary microglia. We observed that fAβ-induced ROS production was dramatically reduced by pretreatment with S-ibuprofen when compared to either fAβ or PMA, a positive control for respiratory burst (p < 0.001; Figure 7). These results indicate that the S-ibuprofen impairs NADPH oxidase function and ROS production in response to Aβ fibrils.

72

Discussion

NSAIDs have received considerable attention as a potential therapy for AD owing to numerous epidemiological studies that provide evidence for an association between the chronic intake of NSAIDs and a decreased risk for AD (Stewart, Kawas et al. 1997; in t'

Veld, Ruitenberg et al. 2001; Zandi, Anthony et al. 2002; Lerner, Miodownik et al.

2008). However, several recent studies found no benefit from NSAID intake on incidence of AD (Aisen, Schafer et al. 2003; Reines, Block et al. 2004; Group, Lyketsos et al. 2007; Arvanitakis, Grodstein et al. 2008; Breitner, Haneuse et al. 2009). Thus, there remains considerable confusion about the effects of NSAIDS in altering AD risk and pathogenesis. In the epidemiological study, Vlad et al. reported that ibuprofen exhibited a strong protective effect, and this protection increases with prolonged usage

(Vlad, Miller et al. 2008). Additionally, in vivo studies in murine models of AD have demonstrated that preventative long-term ibuprofen treatment inhibited brain inflammation, reduced amyloid pathology, decreased Aβ levels, and improved cognition

(Lim, Yang et al. 2000; Lim, Chu et al. 2001; Jantzen, Connor et al. 2002; Yan, Zhang et al. 2003; McKee, Carreras et al. 2008). In the present study, we report that R1.40 mice treated with ibuprofen had a 90% reduction in Aβ plaque number within the parenchyma.

These results obtained in R1.40 mice recapitulate previously reported findings in the

Tg2576 model (Lim, Yang et al. 2000; Jantzen, Connor et al. 2002; Yan, Zhang et al.

2003), but are notable for the near complete absence of amyloid deposits in the brains of these aged mice. In addition, we observe a significant reduction in microglia activation and association with Aβ plaques in the R1.40 mice treated with ibuprofen, a phenomenon

73

previously reported in ibuprofen-treated Tg2576 mice (Lim, Yang et al. 2000; Yan,

Zhang et al. 2003).

We further report that ibuprofen acts to suppress oxidative damage in the AD

brain through its capacity to inhibit NOX2-derived free radical production. The brain’s

high metabolic rate and reduced capacity for cellular regeneration makes the brain

susceptible to oxidative damage. Damage from oxidative stress has been postulated to be

an antecedent event in AD pathogenesis (Pratico, Uryu et al. 2001; Pratico and Sung

2004). Markers of oxidative damage can be detected prior to Aβ deposition in both

brains of humans(Mark, Lovell et al. 1997) and in Tg2576 mice (Pratico, Uryu et al.

2001; Park, Anrather et al. 2005). Interestingly, it has been reported that some NSAIDs exhibit antioxidant activity (Hamburger and McCay 1990; Asanuma, Nishibayashi-

Asanuma et al. 2001). We found that chronic ibuprofen treatment of R1.40 mice results in significant reductions of three independent measures of oxidative damage--lipid peroxidation, tyrosine nitration and protein oxidation, in contrast to Lim and colleagues who reported no change in the level of protein oxidation between ibuprofen and control- treated Tg2576 mice (Lim, Chu et al. 2001). It is significant to note that Lim et al. used a different AD animal model (Tg2576), a shorter treatment paradigm (6 months), and only analyzed a single marker for oxidative damage. Importantly, our current findings provide the first in vivo evidence that chronic ibuprofen treatment can alleviate AD-related oxidative damage.

The microglia phagocytic NADPH oxidase (NOX2) complex has been hypothesized to be a major source of ROS in AD (McDonald, Brunden et al. 1997;

Bianca, Dusi et al. 1999; Qin, Cartier et al. 2006). Microglia and monocytes generate a

74

NOX2-derived respiratory burst in response to fibrillar Aβ peptides (McDonald, Brunden et al. 1997; Bianca, Dusi et al. 1999) that is dependent on Aβ fibril engagement of a microglial cell surface receptor complex (Bamberger, Harris et al. 2003; Wilkinson,

Koenigsknecht-Talboo et al. 2006). The NOX2 enzyme complex localizes to both intracellular and plasma membranes catalyzing the production of superoxide from oxygen. NOX2 is comprised of several different subunits including two integral membrane-bound proteins, p22phox and gp91phox, which form the catalytic subunit

(cytochrome b558) of the complex. In resting cells, the NOX2 enzyme complex has cytosolic components (p47phox, p67phox, p40phox, and Rac1) that are distributed throughout the cytoplasm. Upon stimulation, these cytosolic components are activated by parallel signaling pathways, initiating translocation to the membrane where they interact with the membrane-bound subunits forming the active complex (Lambeth 2004; Bedard and

Krause 2007).

Recently, we have identified a mechanistic linkage between fAβ engagement of the microglial Aβ receptor complex and the initiation of intracellular signaling events regulating oxidase assembly and activation. Tyrosine phosphorylation of Vav, a guanine nucleotide exchange factor (GEF) for Rac1, and Vav’s association with Src-kinases were identified as proximal signaling events critical for ROS production in fAβ-stimulated microglia (Wilkinson, Koenigsknecht-Talboo et al. 2006). Here, we demonstrate that racemic ibuprofen and S-ibuprofen act to inhibit Aβ fibril-stimulated ROS production through disruption of the intracellular signaling pathways leading to NOX2 complex assembly and activation. Treatment with these NSAIDs abrogated Vav tyrosine phosphorylation resulting in the inability of Rac to become activated and translocate to

75

the membrane. We confirmed that NOX2 enzyme activity and ROS production was

indeed reduced in primary microglia following pretreatment with S-ibuprofen.

Pretreatment with R-ibuprofen was without effect. These data argue for the involvement

of a COX-dependent signaling mechanism. Surprisingly, the COX specific inhibitors, sc-

560 (COX-1) or CAY10404 (COX-2) failed to attenuate Vav tyrosine phosphorylation. A parallel effect of S-ibuprofen pretreatment was observed for the inhibition of fA - stimulated p38 phosphorylation. Together, these findings indicate an S-enantiomer- specific, COX-independent action of this drug. Several COX-independent actions of

NSAIDs have been documented; however, COX-independent actions of S-ibuprofen have not, to our knowledge, been reported. While the exact upstream signaling targets modulated by S-ibuprofen remain to be identified, our data suggests that S-ibuprofen may act through disruption of the action of upstream Src kinases, as global inhibition of Src- family tyrosine kinases or inhibition of phosphatidylinositol-3 kinase has been previously shown to attenuate ROS production (Bianca, Dusi et al. 1999; Wilkinson, Koenigsknecht-

Talboo et al. 2006).

The consequence of inhibiting NOX2-derived radicals in AD models has recently been examined. The contribution of NOX2-derived ROS was validated using a NOX2-

deficient (gp91phox null) macrophage cell line, which failed to kill APP-expressing

neuroblastoma cells. Interestingly, Tg2576 mice in which NOX2 was genetically

inactivated do not exhibit alterations in AD plaque pathology at the age of initial

deposition, suggesting that ROS generated by the NOX2 complex do not contribute to

processes affecting initial Aβ deposition. As expected, these animals did not develop

76 oxidative stress, cerebrovascular dysfunction, or behavioral deficits that normally would occur in the Tg2576 mice (Park, Zhou et al. 2008).

A central finding in this current study is the dramatic reduction in murine plaque burden with ibuprofen treatment, a finding consistent with previous reports (Lim, Yang et al. 2000; Lim, Chu et al. 2001; Jantzen, Connor et al. 2002; Yan, Zhang et al. 2003;

McKee, Carreras et al. 2008). The mechanistic basis of this effect remains unclear. A recent renewed interest in this class of drugs for AD treatment stems from findings suggesting that some, but not all, NSAIDs can act independently from their classic anti- inflammatory mechanisms, which may play a role in their disease-modifying actions

(Lehmann, Lenhard et al. 1997; Combs, Johnson et al. 2000; Weggen, Eriksen et al.

2001; Eriksen, Sagi et al. 2003; Zhou, Su et al. 2003; Lleo, Berezovska et al. 2004).

Koenigsknecht et al. reported that the microglial phagocytic function was suppressed in presence of inflammatory cytokines found in the AD brain and that this function could be restored upon treatment of ibuprofen or COX2 inhibitors in vitro (Koenigsknecht-Talboo and Landreth 2005), a finding consistent with those reported by Liang et al. in vivo(Liang, Wang et al. 2005). It remains possible that the reduction in activated microglia and oxidative damage in the brain is secondary to reduction in plaque burden in these animals. The role of microglia in plaque formation and remodeling has been questioned. In a recent paper by Grathwohl et al., the authors demonstrate a loss of microglia had no effect on plaque number and size over a period of 2-4 weeks, they suggest endogenous microglia play no role in formation and maintenance of Aβ plaques.

However, the authors also report a 3-4 fold increase in soluble Aβ levels in the microglia-

77

deficient animals (Grathwohl, Kalin et al. 2009), consistent with a role for microglia in

clearance of Aβ.

In summary, we provide evidence that ibuprofen acts in an enantiomer-specific manner to inhibit NADPH oxidase activation and ROS production. This is associated with a dramatic reduction in oxidative damage and amyloid deposition in a murine model of AD. The existing data from mouse models suggest that ibuprofen acts through multiple independent pathways to affect AD-related pathology. It is important to note that the outcomes of experiments in animal models have not been predictive of the effects of NSAIDs in humans and it remains unclear how these drugs affect AD risk and pathogenesis.

78

Figure 2-1: Chronic ibuprofen treatment reduces AD-related plaque pathology in

B6-R1.40 mice

(A) Sagittal sections from age-matched non-treated (-IBU) and ibuprofen-treated (+IBU)

24-month-old B6-R1.40 mice were immunstained with anti-human Aβ monoclonal antibody 6E10. (B) Average plaque number/section in the parenchyma was reduced by

90% (n=5/group, ***p<0.001) in +IBU animals. Dense core plaques were identified in

(C) -IBU and (D) +IBU mice by Thioflavin-S positive staining (green). Nuclei were visualized with propidium iodide (red). White arrows indicate blood vessels with amyloid deposition.

79

80

Figure 2-2: Chronic ibuprofen treatment reduces microglial activation in B6-R1.40

mice

(A) Representative photomicrograph depicting phenotypically activated microglia

stained with Iba1 (green) adjacent to a 6E10+ plaque (red) in the non-treated (-IBU) but

not the ibuprofen-treated (+IBU) B6-R1.40 mouse. Nuclei are stained with DAPI (blue); scale bar=50 μm. (B) Sagittal sections from -IBU and +IBU-treated B6-R1.40 mice were stained for anti-CD45; scale bar=200μm.

81

82

Figure 2-3: Ibuprofen treatment reduces AD-related oxidative damage.:

(A) Effect of the hAPP transgene on oxidative damage as measured by 4-HNE levels in

brain homogenates. Samples from 24-month-old age-matched C57BL/6 (B6) and B6-

R1.40 mice are shown. (B) Representative immunoblot from individual non-treated (-

IBU) and ibuprofen-treated (+IBU) brain homogenates analyzed for lipid peroxidation measured by 4-HNE protein adduct levels (n=5, **p<0.01). (C) Representative immunoblot from -IBU and +IBU-treated animals analyzed for 3-nitrotyrosine (3-NT) levels (n=5, *p<0.05). Blots were stripped and reprobed with GAPDH as a protein loading control.

83

84

Figure 2-4: Chronic ibuprofen treatment reduced protein oxidation in aged B6-

R1.40 mice:

Protein oxidation was measured using an Oxyblot kit. Representative blot from non- treated (-IBU) and ibuprofen-treated (+IBU) brain homogenates analyzed by immunoblot analysis with an anti-DNP antibody (n=5, *p< 0.05). Blots were stripped and reprobed with GAPDH as a protein loading control.

85

86

Figure 2-5: Fibrillar Aβ-stimulated Vav phosphorylation is inhibited by ibuprofen

pretreatment.

(A) C57Bl/6 microglia were pretreated with ibuprofen (600 μM) for 1hr followed by

incubation in serum-free DMEM-F12 containing NBT +/- fA 25-35 (60 μM) for 30 min.

PMA (390 nM) was a positive control. Superoxide generation was monitored by the

presence of insoluble formazan and visualized on a Leica DMIRB inverted microscope.

Three random fields of cells (>100 cells) were counted (n=6, ***p< 0.001). (B) THP-1

6 cells (5 x 10 cells) were pretreated 1 hr +/- ibuprofen and then stimulated with fAβ25-35

for 3 min. Vav immunoprecipitates were analyzed by Western blot analysis using a phospho-Tyr antibody (4G10). Blots were stripped and reprobed with Vav as a protein- loading control. Band intensity was analyzed as the level of phophorylated Vav normalized to total Vav protein levels and expressed as relative density (n=3, *p<0.05).

(C) Dose response of S-ibuprofen pretreatment on Vav protein-Tyr phosphorylation in

THP-1 cells treated with fAβ25-35 (60 μM) for 3 min. (*p<0.05 at 200 μM and **p<0.01

at 300 μM). R-ibuprofen (200 μM) pretreatment had no effect (n=4). (D) THP-1 cells

pretreated 1 hr with either a COX1 (sc560) or a COX2 (CAY10404) inhibitor were

stimulated with fAβ25-35 for 3 min. Vav immunoprecipitates were analyzed by Western

blot using a phospho-Tyr antibody (4G10). Blots were stripped and reprobed with Vav as

a protein-loading control.

87

88

Figure 2-6: S-ibuprofen disrupts NOX2 complex assembly

(A) Dose response for THP-1 cells pretreated with S-ibuprofen for 1 hr followed by

stimulation with fAβ25-35 (60 μM) for 10 min. Lysates were subjected to differential centrifugation and membrane fractions were immunoblotted for Rac. Cell fractions were also immunoblotted with a flotillin antibody to assess the efficacy of the fractionation procedure. (B) THP-1 cells pretreated with either S- or R-ibuprofen for 1 hr were stimulated with fAβ25-35. Lysates were immunoblotted for phosphorylated p38. Blots

were stripped and reprobed with p38 as a protein loading control. Band intensity was

analyzed as the level of phophorylated p38 normalized to total p38 protein levels and

expressed as relative density (n=4, **p<0.01).

89

90

Figure 2-7: S-ibuprofen inhibits the generation of NOX2-derived radicals in

microglia stimulated with fAβ

Primary C57Bl/6 microglia were preincubated with S-ibuprofen (200 μM) for 1 hr in

serum-free DMEM-F12. NBT was added to the media and microglia were stimulated

with fAβ25-35 (60 μM) or PMA (390 nM) for 30 min. Superoxide anion generation was monitored by the presence of insoluble formazan. Three random fields of cells (>100 cells) were counted (n=3, **p < 0.01).

91

92

Chapter 3 : ApoE-directed Therapeutics Rapidly Clear β-amyloid and Reverse

Deficits in AD Mouse Models

Paige E. Cramer1, John R. Cirrito2, Daniel W. Wesson1,3,C.Y. Daniel Lee1, J. Colleen

Karlo1, Adriana E. Zinn1, Brad T. Casali1, Jessica L. Restivo2, Whitney D. Goebel2 ,

Michael J. James4, Kurt R. Brunden4, Donald A. Wilson3 and Gary E. Landreth1

1. Department of Neurosciences, Case Western Reserve University School of

Medicine, Cleveland, OH 44106 USA

2. Department of Neurology, Hope Center for Neurological Disorders, Knight

Alzheimer’s Disease Research Center, Washington University School of Medicine, St.

Louis, MO 63110 USA

3. Emotional Brain Institute, Nathan Kline Institute for Psychiatric Research and the

New York University School of Medicine, Orangeburg, NY 10962 USA

4. Center of Neurodegenerative Disease Research, Department of Pathology and

Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania,

Philadelphia, PA 19104 USA

93

Acknowledgements: We thank Dr. Mangelsdorf for discussions and M. Pendergast, G.

Casadesus and I. Nagle for technical assistance. This work was supported by the

Blanchette Hooker Rockefeller Foundation, Thome Foundation, Roby and Taft Funds for

Alzheimer’s Research, Painestone Foundation Fritz M. Rottman, PhD, and the NIA,

AG030482-03S1 to GEL; NIDCD, DC003906, RO1-AG037693 to D.A.W; NIA, K01

AG029524, NIA, P50-AG005681, Shmerler family, the Charles F. and Joanne Knight

ADRC at Washington University to J.R.C; and Marian S. Ware Alzheimer Program to

KRB. All data is archived on \\gel-server1. PEC and GEL: US Provisional Patent

Application No.: 61/224,709.

An abbreviated version was accepted for publication at Science on January 20, 2012

94

Abstract:

Alzheimer’s disease is associated with impaired clearance of β-amyloid from the brain, a process normally facilitated by apolipoprotein E (ApoE). ApoE expression is transcriptionally induced through the action of the nuclear receptors peroxisome proliferator activated receptor (PPARγ) and liver X receptors (LXR) in coordination with retinoid X receptors (RXR). Oral administration of the RXR agonist, bexarotene, to a murine model of Alzheimer’s disease resulted in enhanced clearance of soluble Aβ within hours in an apoE-dependent manner. Aβ plaque area was reduced >50% within just 72 hours. Furthermore, bexarotene stimulated the rapid reversal of cognitive, social, and olfactory deficits and improved neural circuit function. Thus, RXR activation stimulates physiological Aβ clearance mechanisms resulting in the very rapid reversal of a broad range of Aβ-induced deficits.

95

Introduction:

The most common forms of Alzheimer’s disease (AD) occur late in life and are typified, at least in part, by accumulation and deposition of β-amyloid (Aβ) within the brain (Querfurth and LaFerla 2010). Elevated levels of Aβ peptides are associated with perturbation of synaptic function and neural network activity that underlie the cognitive deficits observed in the disease (Palop and Mucke 2010). Moreover, Aβ accumulation leads to its deposition into plaques and is postulated to drive a pathologic cascade that ultimately leads to neuronal death. Importantly, it has only very recently been shown that individuals with late-onset AD produce Aβ peptides at normal levels, but have a significantly impaired ability to clear Aβ from brain (Mawuenyega, Sigurdson et al.

2010).

The most important predisposing factor for sporadic AD is allelic variation in the apolipoprotein E (apoE) gene, and possession of an apoE4 allele dramatically increases disease risk (Roses and Saunders 1994). apoE plays critical roles in the clearance and deposition of Aβ peptides. The normal function of apoE is to scaffold the formation of high density lipoprotein (HDL) particles that traffic cholesterol and phospholipids throughout the brain (Kim, Basak et al. 2009). Indeed, apoE-containing HDL particles have recently been shown to promote the proteolytic degradation and clearance of Aβ peptides and the ApoE4 gene product is impaired in this function (Jiang, Lee et al. 2008).

We (Jiang, Lee et al. 2008) and others (Riddell, Zhou et al. 2007; Donkin, Stukas et al.

2010; Fitz, Cronican et al. 2010; Suon, Zhao et al. 2010) have demonstrated that increased brain ApoE/HDL expression is associated with reduced Aβ levels and improved cognitive function in mouse models of AD.

96

The expression of apoE is transcriptionally regulated by the ligand-activated

nuclear receptors, peroxisome proliferator-activated receptor gamma (PPARγ) and liver

X receptors (LXRs) (Chawla, Boisvert et al. 2001). These receptors form obligate

heterodimers with the retinoid X receptors (RXRs) to form functional transcription

factors and are termed “permissive” as their transcriptional activity can be activated by

ligation of either member of the receptor pair (Lefebvre, Benomar et al. 2010).

PPARγ:RXR and LXR:RXR act within a linked metabolic pathway and act in a feed-

forward manner to induce the expression of apoE, its lipid transporters and the genes for

the nuclear receptors themselves (Chawla, Boisvert et al. 2001; Grathwohl, Kalin et al.

2009).

The AD brain is also characterized by a microglial-mediated inflammatory

response that impairs Aβ clearance and exacerbates the primary disease processes (Lee,

Han et al. 2010). It is of particular significance that the PPARs and LXRs act to promote

phagocytosis (Zelcer and Tontonoz 2006; Odegaard, Ricardo-Gonzalez et al. 2007;

Gonzalez, Bensinger et al. 2009; Mukundan, Odegaard et al. 2009) to suppress inflammation (Glass and Ogawa 2006; Pascual and Glass 2006; Ghisletti, Huang et al.

2009) and stimulate the conversion of macrophages/ microglia into ‘alternative’ activation states that favor Aβ clearance.

We reasoned that administration of the RXR agonist bexarotene would act to enhance normal Aβ clearance mechanisms through the simultaneous activation of the

PPARs and LXRs in the brain, leading to an increase in apoE expression, elevated HDL

levels, and reduced brain Aβ. Bexarotene is a highly specific RXR agonist (Farol and

Hymes 2004) and is an FDA approved drug (FDA 1999) with a favorable safety profile

97

(Rizvi, Marshall et al. 1999; Duvic, Hymes et al. 2001) (Rizvi, Marshall et al. 1999;

Duvic, Hymes et al. 2001).

98

Materials and Methods:

Animals

APPswe/PS1Δe9 (APP/PS1)(Jankowsky, Slunt et al. 2005),Tg2576 (Hsiao, Chapman et

al. 1996) or APPPS1-21 (Radde, Bolmont et al. 2006) mice or non-transgenic (NonTg)

littermates of the appropriate genetic background and age were orally gavaged daily for

3, 7, 9, 14, 20 or 90 days with 100 mg/kg/day bexarotene or vehicle (water) .The animals

were sacrificed, and one hemisphere was fixed and processed for immunohistochemistry.

The hippocampus and cortex were removed from the other hemisphere and snap-frozen

until subject to RNA and protein extraction as previously described (Jiang, Lee et al.

2008). All experiments involving animals followed approved protocols by the Case

Western Reserve University School of Medicine, Washington University School of

Medicine, University of Pennsylvania, or the Nathan Kline Institute for Psychiatric

Research’s Institutional Animal Care and Use Committee.

Cell culture/Western blotting:

Primary microglia and astrocytes were prepared from P0–P3 mice and purified microglia

and astrocyte cultures were obtained as previously described (Koenigsknecht-Talboo and

Landreth 2005). For Western blot analysis 4–12% Bis-Tris gels (Invitrogen, Carlsbad,

CA) were used. ApoE lipidation status was monitored by native gel electrophoresis using

4-20% Tris-glycine gels (Invitrogen, Carlsbad, CA). The following primary antibodies were used: anti-human Aβ, 6E10 (Covance, Dedham, MA); anti-ApoE, anti-β-actin

(Santa Cruz Biotechnology, Santa Cruz, CA); anti-ABCA1, anti-ABCG1 (Novus

Biologicals, Littleton, CO).

99

Quantitative Real-time PCR:

Quantification of transcripts of the Aβ degrading enzymes was performed as previously

described (Reed-Geaghan, Savage et al. 2009). Primary microglia were incubated with

the indicated doses of bexarotene for 24h. Cells were harvested and RNA was isolated

using an RNeasy Mini Prep kit (Qiagen, Valencia, CA). RNA samples (0.5 μg) were used

to synthesize cDNA using QuantiTect Reverse Transcription kit (Qiagen). cDNA was

then preamplified for 14 cycles using TaqMan PreAmp Master Mix (Applied

Biosystems) followed by performing real-time PCR in a 10 μl reaction for 40 cycles with the StepOne Plus Real Time PCR system (Applied Biosystems). Primers and FAM probe-sets were selected from the database of Applied Biosystems. Analysis of gene expression was performed using the comparative CT method (Schmittgen and Livak

2008).

Immunohistochemistry and image analysis:

Post-fixed hemispheres were coronally cryostat sectioned (10 μm). For Aβ

immunohistochemistry, methods were performed as previously described (Jiang, Lee et

al. 2008) with analysis of 6 sections/ mouse, obtained 1.2-1.5 mm from the midline,

spaced 0.05 mm apart. Images were analyzed using Image Pro-Plus software (Media

Cybernetics, Silver Spring, MD). Plaque abundance was evaluated by 6E10 and

thioflavin S staining and counterstained with DAPI and propidium Iodide, respectively.

The number of thioflavin S+ plaques and 6E10 plaque area were analyzed by a blinded

observer.

100

Intracellular Aβ clearance:

Primary mouse microglia or astrocytes were incubated with vehicle (DMSO), bexarotene,

T0070907 (10 nM), or 22S-hydroxycholesterol (10µM) for 24 hours at 37°C. Cells were then treated with 2 μg/ml soluble Aβ42 in serum-free DMEM-F12 medium for 20 hours in the presence drug. Purified human plasma apoE (rPeptide, Athens, GA) was applied at the same time as soluble Aβ42. Aβ peptide levels were determined by ELISA following lysis of the cells in 1% SDS and normalized to total protein.

In vivo Microdialysis:

In vivo microdialysis to assess brain ISF Aβ in awake APP/PS1 mice was performed as previously described(Cirrito, May et al. 2003) .Unilateral guide cannula and 2mm microdialysis probes (BR-2, Bioanalytical Systems) were implanted into the hippocampus. Microdialysis perfusion buffer was artificial CSF containing 4% bovine serum albumin with a constant flow rate of 1.0μl/minute. Samples were assessed for Aβx-

40 and Aβx-42 by sandwich ELISA. Basal levels of ISF Aβ were defined as the mean concentration of Aβ over 6-8 hours preceding drug treatment. All ISF Aβ levels were normalized to the basal Aβ concentration for that mouse. Murine ISF Aβ40 levels in non- transgenic were assessed using a similar protocol as APP/PS1 mice except with longer sampling times of 6 hours at a 0.5μl/min flow rate. Compound E was administered intraperitoneally at 20mg/kg.

Bexarotene tissue bioavailability analysis:

101

Young male B6C3F1 mice (n=3) received 5 mg/kg bexarotene dissolved in DMSO via

i.p.administration, with plasma and brain hemispheres collected 1 hour after dosing.

Bexarotenewas detected using multiple reaction monitoring of its specific collision-

induced ion transitions.Samples were separated on an Aquity BEH C18 column and peak

areas were plotted against concentration and a 1/x weighted linear regression curve was

used to quantify tissue bexarotene levels.

Behavioral Analysis:

Olfactory Assay: Tg2576 mice (12-14 mo) were orally gavaged for 3 or 9 days with 100

mg/kg/day bexarotene. Mice were screened for olfactory deficits using the odor cross-

habituation test (Wesson, Levy et al. 2010). Odors (n=4; heptanone, isoamyl acetate,

limonene, and ethyl valerate; Sigma Aldrich, St. Louis,MO) were diluted 1×10−3 in

mineral oil and applied to a cotton-applicator stick enclosed in a piece of plastic tubing.

Odors were delivered for 4 successive trials (1 block), 20sec each, separated by 30sec

inter-trial intervals, by inserting the odor stick into a port on the side of the animal’s

home cage. Testing took place during the light phase of the animals’ (12:12) light:dark

cycle. The duration of time spent investigating, defined as snout-oriented sniffing within

1cm of the odor presentation port, was recorded across all trials by a single observer

blinded to genotypes. Mice were tested in a counter-balanced order.

Contextual Fear Conditioning: Freezing behavior was monitored by automated tracking system (Coulbourn Instruments, USA). In the training phase, mice were individually placed in the shock chamber to explore the environment freely for 2 min. Mice were exposed to the conditioned stimulus (CS: an 85dB sound at 2800Hz) for 30sec. After 2

102

seconds, the unconditioned stimulus (US: 0.56mA) was delivered. After the CS/US

pairing, the mice were kept in the chamber for another 30sec to measure the immediate freezing response. This process was repeated 4 times. Retention tests were performed 24 hours later. Each mouse was returned to the same shock chamber for 2.5 min for contextual freezing measurement in the absence of tone and the percent of time frozen and number of freezes measured.

Morris Water Maze: The spatial version of the water maze task was used to examine cognitive decline and rescue by experimental treatment in all mouse groups following a modified version of previously published methodology (Paylor, Baskall-Baldini et al.

1996). Briefly, mice were tested at the end of pharmacological treatment. Animals were trained in a black circular pool (120 cm), in a well-lighted room replete of visual cues.

Pool water was whitened with non-toxic white dye and temperature was maintained at

23°C. A clear escape platform (10.5 cm in diameter) was located approximately 0.5cm beneath the water level placed in the center of four quadrants (N, S, E, W) of the pool in the same location relative to room visual cues. Animals were tested beginning in different quadrants to control for location bias. Animals were tested for 8 trials per day, subdivided into 2 blocks of 4 trials, over 6 days. Prior to the beginning of testing, the animals were allowed to swim freely in the pool for 30 seconds and then allowed to sit on the escape platform for 30 more seconds. On day 1 of testing, the platform was located in quadrant 4

(northwest quadrant (NW)) for all 8 trials. The animals were placed in the water from one of the 4 start positions at the edge of each quadrant and allowed to swim for 60 seconds.

If the animal did not find the platform during the allocated task, it was guided towards it where it remained for 15 seconds and then immediately placed back into the water from

103

the next start position for the next trial. The exact procedure was followed for four trials

(one from each start position), at which point the mouse was dried and placed back into

its home-cage (warmed with a heating pad) for 30-40 minutes until the start of the next

trial block. Swim time, path length, and swim speed were recorded using a tracking

system and software (Ethovision, Noldus Information Technology, Wageningen, The

Netherlands). On day 5, the platform was placed in the same position as day 1 for the 7

trials and removed on the 8th trial for a probe trial to test for spatial strategy and

retention. During this trial, the platform was removed and the animals were allowed to

swim for 60 seconds without the possibility to escape; percent time spent in the quadrant

where the platform was previously located was measured. On day 6, 24hrs after the last

training session, a second probe trial was repeated as described above, followed by visual

acuity testing. For visual testing, the platform was made visible by bringing it above the

water surface and clearly marked flag; all extra-maze cues were removed completely.

Eight trials in 2 4-trial blocks were administered with the platform moving to a different quadrant for each trial. Animals with escape latencies higher than 30 seconds after 8 trials were considered visually impaired and excluded from the analyses.

Nest construction assay: Nest construction with paper towel material was observed throughout the course of treatment with vehicle or bexarotene as previously described

(Wesson, Levy et al. 2010).

Electrophysiology:

Piriform cortex (PCX) local field potentials (LFPs) were recorded from awake head-fixed

104

mice (12-14mo Tg2576 or non-transgenic) treated daily with either vehicle (water) or

bexarotene. Mice were anesthetized and stainless steel wire electrodes (A-M Systems,

Inc) implanted in the anterior portion of the PCX and over the contralateral neocortex

(the reference electrode). The electrodes were then connected to a head-stage for chronic recordings. The animal was then allowed to recover on a heating blanket for 24 hours and treated daily for 2 days with Carprofen (Rimadyl, Pfizer Animal Care, 5mg/kg S.C.).

Following 3 days of recovery from surgery, mice were positioned in the restraint device

located within an enclosure box outfitted with an exhaust fan. Mice were head-fixed daily

for no more than 1 hour at a time. During fixation a Teflon odor port was positioned

~5mm from the tip of the nose for odor delivery via an automated air-dilution olfactometer. Four odors were presented, in a counterbalanced order, several times each for analysis of odor-evoked LFP activity. LFPs were recorded over 4 days (1 baseline and

3 during treatment) and the data from treatment day 3 analyzed (e.g., Fig. 3 and Fig. S4).

LFP data were recorded along with odor presentation events using CED’s Spike 2

(Cambridge Electro Inc., UK). Data were processed with Fast-Fourier transform analysis

(~2Hz bin resolution) of LFP activity before and during the odor (2 sec epochs).

105

Results

RXR activation stimulates ApoE expression and Aβ degradation.

Bexarotene treatment of primary microglia or astrocytes resulted in the

stimulation of expression of the LXR target genes, ApoE, ABCA1 and ABCG1 (Figure

1A-B). Drug treatment induced astrocytes to secrete ApoE and dramatically stimulated the synthesis of highly lipidated HDL particles (Figure 1C-D).

ApoE containing HDL particles act normally to promote the proteolytic

degradation of soluble forms of Aβ (sAβ) (Jiang, Lee et al. 2008). RXR agonist

treatment of primary microglia and primary astrocytes stimulated the degradation of sAβ

with an EC50 of 0.1nM (Figure 2A-B). Importantly, sAβ degradation was dependent on

LXR and PPARγ activity, as antagonists of these receptors, 22-S hydroxycholesterol, or

T0070907, respectively, inhibited RXR agonist-stimulated intracellular degradation of

sAβ in both microglia (Figure 2C) and astrocytes (Figure 2D). Importantly, the enhanced degradation was dependent upon ApoE, as neither primary microglia, nor astrocytes from ApoE knockout mice exhibited this effect (Figure 2E-F). Control experiments established that the uptake of sAβ was unaffected by bexarotene or by the inhibitors. Bexarotene treatment did not affect the levels of the principal Abeta proteinases, insulin degrading enzyme and neprilysin (Figure 1E-F).

ApoE mediates the rapid reduction in brain Aβ levels.

We evaluated whether RXR activation altered brain interstitial fluid (ISF) Aβ levels by in vivo microdialysis in APPswe/PS1Δe9 mice (Cirrito, May et al. 2003;

Cirrito, Deane et al. 2005; Cirrito, Kang et al. 2008). In awake freely-moving two month old mice, we orally gavaged bexarotene and monitored ISF Aβ concentrations by

106 collection of dialysate every 4 hours from a microdialysis probe positioned in the hippocampus. Strikingly, bexarotene reduced ISF Aβ levels within a few hours of drug administration, reaching a steady state reduction of 25-30% by 24 hrs that was maintained throughout the period of drug exposure (Figure 3A-B). One dose of bexarotene significantly decreased ISF Aβ40 and Aβ42 levels by 25% for over 70 hrs

(Figure 3D), with a return to baseline by 84 hours. Compound E, a potent gamma secretase inhibitor, was administered at 48 hrs, serving as a positive control for inhibition of Aβ production. Importantly, the reduction in ISF Aβ levels was due to increased Aβ clearance, as bexarotene treatment reduced the half life of Aβ from 1.4 to 0.7 hrs (Figure

3C). Bexarotene administration also enhanced clearance of endogenous mouse Aβ from the brain that was entirely reliant upon ApoE, as the drug was without effect in ApoE null mice (Figure 3E). Further, bexarotene is blood-brain barrier permeant, with no differences found between blood and brain drug levels 1hr after a single oral administration (Figure 4A).

Bexarotene stimulates the rapid clearance of Aβ plaques in young and old APP/PS1 mice:

To determine if RXR activation would affect AD-related pathology, we orally administrated bexarotene (100 mg/kg/day; or vehicle) to 6 month old APPswe/PS1Δe9 mice for 3, 7, or 14 days. We observed the rapid removal of both diffuse and compact Aβ plaques in the cortex and hippocampus. (Figure 5). We observed the progressively enhanced expression of apoE, ABCA1, ABCG1 and elevated HDL levels in both the hippocampus and cortex of bexarotene-treated mice (Figure 4 B-C). There was an approximately 30% reduction in soluble Aβ levels throughout the 14 day treatment period

107

(Figure 5A). Insoluble Aβ levels were reduced by 40% after 72 hours and progressively

decreased over the subsequent 14 days (Figure 5A). Remarkably, total (Figure 5B-C)

and Thioflavin S+ Aβ plaques (Figure 5E-F) were reduced by approximately 75% after

14 days of bexarotene treatment. Furthermore, after 3 days of bexarotene treatment we

observed plaque-associated, Aβ-laden microglia, suggesting their participation in the phagocytic removal of Aβ deposits (Figure 5D).

To assess whether bexarotene was able to decrease Aβ burden in older animals with greater plaque deposition, we treated 11 month APP/PS1 mice with bexarotene for 7 days. We found significantly reduced levels of soluble and insoluble Aβ40 and Aβ42

(Figure 6C), a 50% reduction in thioflavin S+ plaques (Figure 6D-C), and a concurrent

increase in expression of apoE, the cholesterol transporters, ABCA1 and ABCG1, and

HDL levels (Figure 6A-B). Thus, the efficacy of acute bexarotene treatment is evident in

both early and later stages of pathogenesis in this mouse model.

Chronic bexarotene treatment reduces soluble levels of Aβ in the APP/PS1 mice:

We also tested the effect of chronic bexarotene treatment (3 months, daily) of

APP/PS1 mice starting from 6 months of age. We found elevated levels of apoE,

ABCA1, ABCG1 and HDL (Figure 7A-B) in the hippocampus and cortex. Bexarotene

reduced soluble Aβ levels by approximately 30%, consistent with its ability to enhance

apoE-dependent Aβ proteolysis (Figure 7C). However, amyloid plaque burden as

assessed by 6E10 or Thioflavin S was unchanged as compared to vehicle-treated controls

in either the cortex or hippocampus of the 9 month APP/PS1 mice (Figure 7D-G).

Bexarotene treatment of an aggressive amyloidogenic mouse model stimulates the clearance of Aβ:

108

To evaluate the robustness of the effect of bexarotene, we treated an aggressive

model of amyloidosis, the APPPS1-21 mouse (14) which possesses high levels of

deposited Aβ at 7-8 months of age. This mouse model generates a 5:1 ratio of

Aβ42:Aβ40. Aβ generation begins 6 weeks of age in these mice. 7-8 month old APPPS1-

21 mice treated for 20 days with bexarotene exhibited a 25-30% reduction of soluble

Aβ40 levels and a 45-50% reduction in both Aβ40 and Aβ42 insoluble Aβ peptides

(Figure 8C). There was a 35% decrease in the number of thioflavin S+ plaques within

the cortex (Figure 8D-E). Similarly, bexarotene treatment enhanced the expression of

ABCA1, ABCG1, apoE and its lipidated forms (Figure 8A-B).

Reversal of Aβ-dependent behavioral deficits with bexarotene treatment:

There is persuasive evidence that the cognitive and behavioral deficits

characterstics of AD arise, in part, from impaired synaptic function due to soluble forms

of Aβ. We employed four different behavioral assays in AD mouse models, each of

which assesses distinct forms of cognitive, social, and sensory behaviors

First, we evaluated associative learning using a contextual fear conditioning

assay. We evaluated the effect of 7 days of bexarotene treatment in APP/PS1Δe9 mice at

a relatively early (6 mo) and late (11 mo) stage of Aβ plaque pathogenesis (Figure 9A-

B). We found that the bexarotene treatment significantly improved learning and memory,

as evidence by the percent time frozen in both the 6 month (Figure 9A) and 11 month old

mice (Figure 9B) in the contextual environment that was statistically indistinguishable to

that of wild type mice at both ages. Similarly, chronic treatment of 6 month old

APP/PS1 mice treated for 90 days (analyzed at 9 mo of age) (Figure 9C), showed drug- induced behavioral improvements in the contextual fear conditioning task. In a separate

109 mouse model, the APPPS1-21 showed similar associative learning improvements in the contextual fear conditioning assay after 20 days of bexarotene treatment (Figure 9E).

Importantly, all mice performed similarly when assessed during the cued portion of the contextual fear conditioning assay.

Second, in order to assess spatial learning and memory, we employed the Morris

Water Maze. APP/PS1 mice treated for 90 days and APPPS1-21 mice treated for 20 days exhibited improved hippocampal function following bexarotene treatment as assessed by the amount of time spent in the target quadrant in the 24 hour retention probe trial compared to the transgenic vehicle-treated mice (Figure 9D and F). Animals were trained for 5 days and tested on day 6. Mice were visually tested following the probe trial and those that were blind were excluded from analysis. Additionally, velocity of swimming and distance traveled during the probe were not statistically significant among groups.

Third, nest construction behavior is an affiliative, social behavior that becomes progressively impaired in Tg2576 mice (Wesson, Levy et al. 2010). Therefore, we tested whether bexarotene treatment restores nesting deficiencies in Tg2576 mice by treating

12-14mo Tg2576 and wild type mice with bexarotene each day and then evaluating nest construction the following morning. Whereas wild type mice reliably made complete nests (chewed materials and placed them into a corner), Tg2576 mice did not (Figure

9G). Following just 3 days of bexarotene treatment, nest construction behavior of the

Tg2576 mice improved to the point of statistical similarity with nontransgenic mice

(Figure 9G).

110

Lastly, one of the earliest manifestations of AD are olfactory sensory impairments

(Murphy 1999) and in Tg2576 mice olfactory dysfunction is highly correlated with Aβ

deposition (Wesson, Levy et al. 2010). We used an odor habituation test to show that

bexarotene treatment for 9 days restores normal odor-guided behaviors in 12-14 mo

Tg2576 mice (Tg2576 baseline v. Tg2576 Bex, Figure 9H). In particular, Tg2576 mice showed robust enhancement of odor habituation behavior (Figure 9H), whereas wild type mice treated with bexarotene did not show behavioral changes on bexarotene compare to baseline (vehicle) performance. Importantly, in a separate set of Tg2576 mice treated with bexarotene for only three days (Figure 11A), there was a robust enhancement of odor habituation behavior (Figure 11B-C), suggesting bexarotene rapid ability to improve behavioral impairments seen in these transgenic mice.

Neural network dysfunction is restored with bexarotene:

The improved behavior observed in bexarotene–treated mice suggests dramatic and global improvements of neural network function in APP mice. Odor-evoked local field potential activity (LFP) in the olfactory cortex (piriform cortex; PCX) is critical to normal odor guided behavior (Wilson, Kadohisa et al. 2006). PCX circuit disruption is implicated in impaired olfactory perception in both humans with AD (Howard, Plailly et al. 2009) and in APP mice (Wesson, Levy et al. 2010). Therefore as a behaviorally- relevant synaptic read-out of circuit status as a function of bexarotene we evaluated PCX activity with chronic electrophysiological recordings 12-14mo Tg2576 mice treated for

72 hrs with either bexarotene or vehicle. We found that Tg2576 mice exhibited significantly less LFP activity within both the beta (15-30 Hz) and gamma band (40-

75Hz) frequencies (both of which are implicated in normal network and circuit function

111 respectively (Kopell, Ermentrout et al. 2000; Wesson, Borkowski et al. 2011) compared to wild type mice (Figure 10A). Tg2576 mice administered bexarotene for 3 days showed nearly-complete restoration of LFP activity (Figure 10B). In these same mice we observed improved behavioral performance in odor habituation following bexarotene treatment (Figure 11), indicating a rapid drug-dependent normalization of local and regional circuit function in the primary olfactory pathway.

112

Discussion

RXR activation stimulates the normal physiological processes through which Aβ

is cleared from the brain. The dependence of soluble Aβ clearance on apoE validates the

mechanistic linkage between the principal genetic risk factor for AD and the cognitive

impairment that characterizes the disease (Jiang, Lee et al. 2008; Donkin, Stukas et al.

2010).

ApoE is the most important risk factor for sporadic forms of AD, although it has

remained enigmatic how the apoE genotype confers altered susceptibility to the disease

(Fan, Donkin et al. 2009; Kim, Basak et al. 2009). The data support a role for apoE in

regulating Aβ deposition and promoting the removal of these peptides from the brain

(DeMattos 2004; Jiang, Lee et al. 2008; Kim, Basak et al. 2009). These data provide

direct validation that apoE plays a critical role in the normal clearance of soluble forms of

Aβ from the brain. This function of apoE is reliant upon its lipidation status, as large

highly lipidated HDL species more efficiently promote the proteolysis of soluble Aβ

(Jiang, Lee et al. 2008), reduce in Aβ deposition (Wahrle, Jiang et al. 2008) and

ameliorate cognitive deficits (Donkin, Stukas et al. 2010; Fitz, Cronican et al. 2010).

Conversely, prevention of apoE lipidation results in elevation of soluble Aβ peptides and

a dramatic increase in plaque deposition that is associated with co-deposition of lipid-

poor apoE in the brain (Hirsch-Reinshagen, Maia et al. 2005; Koldamova, Staufenbiel et

al. 2005; Wahrle, Jiang et al. 2005).

These data provide direct evidence that RXR activation can stimulate and enhance

the normal clearance of sAβ from the brain. There is now persuasive evidence that AD-

associated cognitive impairments arise from the ability of sAβ to interfere with normal

113

synaptic communication (Palop and Mucke 2010). In murine models of AD, associative

learning and memory can be restored through reduction in brain sAβ levels, either

through administration of anti-Aβ antibodies (Dodart, Bales et al. 2002) or upon

provision of a gamma secretase inhibitor (Comery, Martone et al. 2005; Fukumoto,

Takahashi et al. 2010). Bexarotene acts rapidly to facilitate the apoE-dependent clearance

of soluble forms of Aβ, accounting for the extremely rapid change in ISF Aβ metabolism.

We have demonstrated that bexarotene treatment reverses the loss of associative learning

and memory within 7, 20, or 90 days in mice with developing or existent plaque pathology. Wellington and colleagues have recently reported that improved behavior in

APP/PS1 mice was observed upon stimulation of brain apoE and HDL levels that was reliant upon the lipidation status of apoE (Donkin, Stukas et al. 2010). We also observed changes in nesting behavior in Tg2576 mice within 72 hrs of drug administration.

Further, we found a restoration of normal olfactory behaviors and olfactory PCX circuit function in Tg2576 mice following just 3 days of treatment of bexarotene. The reduced odor-evoked activity in the PCX of APP mice was restored to near WT levels in APP mice treated with bexarotene. These data suggest that bexarotene can rapidly reverse synaptic and even cognitive deficits in murine models of AD, even those with advanced pathology.

A striking and unprecedented finding in this study was the extraordinarily rapid removal of both diffuse and compact Aβ plaques in cortex and hippocampus of

APPswe/PS1Δe9 mice after acute treatments of bexarotene. However, the behavioral improvements were poorly correlated with the microglial- mediated removal of insoluble, deposited forms of Aβ.

114

The PPAR and LXR families of nuclear receptors have recently been shown to

act as master regulators of macrophage/microglial phenotype, provoking the conversion

of these cells into M2, alternatively activated, states (Desvergne, Michalik et al. 2006;

Charo 2007; Hong, Quintero-Monzon et al. 2011).These receptors act directly to transrepress NFκB-dependent proinflammatory gene expression (Glass and Ogawa 2006)

and stimulate the expression of genes associated with tissue repair. Significantly, these

receptors act directly to stimulate phagocytosis (Zelcer, Khanlou et al. 2007; Gonzalez,

Bensinger et al. 2009; Mukundan, Odegaard et al. 2009).

In summary, the data demonstrate that RXR activation results in the very rapid

reversal of AD-related phenotypes in three different mouse models of AD over the entire

course of disease pathogenesis. These data suggest that bexarotene acts through distinct

mechanisms to consistently clear soluble forms of Aβ from the brain. We postulate that

the RXRs act to reduce Aβ levels through facilitating apoE-dependent clearance of sAβ, and this likely subserves the rapid reversal of the behavioral and synaptic deficits. In addition, RXR agonists act on macrophages/microglia to suppress inflammation, promote phagocytosis, and this process likely underlies the efficient removal of existing Aβ plaques in the short term treatments with bexarotene. The dual actions of the nuclear

receptors on apoE and the innate immune response are consistent with recent genetic

association analyses implicating them in the etiology of AD (Jones, Holmans et al. 2010).

These data suggest that RXR agonists may be of therapeutic utility in the treatment of

Alzheimer’s disease and its prodromal phases as well as in other disorders characterized

by an overabundance of Aβ, including Down syndrome.

115

Figure 3-1: Bexarotene stimulates the expression of LXR target genes.

Bexarotene stimulates the expression of LXR target genes. Primary microglia (A) and astrocytes (B) were treated for 24 hours with the indicated doses of bexarotene or vehicle

(DMSO) and gene expression measured by Western analysis (Wilcoxon signed rank test

*p<0.05, **p<0.001, n≥4). Secreted ApoE HDL particles were analyzed by native gel electrophoresis (C), and the levels of the indicated particle sizes quantified (D) (Student’s t test *p <0.05, **p<0.01, relative to vehicle treated control, n≥3). Primary microglia (E) and astrocytes (F) were treated for 24 hours with the indicated doses of bexarotene or vehicle and mRNA levels of insulin degrading enzyme (IDE) and neprilysin (NEP) were measured by qPCR.

116

117

Figure 3-2: Bexarotene stimulates the ApoE-dependent intracellular clearance of Aβ through the actions of LXR and PPARγ

Primary microglia (A) and astrocytes (B) or apoE knockout microglia (E) or astrocytes

(F) were treated for 24 hours with bexarotene or vehicle (DMSO) then soluble Aβ42 was

added for an additional 18 hours. Primary microglia (C) and astrocytes (D) were

pretreated with competitive antagonists to LXR (22-S hydroxycholesterol) (22sHC) and

PPARγ (T0070907) (T0) for 2 hours. Remaining intracellular Aβ was measured by

ELISA (Student’s T test; *p<0.05, **p<0.01, ***p< 0.001, n≥3) and fold change is

reported relative to vehicle-treated controls.

118

119

Figure 3-3:ISF levels of Aβ decrease after bexarotene treatment

Interstitial fluid Aβx-40 and Aβx-42 levels were monitored by in vivo hippocampal

microdialysis of pre-depositing 2-3 month old APP/PS1 mice (A). Baseline Aβ levels were monitored for 6 hours, then the mice were orally administered 100 mg/kg/day bexarotene (Bex) or vehicle (water) on three consecutive days with samples collected over a 60 hour period. On day 3 of bexarotene treatment, mice were co-administered a

potent γ-secretase inhibitor (Compound E, 20mg/kg i.p.). (B) The elimination half life of

ISF Aβx-40 was measured in each animal (C). In a separate cohort of 2-3 month old

APP/PS1 mice, baseline ISF Aβ levels were sampled following administration of a single

oral dose of bexarotene (100mg/kg). ISF Aβx-40 and Aβx-42 were sampled every 2-6 hours

for 4 days after treatment (D). Baseline ISF Aβ levels were measured in non-transgenic

(C57Bl/6) and apoE knockout mice (2 mo) with and without bexarotene treatment (E).

ISF Aβx-40 levels were measured between hours 7 and 12 after treatment; n=5/group.

(Student’s t test. *p<0.05, **p<0.01, ***p<0.001).

120

121

Figure 3-4: Bexarotene stimulates the expression of LXR target genes in vivo.

C57Bl/6 mice were injected intraperitoneally with bexarotene (5 mg/kg) and 1 hour later, brain and plasma bexarotene levels were measured (A). APP/PS1 mice (6 mo, n≥6/group) were gavaged with 100 mg/kg/day bexarotene for 3 or 7 days or vehicle (water) for 7 days. Hippocampus and cortex homogenates were analyzed for the expression of

ABCA1, ABCG1 and apoE by Western analysis and quantified (B). Secreted HDL particles from cortex and hippocampus homogenates from animals treated for 3, 7, or 14 days with bexarotene or water (vehicle) were analyzed by native electrophoresis and quantified (C). (Student’s T test; *p<0.05, **p<0.01, ***p<0.001) and fold change is reported relative to vehicle treated controls.

122

123

Figure 3-5: Aβ levels and plaque burden are reduced by bexarotene treatment.

APP/PS1 or non-transgenic (NonTg) mice (6 mo) were orally gavaged for 3, 7 and 14

consecutive days with bexarotene (100 mg/kg/day) or vehicle (water). Soluble and

insoluble Aβ40 and Aβ42 levels were measured by ELISA (A). Representative cortex and hippocampus sections (B and E) of vehicle and 14 day bexarotene treated mice stained with the anti-Aβ antibody, 6E10 (B) or thioflavin S (E) are shown and plaque levels quantified (C and F); n≥5 animals/group (Student’s t test. *p<0.05, **p<0.01,

***p<0.001 Scale bar: cortex 100 μm, hippocampus 200 μm). Representative image of microglia in the cortex of a 6 mo APP/PS1 mouse treated for 3 days with bexarotene (D)

(Red:6E10, Green:Iba1, Blue:DAPI, Scale bar: 10µm).

124

125

Figure 3-6: Short-term treatment of bexarotene in 11 month old APP/PS1 mice

stimulates clearance of Aβ

APP/PS1 mice (11 mo) were orally gavaged for 7 days with bexarotene (100

mg/kg/day) or vehicle (water). Hippocampus and cortex homogenates were analyzed by

Western analysis for the indicated proteins and quantified (A and B). Soluble and

insoluble Aβ40 or Aβ42 levels were measured by ELISA (C). Representative images of sections from the cortex (D) of vehicle- and Bex-treated mice stained with thioflavin S and plaque levels quantified (E); n=5 animals/group (Student’s t test. *p<0.05, **p<0.01,

***p<0.001; Scale bar 100μm).

126

127

Figure 3-7:Chronic bexarotene treatment reduces levels of soluble Aβ.

APP/PS1 mice (9 mo) were orally gavaged for 90 days with bexarotene (bex) (100

mg/kg/day) or vehicle (water). Hippocampus and cortex homogenates were analyzed by

Western analysis for the indicated proteins and quantified (A and B). Soluble and

insoluble Aβ40 or Aβ42 levels were measured by ELISA (C). Images of the cortex and hippocampus (D and E) of vehicle- and bexarotene-treated mice stained with the anti-Aβ

antibody, 6E10 (D) or thioflavin S (E) are shown and plaque levels were quantified (F

and G); n= 10 animals/group (Student’s T test; *p<0.05, ***p<0.001 Scale bar (D) cortex

100 μm, hippocampus 200 μm, (E) all 200 μm).

128

129

Figure 3-8:Bexarotene treatment of an aggressive amyloidogenic mouse model

stimulates the clearance of Aβ

APPPS1-21 mice (7-8 mo) were gavaged daily for 20 days with bexarotene (Bex) (100 mg/kg/day) or vehicle (water). Hippocampus and cortex homogenates were analyzed for the indicated proteins by Western analysis and quantified (A and B). Soluble and insoluble Aβ40 or Aβ42 levels were measured by ELISA (C). Representative images of the cortex (D ) of vehicle- and bexarotene- treated mice stained with thioflavin S are shown

and plaque levels quantified (E); n≥6 animals/group (Student’s t test; *p<0.05,**p<0.01,

***p<0.001) Scale bar (D) 100 μm.

130

131

Figure 3-9: Restoration of memory and cognition with bexarotene treatment.

Contextual fear-learning was assayed in 6 (A) and 11mo old (B) APP/PS1 mice treated

for 7 days, or in 9 mo old, APP/PS1 mice treated for 90 days (C) with vehicle or

bexarotene. APPPS1-21 mice 7-8 mo of age were treated for 20 days and performance

evaluated (E). Total freeze number was recorded in the 2.5 min test trial. Spatial

memory was assessed using the Morris water maze (D,F). Time spent in quadrant with

the hidden platform (NW quadrant) in the retention trial (24 hr post training) of 7-8 mo

old, 20 day-treated APPPS1-21 (D) and 9 mo old, 90 day-treated APP/PS1 mice (F) with

vehicle or bexarotene (Bex) 100mg/kg/day. (Non-transgenic littermates served as

controls (NonTg), n=7-14/group, Student’s t test. *p<0.05, **p<0.01). Nest construction

was monitored visually and quantified in 12-14 mo non-transgenic (NonTg) and Tg2576 mice (G). Baseline data were obtained on day 0, following daily drug treatment and addition of paper towels in clean cages. (2-tailed t test *p<0.05, **p<0.01). Odor habituation behavior in 12-14mo Tg2576 mice tested before (baseline) and after 9 days of bexarotene treatment (H) n=5/group (2-tailed t-test; **p<0.01, ***p<0.001 Tg2576 baseline vs. Tg2576 Bex).

132

133

Figure 3-10: Rescue of cortical network activity with bexarotene.

Tg2576 or non-transgenic (NonTg) mice (12-14mo) gavaged with bexarotene (Bex)

(100mg/kg) or vehicle (H2O) for 3 days following implantation with electrodes into the

PCX for LFP recordings. PCX LFP in response to the odor ethyl valerate in an awake

non-transgenic, bexarotene treated mouse. Displayed is full-band (0-100Hz, top), 15-

35Hz beta (middle) and 35-75Hz gamma band power traces (2nd-order band pass) (A).

PCX odor-evoked response magnitudes (2sec odor/2sec pre-odor) (B). (n=5 mice/group,

4 odor presentations/mouse. *p<0.05, **p<0.01, ***p<0.001, 2-tailed t-tests of mean odor-evoked magnitudes within LFP bins.)

134

135

Figure 3-11: Improvement of odor habituation behavior in 12-14mo old Tg2576 mice treated with bexarotene for 3 days

Data from same mice used for electrophysiology (Fig 10) are shown. Mice (implanted with chronic head-caps for electrophysiology) were orally gavaged vehicle (veh, water) and used for baseline olfactory behavior measures (’baseline’) (A). Subsequently, mice were orally gavaged once daily with either bexarotene (Bex) or vehicle for 3 days and again tested. Line graphs showing normalized odor habituation data across 4 successive odor presentations on baseline and day 3 (B). *p<.05, **p<.001, NonTg+Bex vs.

Tg2576+veh. #p<.05, ##p<.001, NonTg+Bex vs. Tg2576+Bex. $p<.05, Tg2576+veh vs.

Tg2576+Bex. % habituation values by trial #4 (C). Open bars=baseline, closed bars=day

3. All statistics are 2-tailed t-tests.

136

137

Chapter 4 : Discussion

138

The amyloid hypothesis, described by Hardy and Higgins in 1991, suggests that

the principal mechanisms underlying AD pathogenesis are those that either enhance Aβ

production or impair its clearance, promoting Aβ accumulation in the brain. Bateman

and colleagues have demonstrated that in LOAD, the formation of Aβ is not enhanced in

comparison to non-demented individuals, but that the clearance of Aβ is impaired by

about 30% (Bateman, Munsell et al. 2006; Mawuenyega, Sigurdson et al. 2010), such that

there is a gradual accumulation of the soluble, oligomeric forms of Aβ that causes the

synaptic deficits, neurotoxicity and subsequent cognitive impairments that are clinically

manifested in the disease(Walsh, Klyubin et al. 2002). The delayed or disrupted

clearance of Aβ not only leads to the accumulation of soluble and oligomeric species of

Aβ, but also to the formation of fibrilliar Aβ. Aβ plaques have been shown to elicit a phenotypic change in microglia to produce a number of pro-inflammatory cytokines and reactive oxygen species (Bamberger, Harris et al. 2003; Wilkinson, Koenigsknecht-

Talboo et al. 2006; Wilkinson, Cramer et al. 2012).

The primary focus of this thesis has been to elucidate the roles of two different nuclear receptor agonists and their involvement in ameliorating AD associated pathologies. Ibuprofen has been shown to be a nuclear receptor agonist. Peroxisome proliferator activated receptor gamma activation can be mediated by ibuprofen as well as a variety of other non-steroidal anti-inflammatories (NSAIDs) (Lehmann, Lenhard et al.

1997). It was postulated that the anti-inflammatory actions of the NSAIDs may be due to their ability to bind to and activate PPAR. We provide evidence that chronic ibuprofen treatment dramatically reduces the plaque burden in the R1.40 murine model of AD and significantly reduces the amount of oxidative damage in the brains of these mice as well

139

as, alters the activation state of the microglia as measured by CD45 immunoreactivity.

Ibuprofen was also shown to inhibit the Aβ-induced Vav phosphorylation to prevent

RAC1 translocation to the plasma membrane, preventing the formation of the NADPH

oxidase and superoxide formation. These results are likely through a COX-independent mechanism, possibly acting as a PPARγ agonist.

Furthermore and more importantly, we have shown that the activation of the nuclear receptor retinoid X receptor (RXR) using an FDA approved drug, bexarotene, promotes the proteolytic degradation of soluble Aβ by both microglia and astrocytes in an

LXR and PPARγ dependent manner, facilitates the clearance of the amyloid pathology in three different mouse models of AD, ameliorating AD-related behavioral deficits in these mice and improving neuronal circuit activity. Because bexarotene is FDA approved, the likelihood of translating this basic science work into human trials is imminent and very promising. One of the initial surprising findings that emerged from these studies was that the acute activation of RXR, by bexarotene, resulted in the enhanced clearance of dense- core amyloid deposits in the cortex in the APP/PS1 mouse model of AD. This observation suggests that we may have fundamentally misjudged the stability of these

structures. Importantly, we observed phagocytically active microglia containing

significant amounts of intracellular Aβ, and this is likely to be the dominant mechanism

through which Aβ deposits are removed from the brain. It is also possible that reduction

of soluble pools of Aβ alters the equilibrium resulting in release of Aβ from plaque

deposits in the brain. However, there are conflicting data on whether this occurs.

Additionally, from this study, we show that one dose of bexarotene can reduce soluble

levels of ISF Aβ within 6 hours of administration and can maintain a 25% suppression of

140

Aβ for up to 72 hours. Previous data has shown that a small change in ISF levels can

dramatically halt Aβ plaque growth and deposition (Yan, Bero et al. 2009).

In chapters 2 and 3, we provide evidence for nuclear receptor activation to

decrease plaque burden, inflammatory environment and improve the cognitive deficits

associated with Alzheimer’s disease. Though each chapter discusses different

mechanisms, we believe nuclear receptor activation is the underlying theme that resulted

in similar findings. The work documented in this thesis provides an in-depth analysis of

the role of the nuclear receptor RXR in AD pathophysiology. Due to the FDA approval,

the important result from this thesis is the possibility of translating the basic science

research of bexarotene treatment to the clinic. We have raised a number of new

questions from both chapter 2 and 3 that remain to be answered and will be discussed in

detail in this section.

NSAIDs as PPARγ agonists:

Typically, NSAIDs work by inhibiting COX activation, preventing the production

of inflammatory cytokines and ROS. In our study, we show that ibuprofen prevents the assembly of NOX2 in a COX-independent pathway and decreases microglial activation and prevents deposition of amyloid beta. Only a subset of NSAIDs have been shown to bind and activate PPARγ, including ibuprofen. PPARγ activation has not only been show to alter microglial phenotype to an anti-inflammatory ‘alternative’ activation state

(Fuentes, Roszer et al. 2010), but it has also been reported to facilitate Aβ degradation

(Yan, Zhang et al. 2003; Heneka, Sastre et al. 2005). Using the R1.40 mouse model, that begins to develop plaques between 14-15 months of age, we treated mice for 9 months

from 15-24 months of age. It has previously been shown that inflammatory cytokines,

141

reactive oxygen species and nitrogen species formation begins prior to the deposition of

Aβ. Damage from oxidative stress has been postulated to be an antecedent event in AD

pathogenesis (Pratico, Uryu et al. 2001; Pratico and Sung 2004). Markers of oxidative

damage can be detected prior to Aβ deposition in both brains of humans(Mark, Lovell et

al. 1997) and in Tg2576 mice (Pratico, Uryu et al. 2001; Park, Anrather et al. 2005). In addition, there recently has been an increased appreciation of the role that inflammation plays in the pathogenesis of Alzheimer’s disease that has arisen principally from epidemiological studies showing a dramatic effect of long-term NSAID treatment on

Alzheimer’s disease risk. Despite the failure of the clinical trials (Aisen, Schafer et al.

2003; Reines, Block et al. 2004; Group, Lyketsos et al. 2007; Arvanitakis, Grodstein et al. 2008; Breitner, Haneuse et al. 2009), there has been a recent renewed interest in this class of drugs for AD treatment, which stems from findings that some NSAIDs can act independently from their classic anti-inflammatory mechanisms, which may play a role in their disease-modifying actions (Lehmann, Lenhard et al. 1997; Combs, Johnson et al.

2000; Weggen, Eriksen et al. 2001; Eriksen, Sagi et al. 2003; Zhou, Su et al. 2003; Lleo,

Berezovska et al. 2004). These findings raise the question of how NSAIDs might influence other pathogenic features of AD. In our lab, activation of PPARγ has been shown to facilitate the degradation and clearance of Aβ through activating the expression of the Liver X receptor response element, reverse cholesterol transport genes, apoE,

ABCA1 and ABCG1 (Mandrekar-Colucci submitted). Analyzing the R1.40 mice, treated with ibuprofen for 9 months, for protein expression of apoE, ABCA1 and ABCG1 could provide a link for ibuprofen as a PPARγ agonist working in conjunction with LXR, and a rationale for the 90% reduction of Aβ plaque burden, as clearance of soluble Aβ is

142 facilitated by apoE expression and its lipdation (Jiang, Lee et al. 2008). Though this does not provide direct evidence that ibuprofen is a PPARγ agonist, it does allow for a correlation of ibuprofen as an PPARγ agonist. Utilizing PPARγ-/- mice crossed to a murine model of AD, followed by treatment with ibuprofen would test whether ibuprofen is a PPARγ agonist in its ability to decrease Aβ and prevent microglial activation.

Additionally, it would be interesting to look at other markers of ‘classical’ and

‘alternative’ activation of the microglia surrounding the plaques in the non-treated R1.40 mice compared to the microglia surrounding the ibuprofen treated R1.40 mice, as

NSAIDs have been shown to exert anti-inflammatory actions.

Polarization of microglia through PPARγ response elements and other pathways:

In both chapter 2 and chapter 3, the main question that remains unanswered is how the preexisting dense plaques were removed during the treatment of either ibuprofen or bexarotene. A 9-month treatment of ibuprofen reduced plaque burden by 90% in a set of 24 month old R1.40 mice. At 14-15 months, dense Aβ plaques are visible in the

R1.40 mice. Similarly, in 6 month APP/PS1 mice, bexarotene reduced both diffuse and thioflavin S+ plaques in a time dependent manner. After just 3 days of bexarotene treatment, plaque burden in the cortex, as measured by 6E10 immunostaining, was reduced by 50% and insoluble levels of Aβ40 and Aβ42 were reduced by ~50% after 7 days and continued to fall for the subsequent 7 days of bexarotene treatment. This dramatic effect was recapitulated in 12 month APP/PS1 mice, as we documented a 30% fall in insoluble Aβ40 and Aβ42 levels and similar decrease in thioflavin S+ plaques, as well as in the APPPS1-21 mice treated for 20 days with bexarotene, where we

143

documented a 50% fall in insoluble Aβ40 and AB42 levels and 30% decrease in the

number of thioflavin S+ plaques.

The alternative activation of macrophages has been studied very extensively in

the periphery and is becoming crucial to the understanding of disease pathogenesis in the

CNS for a variety of neurodegenerative diseases. PPARγ activation and the subsequent

expression of a variety of markers have been shown to elicit different effects than those

of classical activation markers. However, the actual role of the protein products of the

alternative activation genes remains unknown. Most are thought to be immune-

suppressive and play an integral role in tissue repair. Arginase 1, Found in the

Inflammatory Zone1 (Fizz1), YM1 and TGFβ are four markers of M2 activation, though

there are others including IL-10, that have been shown to be upregulated after PPARγ activation (Mandrekar-Colucci Submitted). Arginase 1 (Arg1) utilizes arginine as a substrate to generate proline and polyamines, both necessary for tissue repair, and diminishes substrate availability for nitric oxide (NO) production by M1 macrophages

(Wu and Morris 1998; Hesse, Modolell et al. 2001). Fizz1 is a resistin-like molecule and

has been shown to be fibrotic, play a role in angiogenesis, allergic reactions and the

inhibition of apoptosis (Holcomb, Kabakoff et al. 2000; Sun, Wang et al. 2008). YM1 is

a chitinase-like, secretory lectin. There is no human homolog, but the human AD brain

does express the closely-related chitinase genes, CHI3L1 and CHI3L2 that may play

similar roles to YM1 such as angiogensis and macrophage recruitment and MMP9

expression (Colton, Mott et al. 2006; Libreros, Garcia-Areas et al. 2011). Transforming

growth factor-β, (TGFβ) is associated with the M2c phenotype. TGFβ plays role in

inducing phagocytosis (von Bernhardi, Tichauer et al. 2011).

144

PPARγ activation has been shown to stimulate macrophages to the M2 alternative

activation state. PPARγ can polarize macrophages to the M2c state of phenotypic

activation, termed ‘acquired deactivation,’ that is associated with increased phagocytosis.

Additionally, many studies have shown macrophage M2 polarization by activation of

PPARγ as well as PPARδ (Bouhlel, Derudas et al. 2007; Charo 2007; Mukundan,

Odegaard et al. 2009). We show in chapter 3 that after 7 days of bexarotene treatment

of 6 month APP/PS1 mice we find abundant plaque-associated microglia with

intracellular Aβ. Initially, determining whether either ibuprofen treatment or bexarotene treatment can similarly polarize primary microglia to a M2 state in vitro as PPARγ activation is able, would be pivotal to the understanding of ibuprofen as a PPARγ agonist

and bexarotene working through the PPARγ/RXR nuclear receptor heterodimer pair.

Previous data has shown that 9-cis RA, an activator of RXR, can prevent the PMA induced expression of TNFα (Zhou, Shen et al. 2010), and can promote phagocytosis in

THP1 cells, a human monoyte cell line (Serghides and Kain 2001) lending support that bexarotene, the RXR agonist, promotes an anti-inflammatory, pro-phagocytic state. We expect that because bexarotene binds to RXR and RXR:PPARγ is a permissive heterodimer and that ibuprofen has been shown to activate PPARγ, that the brains of the ibuprofen or bexarotene treated mice will show an ‘alternative’ activation brain milieu.

However, this experiment could produce a negative outcome, as microglia are heterogeneous in the brain and the regulation of the M2 markers may just be expressed in a subset of the microglial cells. If we found no upregulation of the M2 markers by qRT-

PCR in full brain homogenates, then isolating brain microglia using a percoll gradient from another set of mice treated with ibuprofen or bexarotene would decrease

145

background noise and provide a more concentrated cell population to determine transcript

levels. Lastly and probably most excitingly, using laser capture microdissection would

be the pivotal experiment to determine the polarization phenotype of the plaque- associated microglia, as the technique provides for sampling of mRNA from microglia directly associated with plaques and those in the parenchyma not intimately associated

with plaques. Determining if bexarotene or ibuprofen treatment shifts the microglia

surrounding the plaques to an M2c phagocytically active phenotype and shifts the non-

plaque associated microglial cells to an M2 anti-inflammatory phenotype compared to vehicle treated controls, would not only provide evidence for PPAR:RXR or LXR:RXR

activation and a rationale for the removal of deposited Aβ, but also the data would

provide evidence of microglial heterogeneity in the diseased murine AD brain.

Microglial heterogeneity likely reflects the plasticity and versatility of these cells in

response to their microenvironment. However, questions still remain whether 1) the

polarization of microglia to an M2 activation state can be mediated by bexarotene or

ibuprofen and 2) if there is polarization to the M2c activation state, is the microglial

activation state directly responsible for the phagocytic uptake of Aβ.

Previous literature has suggested that CD45, a tyrosine phosphatase, regulates

immune responses and is thought to be an activation marker of microglia (Tan, Town et

al. 2000; Tan, Town et al. 2000). CD45 expression mediates the suppression of cytokine

induced signaling cascades. Data has suggested that CD45 and the TNF receptor

superfamily member CD40 have been postulated to antagonize each another. Ligation of

CD40 leads to classical activation of microglia, as evidenced by TNF-α and nitric oxide

release (Tan, Town et al. 2000). Zhu and colleagues demonstrate that the PSAPP CD45

146

-/- mouse has significantly increased Aβ plaque burden, increased soluble levels of Aβ, increased IL-1β and TNF-α levels, and decreased phagocytic activity of microglia. These data suggest CD45 plays a role in inhibiting an M1 brain milieu and potentially promoting an alternative activation brain milieu (Zhu, Hou et al. 2011). With this

knowledge, it would be fruitful to analyze the effect of CD45 expression along with other

M2 markers in the brains of mice treated with bexarotene or ibuprofen to see if they

overlap. Understanding the effect of CD45 and PPARγ activation, through RXR or

ibuprofen treatment, may provide another potential therapeutic avenue for AD. It is

interesting to note that 24 month R1.40 mice vehicle-treated exhibit high levels of CD45 immunoreactivity surrounding amyloid plaques, whereas the ibuprofen-treated 24 month old R1.40 mice exhibit little to no CD45 expression. CD45 expression has also been shown to be subdivided into two group, CD45-hi and CD45-low expressing cells. This distinction is often used to distinguish infiltrating monocyte derived macrophages (CD45- hi) and resident microglia (CD45-low) (Ford, Goodsall et al. 1995; Renno, Krakowski et al. 1995; Juedes and Ruddle 2001).

It has also been suggested that the phagocytic activity of microglia is enhanced

after Notch stimulation. Data from Michelucci and colleagues has shown that treatment

of microglia with Jagged1, a ligand of Notch, guides microglia toward a M2-like anti-

inflammatory state, prompting the release of IL-4 and IL-10 (Michelucci, Heurtaux et al.

2009). Neurons and astrocytes are both capable of expressing Jagged1, implicating them

in the polarization of the anti-inflammatory brain milieu. Additionally, it has been shown

that astrocyte expression of Jagged1 is modulated by inflammatory cytokine signaling,

and TGFβ dramatically increases Jagged1 expression on astrocytes, while INFγ and

147

TNFα suppress it (Elyaman, Bradshaw et al. 2007). It seems reasonable that cell to cell contact between microglia and other cells in the CNS provides a means to maintain microglial cells in an anti-inflammatory state. An increased expression of Jagged1 on astrocytes will thus influence the microglial phenotype towards a deactivated or

alternative activation, M2-like differentiation, facilitating phagocytosis. As discussed in a

later section, determining the effects of astrocytes in the polarization of the brain milieu

would be a future direction, all being contingent on whether ibuprofen or bexarotene can

manipulate the gene expression profiles of microglia cells from a pro-inflammatory M1

state to an anti-inflammatory, phagocytic M2 state.

Stimulation of microglial cells with anti-inflammatory cytokines, IL-4, IL-13 and

TGF-β or LXR agonists have been shown to prevent the suppression of pro-inflammatory

cytokines’ effects on phagocytosis (Koenigsknecht-Talboo and Landreth 2005; Zelcer,

Khanlou et al. 2007; Terwel, Steffensen et al. 2011). However, it is not known if RXR

stimulation alone increases microglial phagocytic capacity. Similarly, it is not known

whether RXR activation can rescue Aβ stimulated phagocytosis in microglia exposed to

inflammatory cytokines. To answer this question, the following experiment is proposed:

we will use primary microglia, treated for 24 hours with inflammatory cytokines (INFγ,

TNFα, IL-1β, or LPS). The following 24 hours, the microglia will be treated with the

RXR agonist, bexarotene, in the presence of the inflammatory cytokine. On day 3, the

microglia will be stimulated for 30 minutes in the presence of the inflammatory cytokine

(+/- bexarotene) with fibrillar Aβ, followed by a 30 minute incubation with 1 μm

fluorescent microspheres. The fraction of phagocytically active cells can then be

determined by FACS analysis. Stimulation of microglia with inflammatory cytokines

148 should suppress Aβ-induced phagocytosis of 1μm fluorescent beads. If RXR stimulation by bexarotene can polarize microglia from an M1, ‘classical,’ to an M2, ‘alternative,’ activation status and affect the phagocytic ability of the cells, then we would expect that bexarotene would rescue the phagocytic suppression induced by the pro-inflammatory cytokines and increase the number of cells that have ingested or internalized the beads.

Further analysis by FACS could show the number of beads internalized by each microglia to determine if basally, bexarotene or RXR activation increases phagocytic capacity of microglia.

From these studies, it would be interesting to see if the Aβ containing, phagocytic microglia observed in the 6 month APP/PS1 7 -day bexarotene treated animals, were in fact M2 polarized. In order to determine this, 6 or 12 month old APP/PS1 mice will be treated for 7 days with 100 mg/kg/day bexarotene or vehicle (water). Prior to the initiation of treatment, mice will be systemically injected with methoxy-XO4, a congo- red derivative (Klunk, Bacskai et al. 2002), that traverses the blood brain barrier and labels dense Aβ plaques for long periods of time. This allows for the visualization of the

Aβ plaque before treatment such that if there is phagocytic action, the internalized Aβ can be monitored. After the 7 days of bexarotene treatment, the mice will be sacrificed and the brain microglia will be isolated using a the percoll gradient designed to capture a concentrated, pure population of microglia or in a single cell suspension of brain homogenate by a CD11b antibody and sorted for cells containing methoxy-XO4 labeled

Aβ peptides using FACS analysis. From these cells, the mRNA will be extracted and assessed for M1 and M2 activation markers. We expect that microglia isolated from bexarotene treated mice, containing Aβ, will express genes of alternative activation,

149 while those cells that do not contain Aβ will express a combination of markers of both classical or M1 and alternative or M2 states. We expect that the bexarotene-treated mice will have higher numbers of microglia containing Aβ in comparison to vehicle treated mice.

Infiltration of monocytes and macrophages

The role of peripheral monocytes/macrophages in the AD brain is a very controversial topic. The concept that bone-marrow derived macrophages from the periphery invade the AD brain, home to amyloid plaques and clear Aβ more efficiently than their resident microglial counterparts was first introduced by Simard and Rivest

(Simard, Soulet et al. 2006; Majumdar, Chung et al. 2008). Later studies challenged this concept, claiming the bone marrow chimera and total body irradiation techniques used in the original studies compromised the blood brain barrier (BBB), thus allowing migration of cells into the brain that would not have been able to cross the BBB otherwise (Ajami,

Bennett et al. 2007; Mildner, Schmidt et al. 2007). However, more recent studies using parabiosis or directly injecting bone marrow derived cells into the circulatory system have shown that peripherally derived macrophages cross the BBB, enter the CNS and migrate to areas of damage (Tzeng and Wu 1999; Massengale, Wagers et al. 2005;

Lebson, Nash et al. 2010). In the AD brain, these peripherally derived cells, which are

lo hi + defined as CX3CR1 /Ly6C /CCR2 are attracted to and migrate to amyloid deposits.

This migration is mediated by the high levels of chemokine ligand 2 (CCL2 or MCP1), that is secreted by plaque-associated microglia and astrocytes, and acts as a ligand for

CCR2 (Prinz and Priller 2010). Indeed, bone marrow cells derived from CCR2-/- animals have been shown to be unable to infiltrate the AD brain (El Khoury, Toft et al. 2007).

150

Assuming there is a polarization change within the brains of bexarotene treated

mice, it would be interesting to determine whether the overall inflammatory milieu in Aβ-

depositing APP/PS1mice is due to a change in phenotype of endogenous-resident plaque-

associated microglia from classical to alternative, or whether the polarization is due to an

influx of peripherally derived M2-primed monocytes, which enter the brain and

differentiate into an peripherally derived microglia. Unfortunately, at this point, we do

not have the ability to conclusively answer this question, due to the lack of markers that

differentiate between resident microglia and peripherally derived macrophages. We could attempt to answer the question by isolating CD11b+ bone marrow derived monocytes from donor mice that express a GFP marker, treating the cells with bexarotene, and injecting them biweekly, due to the half life of the bone marrow derived monocytes, into a recipient APP/PS1 mouse by a subcutaneous vascular port (Lebson,

Nash et al. 2010). Animals would have to be sacrificed and brain sections would be immunostained for Iba1, labeling all microglia and 6E10, to label Aβ. The number of

Iba+/GFP+/6e10+ (infiltrating monocytes/microglia) cells would have to be counted and compared to Iba+/GFP-/6E10+ (resident microglia) cells. If RXR activation primes infiltrating monocytes, which then enter the brain, differentiate into M2 microglia and efficiently phagocytose Aβ, all microglial cells that have internalized 6E10 deposits will be GFP positive. We expect that Iba1+/GFP- cells will be unable to take up Aβ since they have not been exposed to the RXR agonist. To further verify RXR action, we could treat the APP/PS1 depositing mice with bexarotene for 3 or 7 days before injecting GFP+

BMDM into the vascular port. In this case, if Iba1+GFP- cells are able to clear amyloid deposits, it would suggest that RXR activation can polarize resident microglia to a M2

151

state. In both reports by, Colton et al. and Jimenez et al., there was noted a microglial

activation status change that may indicate infiltration of peripheral macrophages. Colton

et al. noted an upregulation of both TNFa as well as M2 markers with age in the Tg2576

as well as in the human brain, highlighting the dichotomy of microglia in the brain.

Jimenez noted a shift from M2 marker expression to M1 upregulation as noted by the

increase in TNFα and iNOS markers, though YM1 expression was still elevated in the 18

month PS1xAPP mice. Both studies promote the idea of the microenvironment playing

a role on microglial phenotypic activation.

Long term bexarotene treatment

Because the reduction of insoluble levels of Aβ was so rapid in the bexarotene- treated 6 month APP/PS1 mice, using in vivo multi-photon microscopy to visualize the reduction of Aβ burden would be valuable to determine the onset of phagocytic activity

and the speed of removal of plaques. This tool would also be relevant to determine the temporal control of bexarotene’s effects at diminishing insoluble levels of Aβ. We report

that 6 month APP/PS1 mice treated for 90 days with 100 mg/kg/day bexarotene resulted

in reduced levels of soluble Aβ40 and Aβ42 levels, only a modest reduction in insoluble

Aβ40 and Aβ42 levels and no appreciable difference in plaque area or number of thioflavin

S + plaques in the cortex or hippocampus. Our data coincides with data by Herber et al.

where an LPS intrahippocampal injection resulted in little to no change in amyloid levels

28 days after injection; however analysis at 1, 3, 7 an 14 days resulted in dramatic

reductions of plaque area. These data suggest the phagocytic activity of microglia incited

by the LPS is lost and Aβ production and subsequent accumulation is either enhanced or

Aβ removal mechanisms have shut down (Herber, Roth et al. 2004). Because we

152 reported a return of plaques during the 90 day study, using multi-photon imaging, we would be able to determine at what point the effect of bexarotene changes from being pro-Aβ clearance to mediocre-Aβ clearance. Imaging would occur once a week for 3 months to determine the effect. Using dextran red to label the microvasculature, we could reliably revisit the same plaques, and determine the Aβ plaque size and if new plaques are forming while mice are on long-term treatment of bexarotene, just as reported previously (Yan, Bero et al. 2009; Hefendehl, Wegenast-Braun et al. 2011). It would be more interesting to take the APP/PS1 mouse and cross it to the Iba1-EGFP mouse and inject methoxy-XO4 to label Aβ plaques. With this new tool, we could ascertain in real time, using a thinned cranial window and multi-photon imaging, the effect of bexarotene on phagocytosis of microglia. Additionally, this would allow for an easier isolation of microglia post animal sacrifice to determine the polarization of microglia. However, none of these tools helps us determine why after 90 days of bexarotene treatment, the Aβ plaques were unchanged in terms of area and number.

RXR and cytochrome P450

One hypothesis as to why the plaque number and burden were unchanged in the

90 day bexarotene treatment is that rexinoids are known to increase the enzymes involved in their own catabolism (Wang, Chen et al. 2008). Bexarotene and other rexinoids affect the expression of cytochrome P450 isozymes. Because we treated the mice by oral gavage everyday for 90 days, we likely increased the rate of bexarotene catabolism.

Previous studies have shown that after 4 days of 100 mg/kg/day bexaorotene in rats, the

P450 mediated hepatic metabolism increased to more than 300% that of vehicle. There are a variety of cytochrome p450 isozymes, most of which were elevated after bexarotene

153

treatment. CYP4A, CYP2B1/2 have all be shown to be dramatically upregulated after 4

days of bexarotene treatment (Howell, Shirley et al. 1998). Importantly, CYP3A, the

principal cytochrome in the human liver and small intestine, has been shown to be

involved in bexarotene metabolism, and its autoinduction is likely to lead to the decreased drug concentrations seen after repeated administratsion of bexarotene to rats

(Ligand Pharmaceutical, Inc. data on file). It has been shown that a cytochrome P450 inhibitor prolongs the in vivo half life of all-trans retinoic acid. Surprisingly, it appears that retinoids, ligands activating RAR, do not enhance microsomal hepatic metabolism

(Howell, Shirley et al. 1998). Another interesting finding is that CYP4A induction is also caused by peroxisome proliferators. RXR:PPAR heterodimers activate the signaling pathway and in addition, RXR ligands alone are able to mediate activation of the

PPAR:RXR heterodimer (Perez, Bourguet et al. 2012) .

If RXR activation changes the cytochrome P450 isozymes in humans in a similar way to those that change in rats and mice, the isozymes will alter the metabolism of the drugs themselves. Developing RXR agonists that do not dramatically stimulate levels of the cytochrome P450s would potentially allow for continuous treatment without accelerated drug catabolism and acquired clinical drug resistance. An alternative approach to overcoming the current drawbacks associated with continuous rexinoid treatment in mice, may be to coadminister bexarotene with cytochrome p450 inhibitors, which would act to maintain levels of the drug while blocking its metabolism. One inhibitor of P450 enzyme, CYP3A4, is a constituent of grapefruit juice. Others include ketoconazole, itraconazole, bifonazole and clotrimazole. They bind to cytochrome P450 isozymes and inhibit their catabolic activities (Ahmad 2011).

154

Another way to suppress the RXR cytochrome P450 induction is to treat on a different temporal scale to every other or every third day. We reported oral gavaging AD mouse models daily with 100 mg/kg/day bexarotene. This daily dosing regimen and expected cytochrome P450 induction might explain why after 90 days, we saw little to no insoluble Aβ degradation or clearance result that we reported with shorter treatment periods. Though it is interesting that the mice treated for 90 days did have elevated levels of apoE, ABCA1 and ABCG1 and HDLs, the concentration with which activation of lipid homeostasis genes is dramatically lower (DiBlasio-Smith, Arai et al. 2008) than what may be needed to alter the polarization phenotype of microglia within the brain and body of the AD mouse models. We attempted to determine the temporal scale of treatment by assessing protein elevation after 1 dose of 100mg/kg in a set of APP/PS1 animals and report that protein levels of apoE are elevated after 2 days following one dose of bexarotene (Figure 1). Further, we assessed whether a 14 day consecutive treatment followed by every other day treatment for 14 days would decrease levels of Aβ and enhance behavior and report trends that a 2 week daily treatment followed by a 2 week every other day bexarotene treatment enhances the degradation of Aβ and improves memory in 8 month old APP/PS1 mice in the contextual fear conditioning assay.

Additionally, we still saw enhanced expression in HDL-ApoE. We believe the data is not significant because the n’s are low in these experiments (Figure 2). More investigation is needed in this area to determine if temporal treatment is a valid therapy in mouse models of AD.

155

Aβ load and Cognitive Impairment Reversal

After 14 days of bexarotene treatment or after 9 months of ibuprofen treatment,

murine models of AD, displayed dramatically few Aβ plaques. Determining the kinetics

of plaque regrowth following drug discontinuation would be a novel approach for

treatment as well as provide a further the understanding of plaque deposition. Yan et al.,

Hefendehl et al., and Meyer-Luehmann et al., have all assessed plaque growth in vivo

(Meyer-Luehmann, Spires-Jones et al. 2008; Yan, Bero et al. 2009; Hefendehl,

Wegenast-Braun et al. 2011) however, none has looked after drug treatment to determine

if plaques re-grow a different rates post drug treatment. Determining whether there

remain ‘seeds’ of plaques, promoting faster re-growth would help to guide future drug

treatment strategies. In this experiment, murine AD models would be treated with either

bexarotene or ibuprofen and imaged using 2-photon microscopy over a designated time,

after which the drug would be discontinued. Imaging would continue and plaque growth

would be monitored. In addition to determining the plaque kinetics, it would also be

interesting to determine at what point after drug discontinuation, behavioral deficits

returned, such as in the Wesson et al. paper assessing the return of olfactory behavior

deficits after LXR agonist treatment discontinuation (Wesson, Borkowski et al. 2011).

In order to examine this we would require three cohorts of mice. Initially, the mice

would be treated with either bexarotene or ibuprofen for a designated amount of time--

long enough to reliably show a memory deficit improvement and a decrease in Aβ levels,

both soluble and insoluble. The first cohort would be sacrificed to obtain a baseline measurement of Aβ burden and load, post drug treatment. At this point, the drug treatment would be discontinued in a cohort of mice. During this time, a set of

156

behavioral tasks would be demanded of the mice, including novel object recognition and

olfactory discrimination, behavioral tasks that can be employed repetitively. Once there was a return of memory impairment, the second cohort of mice would be sacrificed and

Aβ levels measured. In addition to the second cohort, a third cohort of mice that were vehicle treated the entire length of the experiment would be sacrificed to measure the baseline Aβ load without any intervention. We hypothesize that a specific concentration

of soluble Aβ elicits the behavioral deficits, such that the values of soluble Aβ between

the vehicle treated and the washout animals will be similar. Because in either of the

bexarotene studies or the ibuprofen studies, there still existed few plaques after treatment,

we hypothesize that the remaining ‘seeds’ of Aβ would enhance or speed up the

reformation of insoluble plaques, leading to higher levels of insoluble Aβ in comparison

to the vehicle treated cohort. It would be interesting to further extend this study with

three additional cohorts of mice, sacrificed in the same fashion as the first three, in order to determine if a second round of drug treatment could again facilitate the removal of Aβ and improve behavior. More interestingly, it would be interesting to determine the time frame of behavioral improvement and the levels of soluble and insoluble Aβ in the double-drug wash-out animals.

SUMO There are many of post translational modifications of proteins acting to either

facilitate or prevent their functions. SUMOylation is a reversible, covalent protein

modification that is mediated by SUMO-conjugating enzymes. Many of the identified

SUMO substrates are transcription factors and their modification by SUMO alters their transcriptional activity. Nuclear receptors can be phosphorylated, acetylated, and

157

ubiquitinated to regulate their transcriptional activation and/or stability. Additionally,

SUMOylation can affect degradation, stability, localization, and most importantly,

transcriptional activity of target proteins. Lee et al. demonstrates that in neonatal rat

astrocytes, ligand application lead to differential SUMOylation of LXRα and LXRβ, such

that LXRα was SUMOylated via HDAC4 with SUMO2 and LXRβ was SUMOylated by

PIAS1 with SUMO1. This SUMOylation of LXRα and LXRβ prevented INFγ activated

STAT1 mediated gene transcription of proinflammatory cytokines (Lee, Park et al. 2009).

In another study, Ghisletti et al., demonstrated that in macrophages, after ligand binding,

LXRα and LXRβ are both SUMOylated by HDAC4 with SUMO2. In this case, LXRα/β

were present on the NfκB promoter, preventing LPS activated NfκB gene transcription by

preventing NCOR release and subsequent inflammatory gene transcription(Ghisletti,

Huang et al. 2007). PPARγ has also been shown to SUMOylated with SUMO1 via

PIAS1 in macrophages with the concurrent SUMOylation of RXRα by SUMO1 via

PIAS1 (Wadosky and Willis 2012). It is not know if RXR can be SUMOylated differentially depending on the ligand, cell type, and the heterodimer receptor RXR partners with or inflammatory stimuli. Determining the ability of nuclear receptors to transrepress a variety of pro-inflammatory genes in the AD brain would promote the development of receptor specific agonists. Additionally, using a cell culture system, we could determine if bexarotene promotes LXRα, LXRβ or PPARγ differential

SUMOylation. We could also determine if RXR is SUMOylated by the same E3 ligase as its receptor partner. In another study by Choi et al., it was initially shown that RXRα was SUMOylated with SUMO1, but in addition, it was determined that SUSP1, a

SUMO-specific protease, could remove SUMO1 from RXRα, reversing the sumoylation-

158

mediate repression of RXRα activity. SUSP1 was unable to deSUMOylate PPARγ,

suggesting it is unique for RXRα to control RXR dependent transcription; however LXRs

were not identified or discussed. Reversible SUMO modification is a potential

mechanism for the control of nuclear receptor transcriptional activity.

RXR and ApoE isoforms

One of the significant findings of this research was the ability of RXR activation to induce the LXR pathway driving expression of LXR target gene, apoE and facilitating

Aβ degradation. While mice have only one isoform of apoE, humans have three apoE isoforms: apoε2, apoε3, apoε4, which differ by only 2 amino acids, cysteine and arginine.

Apoε4 is one of the major genetic risk factors, and should be considered a gene associated with the development of LOAD, while apoε2 is considered protective for

AD(Kim, Basak et al. 2009). Studies have shown that human apoE targeted replace mice show differing levels of amyloid deposition. Mice expressing the apoε4 exhibit the highest level of amyloid deposition, while those expressing apoε2 have significantly fewer deposits (Fagan, Watson et al. 2002; Castellano, Kim et al. 2011). Data from our lab has shown that soluble Aβ degradation is affected by apoε isoform in a similar manner as well (Jiang, Lee et al. 2008). Additionally, the turnover of apoE is dependent on its lipidation status, and its ability to become lipidated is isoform dependent, such that apoε2 and apoε3 particles are capable of being loaded with large amounts of cholesterol and phospholipids forming larger HDL particles, whereas apoε4 particles are much smaller in size and are turned over more quickly (Riddell, Zhou et al. 2008). Similarly, our lab has shown that the larger the HDL particle, the more efficient in promoting Aβ degradation (Jiang, Lee et al. 2008). The apoε4 isoform has been suggested to be a loss

159 of a protective function or a gain of a toxic function. We believe the apoε4 isoform is a loss of a protective function that can be reversed by the treatment of bexarotene.

Assessing the effect of bexarotene on Aβ burden and behavior using the targeted replacement apoε2, 3 and 4 mice will answer whether the drug effect is isoform independent. We predict that treatment of mice harboring apoε4 with bexarotene, will have decreased Aβ load following treatment, as bexarotene will enhance the expression of apoE and its lipidation status in the brains of the apoε4/AD mice, facilitating the clearance of soluble Aβ. In vitro, we show bexarotene treatment enhances the lipidation of apoE regardless of isoform. In addition, bexarotene induces the expression of ABCA1 in the three astrocytoma knock-in apoE cell lines and improves the lipidation of apoE regardless of isoform (Figure 3).

RXR activation in astrocytes

An exciting finding of this study is that bexarotene treatment enhanced astrocytic degradation of Aβ. We provide in vitro data demonstrating the ability of astrocytes to take up soluble Aβ species and degrade them in an apoE dependent mechanism, similar to microglia. Additionally, we show astrocytes are just as capable of degrading soluble

Aβ in the presence of the same concentration of bexarotene as microglia. Because astrocytes make up the largest portion of glial cells in the brain, 85%, and that microglia only comprise between 5-10% of the total cells in the brain, a small contribution from astrocytes generates a large impact in soluble amyloid clearance over the entire brain.

Astrocytes play a very fundamental role in brain homeostasis. They’re typically known to provide metabolic support for neurons, protecting them by responding to a variety of neurotransmitters, neuropeptides, growth factors and cytokines. In addition,

160

they are crucial to forming the blood brain barrier. Typically, microglia have been

focused on as the dominant immune effector cell of the central nervous system, however,

astrocytes also express the toll-like receptors (El-Hage, Podhaizer et al. 2011). Activated microglia facilitate the activation of astrocytes, including proliferation, release of cytokines and calcium waves, and in turn, activated astrocytes can activate microglia that are distant to the site of activation, be it an Aβ plaque or injury(Gahtan and Overmier

1999).

Data from our lab has shown that PPARγ activation, by pioglitazone, in 12 month old APP/PS1 mice induces astrocytes to phagocytose fibrillar Aβ in vivo. Despite previous work showing astrocytes’ ability to take up Aβ (Wyss-Coray, Loike et al. 2003;

Koistinaho, Lin et al. 2004), our lab is the first to show a drug dependent effect on

astrocytic uptake of Aβ. Based on this data, we postulate that just as pioglitazone can

facilitate fibrillar Aβ uptake, bexarotene can promote astrocytic uptake of fibrillar Aβ, just as we have seen it does with microglia. We believe that astrocytes are facilitating Aβ clearance through two mechanisms: apoE dependent degradation of soluble Aβ, through non-saturable macropinocytosis and through a phagocytic mechanism allowing larger Aβ particles to be taken up and degraded.

Nuclear Receptor activation and other diseases

RXR activation by various rexinoids has shown that RXR is involved in a variety of other CNS disorders. RXRγ inactivation has been linked to affective disorders like depression. Krzyzosiak et al., showed that in RXRγ-/- mice display signs of despair and anhedonia through a downregulation of expression of the dopamine D2 receptor in the

161

nucleus accumbens and altered serotonin signaling (Krzyzosiak, Szyszka-Niagolov et al.

2010).

Recently, bexarotene was used in an open pilot trial for schizophrenia.

Schizophrenia patients who were only partially responsive to psychiatric drugs received a low dose of bexarotene (75 mg/day) for 6 weeks. At the end of the 6 weeks, all patients had improved in the Positive and Negative symptom scale, suggesting bexarotene’s utility in schizophrenia disorder. The rationale for using bexarotene lies in its activation

of the retinoid signaling pathway. Patients with schizophrenia and their families have

similar congenital anomalies, including heart defects,that are found in dysfunctional

retinoid signaling. RXR activation has been shown to activate the dopamine D2 receptors

that are known to play a role in schizophrenia and depression (Lerner, Miodownik et al.

2008).

RXR activation has also been linked to facilitating dopamine cell survival in

models of Parkinsons disease, by providing trophic support for dopamine neurons. The

RXR ligand, LG268 was shown to rescue primary dopamine neurons from hypoxia and

the dopamine neurotoxic analogue, 6-OHDA, suggesting RXR activation in the

Parkinson disease therapeutic space. (Friling, Bergsland et al. 2009)

In another study focused on RXR signaling, it was shown by Huang et al., that

RXRγ accelerates CNS remyelination. Knocking down RXRγ expression by RNA

interference or antagonizing RXRγ, the authors showed that oligodendrocyte differentiation was delayed. The researchers showed that agonizing RXR with 9-cis

retinoic acid promoted oligodendrocyte differentiation and resulted in increased

remyelination of axons after a demyelination insult in mice (Huang, Jarjour et al. 2011).

162

Regulation of RXR activity holds promise to contribute to alternative strategies to a

variety of disorders in the CNS.

Also, activating PPARγ has been shown to prevent post-ischemic inflammation and neuronal damage both in in vitro and in vivo studies of stroke (Collino, Patel et al.

2008). There are broad implications of RXR involvement as PPARγ heterodimerizes with

RXR. Testing the RXR agonist, bexarotene or the PPARγ agonist, ibuprofen against

different CNS disorders may provide alternative approaches to diseases and disorders of

the CNS.

Clinical Trials for AD With the population aging, the number of cases of AD is expected to triple in the

United States over the next 40 years causing a sense urgency to develop disease modifying therapies for AD. Despite murine models of the disease being cured of their

AD-like pathologies, unfortunately, the results have not translated to humans.

Unfortunately, there is little to no neuronal loss that is associated with pathology in mouse models of the disease, making them poor models of AD. With that knowledge, current trials for AD include approaches to target Aβ by decreasing production (Bateman,

Siemers et al. 2009), enhancing clearance (Lemere and Masliah 2010; Rinne, Brooks et

al. 2010), preventing Aβ plaque formation (Aisen, Saumier et al. 2006), and other attempts to ameliorate the toxic effects of the amyloid cascade. The advancement towards a disease modifying therapy in humans has not been successful for a variety of reasons. There has been a recent reevaluation of clinical trials of therapies for AD as the knowledge for treating patients with moderate to severe and even mild AD have proven to be ineffective as the neuronal loss and subsequent memory loss in these patients is

163

irrevocable. Measures of a successful drug outcome are also controversial. Biomarker

trials are now becoming more popular as they are definitive assessments of changes in the brain, though even biomarkers have their flaws, as reducing Aβ burden has been shown to not correlate with memory improvements after immunization trials (Kurz and

Perneczky 2011). Testing pre-symptomatic individuals whose likelihood of developing

AD is high is gaining attention as a new strategy for therapeutic testing. The goal of the

Alzheimer’s Prevention Initiative, which was initially developed from Banner

Alzheimer’s Institute, is to test pre-symptomatic patients whose likelihood of developing

AD is high, with promising AD treatments to delay, halt or even prevent the disease

(Reiman, Langbaum et al. 2011). These at-risk individuals offer a potential proof of

concept for pre-symptomatic disease modification in AD.

Conclusions

In conclusion, our results have demonstrated the ability of both ibuprofen, the

PPARγ agonist, and bexarotene, the RXR agonist, to ameliorate one of the hallmarks of

Alzheimer’s disease, the amyloid beta plaques. Ibuprofen reduced the amyloid burden by

90% and reduced the inflammation and oxidative damage in a mouse model of

Alzheimer’s disease. By inhibiting Vav phosphorylation, ibuprofen prevented the

formation of the NADPH oxidase, reducing the inflammatory milieu and associated

oxidative damage of proteins in the brains of the R1.40 mouse model of AD. Bexarotene

rapidly reversed both the Aβ plaque pathology and corrected a broad range of deficits in

cognition and memory, including neural circuit function in various mouse models of AD.

The mechanisms through which the rapid amyloid removal was achieved by bexarotene

was based on the induction of LXR target genes, apoE, ABCA1 and ABCG1, mediating

164

the degradation of soluble Aβ by increasing the size particles of lipidated apoE. A

second mechanism of RXR activation in removal of amyloid beta plaques appears to be

through an increase in phagocytic activity, possibly mediated by polarizing the microglial

cell population to the M2 phenotype, specifically the M2c phenotype. Additionally, RXR

activation facilitated astrocyte degradation of amyloid beta.

The primary significance of this research is that both ibuprofen and bexarotene

are FDA approved drugs. Though ibuprofen has been clinically tested for Alzheimer’s

disease, bexarotene has not yet. Fortunately, translating this basic science research into

clinical trials will happen at much faster than any non-FDA approved compounds.

Though bexarotene was not developed for Alzheimer’s disease therapy, it has now been

repurposed. It will be crucial to determine the effects of bexarotene use in humans carrying all three apoE isoforms and in pre-symptomatic patients with increased likelihood to develop the disease. If the same mechanism exists in humans as we have shown in mice, we believe we have developed a new ‘disease modifying’ therapy for

Alzheimer’s disease.

165

4-1 Time course of Reverse Cholesterol Transport proteins Six month old C57Bl/6 mice were treated with bexarotene 100 mg/kg and then sacrificed at different time points following dose. Protein levels from cortex and hippocampus homogenates were analyzed for ABCA1, ABCG1 and ApoE expression by Western blot.

Each lane represents an individual animal.

166

167

4-2 One month study: 14 consecutive/14 every other day Bexarotene treatment 6 month old APP/PS1 mice were treated for 2 weeks everyday followed by 2 weeks every other day with 100 mg/kg bexarotene. Soluble and insoluble Aβ was measured by

ELISA (A). Hippocampal memory was assessed by contextual fear conditioning (B).

ApoE-HDL levels were analyzed by native gel electrophoresis of cortex and hippocampal homogenates (C).

168

169

4-3 Bexarotene and ApoE isoform Human astrocytoma knock in apoE2, apoE3, or apoE4 cell lines were treated with or without 10 nM Bexarotene for 24 hours and Western blotted for ABCA1 or by native gel electrophoresis for ApoE-HDL.

170

171

Chapter 5: Works Cited

172

Ahmad, M. (2011). "Study on Cytochrome P-450 Dependent Retinoic Acid Metabolism and its Inhibitors as Potential Agents for Cancer Therapy." Sci Pharm 79(4): 921-935. Ahuja, H. S., A. Szanto, et al. (2003). "The retinoid X receptor and its ligands: versatile regulators of metabolic function, cell differentiation and cell death." J Biol Regul Homeost Agents 17(1): 29-45. Aisen, P. S., D. Saumier, et al. (2006). "A Phase II study targeting amyloid-beta with 3APS in mild- to-moderate Alzheimer disease." Neurology 67(10): 1757-1763. Aisen, P. S., K. A. Schafer, et al. (2003). "Effects of rofecoxib or naproxen vs placebo on Alzheimer disease progression: a randomized controlled trial." Jama 289(21): 2819- 2826. Ajami, B., J. L. Bennett, et al. (2007). "Local self-renewal can sustain CNS microglia maintenance and function throughout adult life." Nat Neurosci 10(12): 1538-1543. Akiyama, H., S. Barger, et al. (2000). "Inflammation and Alzheimer's disease." Neurobiol Aging 21(3): 383-421. Akiyama, H. and P. L. McGeer (1990). "Brain microglia constitutively express beta-2 integrins." J Neuroimmunol 30(1): 81-93. Akiyama, H., H. Mori, et al. (1999). "Occurrence of the diffuse amyloid beta-protein (Abeta) deposits with numerous Abeta-containing glial cells in the cerebral cortex of patients with Alzheimer's disease." Glia 25(4): 324-331. Andersson, S., N. Gustafsson, et al. (2005). "Inactivation of liver X receptor beta leads to adult- onset motor neuron degeneration in male mice." Proc Natl Acad Sci U S A 102(10): 3857-3862. Aranda, A. and A. Pascual (2001). "Nuclear hormone receptors and gene expression." Physiol Rev 81(3): 1269-1304. Arvanitakis, Z., F. Grodstein, et al. (2008). "Relation of NSAIDs to incident AD, change in cognitive function, and AD pathology." Neurology 70(23): 2219-2225. Arvanitakis, Z., R. S. Wilson, et al. (2004). "Diabetes mellitus and risk of Alzheimer disease and decline in cognitive function." Arch Neurol 61(5): 661-666. Asanuma, M., S. Nishibayashi-Asanuma, et al. (2001). "Neuroprotective effects of non-steroidal anti-inflammatory drugs by direct scavenging of nitric oxide radicals." J Neurochem 76(6): 1895-1904. Bamberger, M. E., M. E. Harris, et al. (2003). "A cell surface receptor complex for fibrillar beta- amyloid mediates microglial activation." J Neurosci 23(7): 2665-2674. Bamberger, M. E. and G. E. Landreth (2002). "Inflammation, apoptosis, and Alzheimer's disease." Neuroscientist 8(3): 276-283. Banati, R. B., J. Gehrmann, et al. (1993). "Cytotoxicity of microglia." Glia 7(1): 111-118. Bateman, R. J., L. Y. Munsell, et al. (2006). "Human amyloid-beta synthesis and clearance rates as measured in cerebrospinal fluid in vivo." Nat Med 12(7): 856-861. Bateman, R. J., E. R. Siemers, et al. (2009). "A gamma-secretase inhibitor decreases amyloid-beta production in the central nervous system." Ann Neurol 66(1): 48-54. Beach, T. G., C. L. Maarouf, et al. (2011). "Reduced clinical and postmortem measures of cardiac pathology in subjects with advanced Alzheimer's Disease." BMC Geriatr 11: 3. Bedard, K. and K. H. Krause (2007). "The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology." Physiol Rev 87(1): 245-313. Berger, J. and J. A. Wagner (2002). "Physiological and therapeutic roles of peroxisome proliferator-activated receptors." Diabetes Technol Ther 4(2): 163-174. Bertram, L. and R. E. Tanzi (2008). "Thirty years of Alzheimer's disease genetics: the implications of systematic meta-analyses." Nat Rev Neurosci 9(10): 768-778.

173

Bianca, V. D., S. Dusi, et al. (1999). "beta-amyloid activates the O-2 forming NADPH oxidase in microglia, monocytes, and neutrophils. A possible inflammatory mechanism of neuronal damage in Alzheimer's disease." J Biol Chem 274(22): 15493-15499. Birks, J. (2006). "Cholinesterase inhibitors for Alzheimer's disease." Cochrane Database Syst Rev(1): CD005593. Block, M. L., L. Zecca, et al. (2007). "Microglia-mediated neurotoxicity: uncovering the molecular mechanisms." Nat Rev Neurosci 8(1): 57-69. Bolmont, T., F. Haiss, et al. (2008). "Dynamics of the microglial/amyloid interaction indicate a role in plaque maintenance." J Neurosci 28(16): 4283-4292. Bornemann, K. D., K. H. Wiederhold, et al. (2001). "Abeta-induced inflammatory processes in microglia cells of APP23 transgenic mice." Am J Pathol 158(1): 63-73. Botting, R. M. (2010). "Vane's discovery of the mechanism of action of aspirin changed our understanding of its clinical pharmacology." Pharmacol Rep 62(3): 518-525. Bouhlel, M. A., B. Derudas, et al. (2007). "PPARgamma activation primes human monocytes into alternative M2 macrophages with anti-inflammatory properties." Cell Metab 6(2): 137- 143. Bradford, M. M. (1976). "A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding." Anal Biochem 72: 248-254. Breitner, J. C., S. J. Haneuse, et al. (2009). "Risk of dementia and AD with prior exposure to NSAIDs in an elderly community-based cohort." Neurology 72(22): 1899-1905. Bulloch, K., M. M. Miller, et al. (2008). "CD11c/EYFP transgene illuminates a discrete network of dendritic cells within the embryonic, neonatal, adult, and injured mouse brain." J Comp Neurol 508(5): 687-710. Burdick, D., B. Soreghan, et al. (1992). "Assembly and aggregation properties of synthetic Alzheimer's A4/beta amyloid peptide analogs." J Biol Chem 267(1): 546-554. Castellano, J. M., J. Kim, et al. (2011). "Human apoE isoforms differentially regulate brain amyloid-beta peptide clearance." Sci Transl Med 3(89): 89ra57. Castrillo, A. and P. Tontonoz (2004). "Nuclear receptors in macrophage biology: at the crossroads of lipid metabolism and inflammation." Annu Rev Cell Dev Biol 20: 455-480. Charo, I. F. (2007). "Macrophage polarization and insulin resistance: PPARgamma in control." Cell Metab 6(2): 96-98. Chawla, A., W. A. Boisvert, et al. (2001). "A PPAR gamma-LXR-ABCA1 pathway in macrophages is involved in cholesterol efflux and atherogenesis." Mol Cell 7(1): 161-171. Chen, J., L. G. Costa, et al. (2011). "Retinoic acid isomers up-regulate ATP binding cassette A1 and G1 and cholesterol efflux in rat astrocytes: implications for their therapeutic and teratogenic effects." J Pharmacol Exp Ther 338(3): 870-878. Choi, S. H., S. Aid, et al. (2012). "Inhibition of NADPH oxidase promotes alternative and anti- inflammatory microglial activation during neuroinflammation." J Neurochem 120(2): 292-301. Chung, H., M. I. Brazil, et al. (1999). "Uptake, degradation, and release of fibrillar and soluble forms of Alzheimer's amyloid beta-peptide by microglial cells." J Biol Chem 274(45): 32301-32308. Cirrito, J. R., R. Deane, et al. (2005). "P-glycoprotein deficiency at the blood-brain barrier increases amyloid-beta deposition in an Alzheimer disease mouse model." J Clin Invest 115(11): 3285-3290. Cirrito, J. R., J. E. Kang, et al. (2008). "Endocytosis is required for synaptic activity-dependent release of amyloid-beta in vivo." Neuron 58(1): 42-51.

174

Cirrito, J. R., P. C. May, et al. (2003). "In vivo assessment of brain interstitial fluid with microdialysis reveals plaque-associated changes in amyloid-beta metabolism and half- life." J Neurosci 23(26): 8844-8853. Cirrito, J. R., K. A. Yamada, et al. (2005). "Synaptic activity regulates interstitial fluid amyloid-beta levels in vivo." Neuron 48(6): 913-922. Cleary, J. P., D. M. Walsh, et al. (2005). "Natural oligomers of the amyloid-beta protein specifically disrupt cognitive function." Nat Neurosci 8(1): 79-84. Collino, M., N. S. Patel, et al. (2008). "PPARs as new therapeutic targets for the treatment of cerebral ischemia/reperfusion injury." Ther Adv Cardiovasc Dis 2(3): 179-197. Colton, C. A., R. T. Mott, et al. (2006). "Expression profiles for macrophage alternative activation genes in AD and in mouse models of AD." J Neuroinflammation 3: 27. Combs, C., P. Bates, et al. (2001). "Regulation of beta-amyloid stimulated proinflammatory responses by peroxisome proliferator-activated receptor alpha." Neurochem. Intl. Combs, C. K., D. E. Johnson, et al. (1999). "Identification of microglial signal transduction pathways mediating a neurotoxic response to amyloidogenic fragments of beta-amyloid and prion proteins." J Neurosci 19(3): 928-939. Combs, C. K., D. E. Johnson, et al. (2000). "Inflammatory mechanisms in Alzheimer's disease: inhibition of beta-amyloid-stimulated proinflammatory responses and neurotoxicity by PPARgamma agonists." J Neurosci 20(2): 558-567. Comery, T. A., R. L. Martone, et al. (2005). "Acute gamma-secretase inhibition improves contextual fear conditioning in the Tg2576 mouse model of Alzheimer's disease." J Neurosci 25(39): 8898-8902. Corcoran, J. P., P. L. So, et al. (2004). "Disruption of the retinoid signalling pathway causes a deposition of amyloid beta in the adult rat brain." Eur J Neurosci 20(4): 896-902. Corder, E. H., A. M. Saunders, et al. (1993). "Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families." Science 261(5123): 921-923. Cordy, J. M., I. Hussain, et al. (2003). "Exclusively targeting beta-secretase to lipid rafts by GPI- anchor addition up-regulates beta-site processing of the amyloid precursor protein." Proc Natl Acad Sci U S A 100(20): 11735-11740. Cramer, P. E., J. R. Cirrito, et al. (2012). "ApoE-Directed Therapeutics Rapidly Clear beta-Amyloid and Reverse Deficits in AD Mouse Models." Science. Davalos, D., J. Grutzendler, et al. (2005). "ATP mediates rapid microglial response to local brain injury in vivo." Nat Neurosci 8(6): 752-758. Deane, R., A. Sagare, et al. (2008). "apoE isoform-specific disruption of amyloid beta peptide clearance from mouse brain." J Clin Invest 118(12): 4002-4013. Deane, R., Z. Wu, et al. (2004). "RAGE (yin) versus LRP (yang) balance regulates alzheimer amyloid beta-peptide clearance through transport across the blood-brain barrier." Stroke 35(11 Suppl 1): 2628-2631. Deane, R. and B. V. Zlokovic (2007). "Role of the blood-brain barrier in the pathogenesis of Alzheimer's disease." Curr Alzheimer Res 4(2): 191-197. DeMattos, R. B. (2004). "Apolipoprotein E dose-dependent modulation of beta-amyloid deposition in a transgenic mouse model of Alzheimer's disease." J Mol Neurosci 23(3): 255-262. Desvergne, B., L. Michalik, et al. (2006). "Transcriptional regulation of metabolism." Physiol Rev 86(2): 465-514. Detmers, P. A., D. Zhou, et al. (1998). "Role of stress-activated mitogen-activated protein kinase (p38) in beta 2-integrin-dependent neutrophil adhesion and the adhesion-dependent oxidative burst." J Immunol 161(4): 1921-1929.

175

DiBlasio-Smith, E. A., M. Arai, et al. (2008). "Discovery and implementation of transcriptional biomarkers of synthetic LXR agonists in peripheral blood cells." J Transl Med 6: 59. Ding, Y., A. Qiao, et al. (2008). "Retinoic acid attenuates beta-amyloid deposition and rescues memory deficits in an Alzheimer's disease transgenic mouse model." J Neurosci 28(45): 11622-11634. Dodart, J. C., K. R. Bales, et al. (2002). "Immunization reverses memory deficits without reducing brain Abeta burden in Alzheimer's disease model." Nat Neurosci 5(5): 452-457. Dolev, I. and D. M. Michaelson (2004). "A nontransgenic mouse model shows inducible amyloid- beta (Abeta) peptide deposition and elucidates the role of apolipoprotein E in the amyloid cascade." Proc Natl Acad Sci U S A 101(38): 13909-13914. Dolle, P. (2009). "Developmental expression of retinoic acid receptors (RARs)." Nucl Recept Signal 7: e006. Donkin, J. J., S. Stukas, et al. (2010). "ATP-binding cassette transporter A1 mediates the beneficial effects of the liver-X-receptor agonist GW3965 on object recognition memory and amyloid burden in APP/PS1 mice." J Biol Chem. Donkin, J. J., S. Stukas, et al. (2010). "ATP-binding cassette transporter A1 mediates the beneficial effects of the liver X receptor agonist GW3965 on object recognition memory and amyloid burden in amyloid precursor protein/presenilin 1 mice." J Biol Chem 285(44): 34144-34154. Dou, Y., H. J. Wu, et al. (2012). "Microglial migration mediated by ATP-induced ATP release from lysosomes." Cell Res. Duan, R. S., X. Yang, et al. (2008). "Decreased fractalkine and increased IP-10 expression in aged brain of APP(swe) transgenic mice." Neurochem Res 33(6): 1085-1089. Duvic, M., K. Hymes, et al. (2001). "Bexarotene is effective and safe for treatment of refractory advanced-stage cutaneous T-cell lymphoma: multinational phase II-III trial results." J Clin Oncol 19(9): 2456-2471. Eckman, E. A., M. Watson, et al. (2003). "Alzheimer's disease beta-amyloid peptide is increased in mice deficient in endothelin-converting enzyme." J Biol Chem 278(4): 2081-2084. Ehehalt, R., P. Keller, et al. (2003). "Amyloidogenic processing of the Alzheimer beta-amyloid precursor protein depends on lipid rafts." J Cell Biol 160(1): 113-123. El-Hage, N., E. M. Podhaizer, et al. (2011). "Toll-like receptor expression and activation in astroglia: differential regulation by HIV-1 Tat, gp120, and morphine." Immunol Invest 40(5): 498-522. El Khoury, J., M. Toft, et al. (2007). "Ccr2 deficiency impairs microglial accumulation and accelerates progression of Alzheimer-like disease." Nat Med 13(4): 432-438. Elyaman, W., E. M. Bradshaw, et al. (2007). "JAGGED1 and delta1 differentially regulate the outcome of experimental autoimmune encephalomyelitis." J Immunol 179(9): 5990- 5998. Eriksen, J. L., S. A. Sagi, et al. (2003). "NSAIDs and enantiomers of flurbiprofen target gamma- secretase and lower Abeta 42 in vivo." J Clin Invest 112(3): 440-449. Escribano, L., A. M. Simon, et al. (2010). "Rosiglitazone rescues memory impairment in Alzheimer's transgenic mice: mechanisms involving a reduced amyloid and tau pathology." Neuropsychopharmacology 35(7): 1593-1604. Evans, D. A., D. A. Bennett, et al. (2003). "Incidence of Alzheimer disease in a biracial urban community: relation to apolipoprotein E allele status." Arch Neurol 60(2): 185-189. Evans, D. A., L. E. Hebert, et al. (1997). "Education and other measures of socioeconomic status and risk of incident Alzheimer disease in a defined population of older persons." Arch Neurol 54(11): 1399-1405.

176

Fagan, A. M., M. Watson, et al. (2002). "Human and murine ApoE markedly alters A beta metabolism before and after plaque formation in a mouse model of Alzheimer's disease." Neurobiol Dis 9(3): 305-318. Fan, J., J. Donkin, et al. (2009). "Greasing the wheels of Abeta clearance in Alzheimer's disease: the role of lipids and apolipoprotein E." Biofactors 35(3): 239-248. Farol, L. T. and K. B. Hymes (2004). "Bexarotene: a clinical review." Expert Rev Anticancer Ther 4(2): 180-188. Farooqui, A. A., T. Farooqui, et al. (2011). "Metabolic syndrome as a risk factor for neurological disorders." Cell Mol Life Sci. Farris, W., S. Mansourian, et al. (2003). "Insulin-degrading enzyme regulates the levels of insulin, amyloid beta-protein, and the beta-amyloid precursor protein intracellular domain in vivo." Proc Natl Acad Sci U S A 100(7): 4162-4167. Fassbender, K., S. Walter, et al. (2004). "The LPS receptor (CD14) links innate immunity with Alzheimer's disease." FASEB J 18(1): 203-205. FDA (1999). "Bexarotene, New Drug Application 21-055 " http://www.accessdata.fda.gov/drugsatfda_docs/nda/99/21055_Targretin.cfm. Festuccia, W. T., S. Oztezcan, et al. (2008). "Peroxisome proliferator-activated receptor-gamma- mediated positive energy balance in the rat is associated with reduced sympathetic drive to adipose tissues and thyroid status." Endocrinology 149(5): 2121-2130. Fitz, N. F., A. Cronican, et al. (2010). "Liver X receptor agonist treatment ameliorates amyloid pathology and memory deficits caused by high-fat diet in APP23 mice." J Neurosci 30(20): 6862-6872. Fitzgerald, M. L., K. J. Moore, et al. (2002). "Nuclear hormone receptors and cholesterol trafficking: the orphans find a new home." J Mol Med (Berl) 80(5): 271-281. Ford, A. L., A. L. Goodsall, et al. (1995). "Normal adult ramified microglia separated from other central nervous system macrophages by flow cytometric sorting. Phenotypic differences defined and direct ex vivo antigen presentation to myelin basic protein-reactive CD4+ T cells compared." J Immunol 154(9): 4309-4321. Forman, B. M., J. Chen, et al. (1997). "Hypolipidemic drugs, polyunsaturated fatty acids, and eicosanoids are ligands for peroxisome proliferator-activated receptors alpha and delta." Proc Natl Acad Sci U S A 94(9): 4312-4317. Frautschy, S. A., F. Yang, et al. (1998). "Microglial response to amyloid plaques in APPsw transgenic mice." Am J Pathol 152(1): 307-317. Friling, S., M. Bergsland, et al. (2009). "Activation of Retinoid X Receptor increases dopamine cell survival in models for Parkinson's disease." BMC Neurosci 10: 146. Fuentes, L., T. Roszer, et al. (2010). "Inflammatory mediators and insulin resistance in obesity: role of nuclear receptor signaling in macrophages." Mediators Inflamm 2010: 219583. Fukumoto, H., H. Takahashi, et al. (2010). "A noncompetitive BACE1 inhibitor TAK-070 ameliorates Abeta pathology and behavioral deficits in a mouse model of Alzheimer's disease." J Neurosci 30(33): 11157-11166. Funato, H., M. Yoshimura, et al. (1998). "Astrocytes containing amyloid beta-protein (Abeta)- positive granules are associated with Abeta40-positive diffuse plaques in the aged human brain." Am J Pathol 152(4): 983-992. Gahtan, E. and J. B. Overmier (1999). "Inflammatory pathogenesis in Alzheimer's disease: biological mechanisms and cognitive sequeli." Neurosci Biobehav Rev 23(5): 615-633. Ganguli, M., H. H. Dodge, et al. (2005). "Alzheimer disease and mortality: a 15-year epidemiological study." Arch Neurol 62(5): 779-784.

177

Geldmacher, D. S., T. Fritsch, et al. (2011). "A randomized pilot clinical trial of the safety of pioglitazone in treatment of patients with Alzheimer disease." Arch Neurol 68(1): 45-50. Ghisletti, S., W. Huang, et al. (2009). "Cooperative NCoR/SMRT interactions establish a corepressor-based strategy for integration of inflammatory and anti-inflammatory signaling pathways." Genes Dev 23(6): 681-693. Ghisletti, S., W. Huang, et al. (2007). "Parallel SUMOylation-dependent pathways mediate gene- and signal-specific transrepression by LXRs and PPARgamma." Mol Cell 25(1): 57-70. Giacobini, E. (1994). "Therapy for Alzheimer's disease. Symptomatic or neuroprotective?" Mol Neurobiol 9(1-3): 115-118. Ginhoux, F., M. Greter, et al. (2010). "Fate mapping analysis reveals that adult microglia derive from primitive macrophages." Science 330(6005): 841-845. Glass, C. K. and S. Ogawa (2006). "Combinatorial roles of nuclear receptors in inflammation and immunity." Nat Rev Immunol 6(1): 44-55. Glass, C. K. and M. G. Rosenfeld (2000). "The coregulator exchange in transcriptional functions of nuclear receptors." Genes Dev 14(2): 121-141. Glass, C. K. and K. Saijo (2010). "Nuclear receptor transrepression pathways that regulate inflammation in macrophages and T cells." Nat Rev Immunol 10(5): 365-376. Glenner, G. G. and C. W. Wong (1984). "Alzheimer's disease and Down's syndrome: sharing of a unique cerebrovascular amyloid fibril protein." Biochem Biophys Res Commun 122(3): 1131-1135. Glenner, G. G. and C. W. Wong (1984). "Alzheimer's disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein." Biochem Biophys Res Commun 120(3): 885-890. Gold, M., C. Alderton, et al. (2010). "Rosiglitazone monotherapy in mild-to-moderate Alzheimer's disease: results from a randomized, double-blind, placebo-controlled phase III study." Dement Geriatr Cogn Disord 30(2): 131-146. Goldgaber, D., M. I. Lerman, et al. (1987). "Characterization and chromosomal localization of a cDNA encoding brain amyloid of Alzheimer's disease." Science 235(4791): 877-880. Goldgaber, D., M. I. Lerman, et al. (1987). "Isolation, characterization, and chromosomal localization of human brain cDNA clones coding for the precursor of the amyloid of brain in Alzheimer's disease, Down's syndrome and aging." J Neural Transm Suppl 24: 23-28. Gonzalez, N. A., S. J. Bensinger, et al. (2009). "Apoptotic cells promote their own clearance and immune tolerance through activation of the nuclear receptor LXR." Immunity 31(2): 245- 258. Gordon, S. (2003). "Alternative activation of macrophages." Nat Rev Immunol 3(1): 23-35. Gorelick, P. B. (2004). "Risk factors for vascular dementia and Alzheimer disease." Stroke 35(11 Suppl 1): 2620-2622. Grathwohl, S. A., R. E. Kalin, et al. (2009). "Formation and maintenance of Alzheimer's disease beta-amyloid plaques in the absence of microglia." Nat Neurosci 12(11): 1361-1363. Grimm, M. O., I. Tomic, et al. (2002). "Potential external source of A beta in biological samples." Nat Cell Biol 4(7): E164-165; author reply E165-166. Group, A. R., C. G. Lyketsos, et al. (2007). "Naproxen and celecoxib do not prevent AD in early results from a randomized controlled trial." Neurology 68(21): 1800-1808. Haass, C., A. Y. Hung, et al. (1993). "beta-Amyloid peptide and a 3-kDa fragment are derived by distinct cellular mechanisms." J Biol Chem 268(5): 3021-3024. Haass, C., M. G. Schlossmacher, et al. (1992). "Amyloid beta-peptide is produced by cultured cells during normal metabolism." Nature 359(6393): 322-325.

178

Hamburger, S. A. and P. B. McCay (1990). "Spin trapping of ibuprofen radicals: evidence that ibuprofen is a hydroxyl radical scavenger." Free Radic Res Commun 9(3-6): 337-342. Hanisch, U. K. and H. Kettenmann (2007). "Microglia: active sensor and versatile effector cells in the normal and pathologic brain." Nat Neurosci 10(11): 1387-1394. Hardy, J. A. and G. A. Higgins (1992). "Alzheimer's disease: the amyloid cascade hypothesis." Science 256(5054): 184-185. Harrington, C., S. Sawchak, et al. (2011). "Rosiglitazone does not improve cognition or global function when used as adjunctive therapy to AChE inhibitors in mild-to-moderate Alzheimer's disease: two phase 3 studies." Curr Alzheimer Res 8(5): 592-606. Haynes, S. E., G. Hollopeter, et al. (2006). "The P2Y12 receptor regulates microglial activation by extracellular nucleotides." Nat Neurosci 9(12): 1512-1519. Hebert, L. E., P. A. Scherr, et al. (2003). "Alzheimer disease in the US population: prevalence estimates using the 2000 census." Arch Neurol 60(8): 1119-1122. Hefendehl, J. K., B. M. Wegenast-Braun, et al. (2011). "Long-term in vivo imaging of beta- amyloid plaque appearance and growth in a mouse model of cerebral beta- amyloidosis." J Neurosci 31(2): 624-629. Hegele, R. A. (2005). "Retinoid X receptor heterodimers in the metabolic syndrome." N Engl J Med 353(19): 2088. Helzner, E. P., N. Scarmeas, et al. (2008). "Survival in Alzheimer disease: a multiethnic, population-based study of incident cases." Neurology 71(19): 1489-1495. Heneka, M. T., M. Sastre, et al. (2005). "Acute treatment with the PPARgamma agonist pioglitazone and ibuprofen reduces glial inflammation and Abeta1-42 levels in APPV717I transgenic mice." Brain 128(Pt 6): 1442-1453. Herber, D. L., L. M. Roth, et al. (2004). "Time-dependent reduction in Abeta levels after intracranial LPS administration in APP transgenic mice." Exp Neurol 190(1): 245-253. Hesse, M., M. Modolell, et al. (2001). "Differential regulation of nitric oxide synthase-2 and arginase-1 by type 1/type 2 cytokines in vivo: granulomatous pathology is shaped by the pattern of L-arginine metabolism." J Immunol 167(11): 6533-6544. Hirsch-Reinshagen, V., L. F. Maia, et al. (2005). "The absence of ABCA1 decreases soluble ApoE levels but does not diminish amyloid deposition in two murine models of Alzheimer disease." J Biol Chem 280(52): 43243-43256. Hirsch-Reinshagen, V., S. Zhou, et al. (2004). "Deficiency of ABCA1 impairs apolipoprotein E metabolism in brain." J Biol Chem 279(39): 41197-41207. Hoi, C. P., Y. P. Ho, et al. (2010). "Neuroprotective effect of honokiol and magnolol, compounds from Magnolia officinalis, on beta-amyloid-induced toxicity in PC12 cells." Phytother Res 24(10): 1538-1542. Holcomb, I. N., R. C. Kabakoff, et al. (2000). "FIZZ1, a novel cysteine-rich secreted protein associated with pulmonary inflammation, defines a new gene family." EMBO J 19(15): 4046-4055. Holtzman, D. M., K. R. Bales, et al. (1999). "Expression of human apolipoprotein E reduces amyloid-beta deposition in a mouse model of Alzheimer's disease." J Clin Invest 103(6): R15-R21. Hong, S., O. Quintero-Monzon, et al. (2011). "Dynamic analysis of amyloid beta-protein in behaving mice reveals opposing changes in ISF versus parenchymal Abeta during age- related plaque formation." J Neurosci 31(44): 15861-15869. Howard, J. D., J. Plailly, et al. (2009). "Odor quality coding and categorization in human posterior piriform cortex." Nat Neurosci 12(7): 932-938.

179

Howell, S. R., M. A. Shirley, et al. (1998). "Effects of retinoid treatment of rats on hepatic microsomal metabolism and cytochromes P450. Correlation between retinoic acid receptor/retinoid x receptor selectivity and effects on metabolic enzymes." Drug Metab Dispos 26(3): 234-239. Hoyer, S. (2002). "The aging brain. Changes in the neuronal insulin/insulin receptor signal transduction cascade trigger late-onset sporadic Alzheimer disease (SAD). A mini- review." J Neural Transm 109(7-8): 991-1002. Hsiao, K., P. Chapman, et al. (1996). "Correlative memory deficits, Abeta elevation, and amyloid plaques in transgenic mice." Science 274(5284): 99-102. Huang, J. K., A. A. Jarjour, et al. (2011). "Retinoid X receptor gamma signaling accelerates CNS remyelination." Nat Neurosci 14(1): 45-53. in t' Veld, B. A., A. Ruitenberg, et al. (2001). "Nonsteroidal antiinflammatory drugs and the risk of Alzheimer's disease." N Engl J Med 345(21): 1515-1521. Itagaki, S., P. L. McGeer, et al. (1989). "Relationship of microglia and astrocytes to amyloid deposits of Alzheimer disease." J Neuroimmunol 24(3): 173-182. Iwata, N., H. Mizukami, et al. (2004). "Presynaptic localization of neprilysin contributes to efficient clearance of amyloid-beta peptide in mouse brain." J Neurosci 24(4): 991-998. Jankowsky, J. L., H. H. Slunt, et al. (2005). "Persistent amyloidosis following suppression of Abeta production in a transgenic model of Alzheimer disease." PLoS Med 2(12): e355. Jantzen, P. T., K. E. Connor, et al. (2002). "Microglial activation and beta -amyloid deposit reduction caused by a nitric oxide-releasing nonsteroidal anti-inflammatory drug in amyloid precursor protein plus presenilin-1 transgenic mice." J Neurosci 22(6): 2246- 2254. Jarvis, C. I., M. B. Goncalves, et al. (2010). "Retinoic acid receptor-alpha signalling antagonizes both intracellular and extracellular amyloid-beta production and prevents neuronal cell death caused by amyloid-beta." Eur J Neurosci 32(8): 1246-1255. Jiang, Q., C. Y. Lee, et al. (2008). "ApoE promotes the proteolytic degradation of Abeta." Neuron 58(5): 681-693. Jimenez, S., D. Baglietto-Vargas, et al. (2008). "Inflammatory response in the hippocampus of PS1M146L/APP751SL mouse model of Alzheimer's disease: age-dependent switch in the microglial phenotype from alternative to classic." J Neurosci 28(45): 11650-11661. Jones, L., P. A. Holmans, et al. (2010). "Genetic evidence implicates the immune system and cholesterol metabolism in the aetiology of Alzheimer's disease." PLoS One 5(11): e13950. Joseph, S. B., A. Castrillo, et al. (2003). "Reciprocal regulation of inflammation and lipid metabolism by liver X receptors." Nat Med 9(2): 213-219. Juedes, A. E. and N. H. Ruddle (2001). "Resident and infiltrating central nervous system APCs regulate the emergence and resolution of experimental autoimmune encephalomyelitis." J Immunol 166(8): 5168-5175. Jung, C. G., H. Horike, et al. (2010). "Honokiol increases ABCA1 expression level by activating retinoid X receptor beta." Biol Pharm Bull 33(7): 1105-1111. Kalaany, N. Y. and D. J. Mangelsdorf (2006). "LXRS and FXR: the yin and yang of cholesterol and fat metabolism." Annu Rev Physiol 68: 159-191. Kane, M. A. (2012). "Analysis, occurrence, and function of 9-cis-retinoic acid." Biochim Biophys Acta 1821(1): 10-20. Kang, J., H. G. Lemaire, et al. (1987). "The precursor of Alzheimer's disease amyloid A4 protein resembles a cell-surface receptor." Nature 325(6106): 733-736.

180

Katzman, R. (1994). "Apolipoprotein E and Alzheimer's disease." Curr Opin Neurobiol 4(5): 703- 707. Kim, J., J. M. Basak, et al. (2009). "The role of apolipoprotein E in Alzheimer's disease." Neuron 63(3): 287-303. Kiyota, T., S. Okuyama, et al. (2010). "CNS expression of anti-inflammatory cytokine interleukin-4 attenuates Alzheimer's disease-like pathogenesis in APP+PS1 bigenic mice." FASEB J 24(8): 3093-3102. Klunk, W. E., B. J. Bacskai, et al. (2002). "Imaging Abeta plaques in living transgenic mice with multiphoton microscopy and methoxy-X04, a systemically administered Congo red derivative." J Neuropathol Exp Neurol 61(9): 797-805. Klunk, W. E., H. Engler, et al. (2004). "Imaging brain amyloid in Alzheimer's disease with Pittsburgh Compound-B." Ann Neurol 55(3): 306-319. Koenigsknecht-Talboo, J. and G. E. Landreth (2005). "Microglial phagocytosis induced by fibrillar beta-amyloid and IgGs are differentially regulated by proinflammatory cytokines." J Neurosci 25(36): 8240-8249. Koenigsknecht, J. and G. Landreth (2004). "Microglial phagocytosis of fibrillar beta-amyloid through a beta1 integrin-dependent mechanism." J Neurosci 24(44): 9838-9846. Koistinaho, M., S. Lin, et al. (2004). "Apolipoprotein E promotes astrocyte colocalization and degradation of deposited amyloid-beta peptides." Nat Med 10(7): 719-726. Koldamova, R., M. Staufenbiel, et al. (2005). "Lack of ABCA1 considerably decreases brain ApoE level and increases amyloid deposition in APP23 mice." J Biol Chem 280(52): 43224- 43235. Koldamova, R. P., I. M. Lefterov, et al. (2003). "22R-hydroxycholesterol and 9-cis-retinoic acid induce ATP-binding cassette transporter A1 expression and cholesterol efflux in brain cells and decrease amyloid beta secretion." J Biol Chem 278(15): 13244-13256. Kopell, N., G. B. Ermentrout, et al. (2000). "Gamma rhythms and beta rhythms have different synchronization properties." Proceedings of the National Academy of Sciences of the United States of America 97(4): 1867-1872. Kosik, K. S., C. L. Joachim, et al. (1986). "Microtubule-associated protein tau (tau) is a major antigenic component of paired helical filaments in Alzheimer disease." Proc Natl Acad Sci U S A 83(11): 4044-4048. Kotani, H., H. Tanabe, et al. (2012). "A naturally occurring rexinoid, honokiol, can serve as a regulator of various retinoid x receptor heterodimers." Biol Pharm Bull 35(1): 1-9. Kotani, H., H. Tanabe, et al. (2010). "Identification of a naturally occurring rexinoid, honokiol, that activates the retinoid X receptor." J Nat Prod 73(8): 1332-1336. Kotilinek, L. A., M. A. Westerman, et al. (2008). "Cyclooxygenase-2 inhibition improves amyloid- beta-mediated suppression of memory and synaptic plasticity." Brain 131(Pt 3): 651- 664. Krausgruber, T., K. Blazek, et al. (2011). "IRF5 promotes inflammatory macrophage polarization and TH1-TH17 responses." Nat Immunol 12(3): 231-238. Krzyzosiak, A., M. Szyszka-Niagolov, et al. (2010). "Retinoid x receptor gamma control of affective behaviors involves dopaminergic signaling in mice." Neuron 66(6): 908-920. Kukull, W. A., R. Higdon, et al. (2002). "Dementia and Alzheimer disease incidence: a prospective cohort study." Arch Neurol 59(11): 1737-1746. Kulnane, L. S. and B. T. Lamb (2001). "Neuropathological characterization of mutant amyloid precursor protein yeast artificial chromosome transgenic mice." Neurobiol Dis 8(6): 982- 992.

181

Kurz, A. and R. Perneczky (2011). "Amyloid clearance as a treatment target against Alzheimer's disease." J Alzheimers Dis 24 Suppl 2: 61-73. LaDu, M. J., T. M. Pederson, et al. (1995). "Purification of apolipoprotein E attenuates isoform- specific binding to beta-amyloid." J Biol Chem 270(16): 9039-9042. Lalloyer, F., C. Fievet, et al. (2006). "The RXR agonist bexarotene improves cholesterol homeostasis and inhibits atherosclerosis progression in a mouse model of mixed dyslipidemia." Arterioscler Thromb Vasc Biol 26(12): 2731-2737. Lamb, B. T., K. A. Bardel, et al. (1999). "Amyloid production and deposition in mutant amyloid precursor protein and presenilin-1 yeast artificial chromosome transgenic mice." Nat Neurosci 2(8): 695-697. Lambeth, J. D. (2004). "NOX enzymes and the biology of reactive oxygen." Nat Rev Immunol 4(3): 181-189. Lammich, S., E. Kojro, et al. (1999). "Constitutive and regulated alpha-secretase cleavage of Alzheimer's amyloid precursor protein by a disintegrin metalloprotease." Proc Natl Acad Sci U S A 96(7): 3922-3927. Landreth, G., Q. Jiang, et al. (2008). "PPARgamma agonists as therapeutics for the treatment of Alzheimer's disease." Neurotherapeutics 5(3): 481-489. Landreth, G. E. and E. G. Reed-Geaghan (2009). "Toll-like receptors in Alzheimer's disease." Curr Top Microbiol Immunol 336: 137-153. Lawson, L. J., V. H. Perry, et al. (1990). "Heterogeneity in the distribution and morphology of microglia in the normal adult mouse brain." Neuroscience 39(1): 151-170. Lawson, L. J., V. H. Perry, et al. (1992). "Turnover of resident microglia in the normal adult mouse brain." Neuroscience 48(2): 405-415. Lebson, L., K. Nash, et al. (2010). "Trafficking CD11b-positive blood cells deliver therapeutic genes to the brain of amyloid-depositing transgenic mice." J Neurosci 30(29): 9651- 9658. Lee, J. H., S. M. Park, et al. (2009). "Differential SUMOylation of LXRalpha and LXRbeta mediates transrepression of STAT1 inflammatory signaling in IFN-gamma-stimulated brain astrocytes." Mol Cell 35(6): 806-817. Lee, Y. J., S. B. Han, et al. (2010). "Inflammation and Alzheimer's disease." Arch Pharm Res 33(10): 1539-1556. Lefebvre, P., Y. Benomar, et al. (2010). "Retinoid X receptors: common heterodimerization partners with distinct functions." Trends Endocrinol Metab. Lehman, E. J., L. S. Kulnane, et al. (2003). "Genetic background regulates beta-amyloid precursor protein processing and beta-amyloid deposition in the mouse." Hum Mol Genet 12(22): 2949-2956. Lehmann, J. M., J. M. Lenhard, et al. (1997). "Peroxisome proliferator-activated receptors alpha and gamma are activated by indomethacin and other non-steroidal anti-inflammatory drugs." J Biol Chem 272(6): 3406-3410. Lehrke, M. and M. A. Lazar (2005). "The many faces of PPARgamma." Cell 123(6): 993-999. Leissring, M. A., W. Farris, et al. (2003). "Enhanced proteolysis of beta-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death." Neuron 40(6): 1087-1093. Lemere, C. A. and E. Masliah (2010). "Can Alzheimer disease be prevented by amyloid-beta immunotherapy?" Nat Rev Neurol 6(2): 108-119. Lerner, V., C. Miodownik, et al. (2008). "Bexarotene as add-on to antipsychotic treatment in schizophrenia patients: a pilot open-label trial." Clin Neuropharmacol 31(1): 25-33.

182

Liang, X., Q. Wang, et al. (2005). "Deletion of the prostaglandin E2 EP2 receptor reduces oxidative damage and amyloid burden in a model of Alzheimer's disease." J Neurosci 25(44): 10180-10187. Liang, Y., S. Lin, et al. (2004). "A liver X receptor and retinoid X receptor heterodimer mediates apolipoprotein E expression, secretion and cholesterol homeostasis in astrocytes." J Neurochem 88(3): 623-634. Libreros, S., R. Garcia-Areas, et al. (2011). "Induction of proinflammatory mediators by CHI3L1 is reduced by chitin treatment: Decreased tumor metastasis in a breast cancer model." Int J Cancer. Lim, G. P., T. Chu, et al. (2001). "The curry spice curcumin reduces oxidative damage and amyloid pathology in an Alzheimer transgenic mouse." J Neurosci 21(21): 8370-8377. Lim, G. P., F. Yang, et al. (2000). "Ibuprofen suppresses plaque pathology and inflammation in a mouse model for Alzheimer's disease." J Neurosci 20(15): 5709-5714. Liu, R. M., T. van Groen, et al. (2011). "Knockout of plasminogen activator inhibitor 1 gene reduces amyloid beta peptide burden in a mouse model of Alzheimer's disease." Neurobiol Aging 32(6): 1079-1089. Liu, Y., S. Walter, et al. (2005). "LPS receptor (CD14): a receptor for phagocytosis of Alzheimer's amyloid peptide." Brain 128(Pt 8): 1778-1789. Lleo, A., O. Berezovska, et al. (2004). "Nonsteroidal anti-inflammatory drugs lower Abeta42 and change presenilin 1 conformation." Nat Med 10(10): 1065-1066. Lorenzo, A. and B. A. Yankner (1994). "Beta-amyloid neurotoxicity requires fibril formation and is inhibited by congo red." Proc Natl Acad Sci U S A 91(25): 12243-12247. Lyketsos, C. G., J. C. Breitner, et al. (2007). "Naproxen and celecoxib do not prevent AD in early results from a randomized controlled trial." Neurology 68(21): 1800-1808. Mackenzie, I. R., C. Hao, et al. (1995). "Role of microglia in senile plaque formation." Neurobiol Aging 16(5): 797-804. Majumdar, A., H. Chung, et al. (2008). "Degradation of fibrillar forms of Alzheimer's amyloid beta-peptide by macrophages." Neurobiol Aging 29(5): 707-715. Malm, T. M., M. Koistinaho, et al. (2005). "Bone-marrow-derived cells contribute to the recruitment of microglial cells in response to beta-amyloid deposition in APP/PS1 double transgenic Alzheimer mice." Neurobiol Dis 18(1): 134-142. Mandrekar, S., Q. Jiang, et al. (2009). "Microglia mediate the clearance of soluble Abeta through fluid phase macropinocytosis." J Neurosci 29(13): 4252-4262. Mantovani, A., A. Sica, et al. (2004). "The chemokine system in diverse forms of macrophage activation and polarization." Trends Immunol 25(12): 677-686. Mark, R. J., M. A. Lovell, et al. (1997). "A role for 4-hydroxynonenal, an aldehydic product of lipid peroxidation, in disruption of ion homeostasis and neuronal death induced by amyloid beta-peptide." J Neurochem 68(1): 255-264. Markesbery, W. R. (1997). "Oxidative stress hypothesis in Alzheimer's disease." Free Radic Biol Med 23(1): 134-147. Marr, R. A., E. Rockenstein, et al. (2003). "Neprilysin gene transfer reduces human amyloid pathology in transgenic mice." J Neurosci 23(6): 1992-1996. Martinez, F. O., L. Helming, et al. (2009). "Alternative activation of macrophages: an immunologic functional perspective." Annu Rev Immunol 27: 451-483. Mascrez, B., N. B. Ghyselinck, et al. (2009). "A transcriptionally silent RXRalpha supports early embryonic morphogenesis and heart development." Proc Natl Acad Sci U S A 106(11): 4272-4277.

183

Masliah, E., M. Mallory, et al. (1991). "Immunoreactivity of CD45, a protein phosphotyrosine phosphatase, in Alzheimer's disease." Acta Neuropathol 83(1): 12-20. Massengale, M., A. J. Wagers, et al. (2005). "Hematopoietic cells maintain hematopoietic fates upon entering the brain." J Exp Med 201(10): 1579-1589. Mawuenyega, K. G., W. Sigurdson, et al. (2010). "Decreased Clearance of CNS {beta}-Amyloid in Alzheimer's Disease." Science 330: 1774-1776. Mawuenyega, K. G., W. Sigurdson, et al. (2010). "Decreased clearance of CNS beta-amyloid in Alzheimer's disease." Science 330(6012): 1774. McDonald, D. R., M. E. Bamberger, et al. (1998). "beta-Amyloid fibrils activate parallel mitogen- activated protein kinase pathways in microglia and THP1 monocytes." J Neurosci 18(12): 4451-4460. McDonald, D. R., K. R. Brunden, et al. (1997). "Amyloid fibrils activate tyrosine kinase-dependent signaling and superoxide production in microglia." J Neurosci 17(7): 2284-2294. McGeer, P. L. and E. G. McGeer (2001). "Polymorphisms in inflammatory genes and the risk of Alzheimer disease." Arch Neurol 58(11): 1790-1792. McKee, A. C., I. Carreras, et al. (2008). "Ibuprofen reduces Abeta, hyperphosphorylated tau and memory deficits in Alzheimer mice." Brain Res 1207: 225-236. McKenna, N. J. and B. W. O'Malley (2002). "Minireview: nuclear receptor coactivators--an update." Endocrinology 143(7): 2461-2465. Meyer-Luehmann, M., T. L. Spires-Jones, et al. (2008). "Rapid appearance and local toxicity of amyloid-beta plaques in a mouse model of Alzheimer's disease." Nature 451(7179): 720- 724. Michelucci, A., T. Heurtaux, et al. (2009). "Characterization of the microglial phenotype under specific pro-inflammatory and anti-inflammatory conditions: Effects of oligomeric and fibrillar amyloid-beta." J Neuroimmunol 210(1-2): 3-12. Mildner, A., B. Schlevogt, et al. (2011). "Distinct and non-redundant roles of microglia and myeloid subsets in mouse models of Alzheimer's disease." J Neurosci 31(31): 11159- 11171. Mildner, A., H. Schmidt, et al. (2007). "Microglia in the adult brain arise from Ly-6ChiCCR2+ monocytes only under defined host conditions." Nat Neurosci 10(12): 1544-1553. Miller, B. C., E. A. Eckman, et al. (2003). "Amyloid-beta peptide levels in brain are inversely correlated with insulysin activity levels in vivo." Proc Natl Acad Sci U S A 100(10): 6221- 6226. Mohmmad Abdul, H., R. Sultana, et al. (2006). "Mutations in amyloid precursor protein and presenilin-1 genes increase the basal oxidative stress in murine neuronal cells and lead to increased sensitivity to oxidative stress mediated by amyloid beta-peptide (1-42), HO and kainic acid: implications for Alzheimer's disease." J Neurochem 96(5): 1322-1335. Moreno, S., S. Farioli-Vecchioli, et al. (2004). "Immunolocalization of peroxisome proliferator- activated receptors and retinoid X receptors in the adult rat CNS." Neuroscience 123(1): 131-145. Morgan, D., M. N. Gordon, et al. (2005). "Dynamic complexity of the microglial activation response in transgenic models of amyloid deposition: implications for Alzheimer therapeutics." J Neuropathol Exp Neurol 64(9): 743-753. Mosser, D. M. (2003). "The many faces of macrophage activation." J Leukoc Biol 73(2): 209-212. Mosser, D. M. and J. P. Edwards (2008). "Exploring the full spectrum of macrophage activation." Nat Rev Immunol 8(12): 958-969.

184

Mucke, L., E. Masliah, et al. (2000). "High-level neuronal expression of abeta 1-42 in wild-type human amyloid protein precursor transgenic mice: synaptotoxicity without plaque formation." J Neurosci 20(11): 4050-4058. Mukherjee, R., J. Strasser, et al. (1998). "RXR agonists activate PPARalpha-inducible genes, lower triglycerides, and raise HDL levels in vivo." Arterioscler Thromb Vasc Biol 18(2): 272-276. Mukundan, L., J. I. Odegaard, et al. (2009). "PPAR-delta senses and orchestrates clearance of apoptotic cells to promote tolerance." Nat Med 15(11): 1266-1272. Mullan, M., F. Crawford, et al. (1992). "A pathogenic mutation for probable Alzheimer's disease in the APP gene at the N-terminus of beta-amyloid." Nat Genet 1(5): 345-347. Murphy, C. (1999). "Loss of olfactory function in dementing disease." Physiol Behav 66(2): 177- 182. Nimmerjahn, A., F. Kirchhoff, et al. (2005). "Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo." Science 308(5726): 1314-1318. Nishimaki-Mogami, T., N. Tamehiro, et al. (2008). "The RXR agonists PA024 and HX630 have different abilities to activate LXR/RXR and to induce ABCA1 expression in macrophage cell lines." Biochem Pharmacol 76(8): 1006-1013. Nomoto, M., Y. Takeda, et al. (2012). "Dysfunction of the RAR/RXR signaling pathway in the forebrain impairs hippocampal memory and synaptic plasticity." Mol Brain 5(1): 8. Nunez, V., D. Alameda, et al. (2010). "Retinoid X receptor alpha controls innate inflammatory responses through the up-regulation of chemokine expression." Proc Natl Acad Sci U S A 107(23): 10626-10631. Nunomura, A., G. Perry, et al. (1999). "RNA oxidation is a prominent feature of vulnerable neurons in Alzheimer's disease." J Neurosci 19(6): 1959-1964. Odegaard, J. I., R. R. Ricardo-Gonzalez, et al. (2007). "Macrophage-specific PPARgamma controls alternative activation and improves insulin resistance." Nature 447(7148): 1116-1120. Olichney, J. M., M. N. Sabbagh, et al. (1997). "The impact of apolipoprotein E4 on cause of death in Alzheimer's disease." Neurology 49(1): 76-81. Ono, K., Y. Yoshiike, et al. (2004). "Vitamin A exhibits potent antiamyloidogenic and fibril- destabilizing effects in vitro." Exp Neurol 189(2): 380-392. Osenkowski, P., W. Ye, et al. (2008). "Direct and potent regulation of gamma-secretase by its lipid microenvironment." J Biol Chem 283(33): 22529-22540. Palop, J. J. and L. Mucke (2010). "Amyloid-beta-induced neuronal dysfunction in Alzheimer's disease: from synapses toward neural networks." Nat Neurosci 13(7): 812-818. Paresce, D. M., H. Chung, et al. (1997). "Slow degradation of aggregates of the Alzheimer's disease amyloid beta-protein by microglial cells." J Biol Chem 272(46): 29390-29397. Park, L., J. Anrather, et al. (2005). "NADPH-oxidase-derived reactive oxygen species mediate the cerebrovascular dysfunction induced by the amyloid beta peptide." J Neurosci 25(7): 1769-1777. Park, L., P. Zhou, et al. (2008). "Nox2-derived radicals contribute to neurovascular and behavioral dysfunction in mice overexpressing the amyloid precursor protein." Proc Natl Acad Sci U S A 105(4): 1347-1352. Pascual, G., A. L. Fong, et al. (2005). "A SUMOylation-dependent pathway mediates transrepression of inflammatory response genes by PPAR-gamma." Nature 437(7059): 759-763. Pascual, G. and C. K. Glass (2006). "Nuclear receptors versus inflammation: mechanisms of transrepression." Trends Endocrinol Metab 17(8): 321-327.

185

Paylor, R., L. Baskall-Baldini, et al. (1996). "Developmental differences in place-learning performance between C57BL/6 and DBA/2 mice parallel the ontogeny of hippocampal protein kinase C." Behav Neurosci 110(6): 1415-1425. Pedersen, W. A. and E. R. Flynn (2004). "Insulin resistance contributes to aberrant stress responses in the Tg2576 mouse model of Alzheimer's disease." Neurobiol Dis 17(3): 500- 506. Pedersen, W. A., P. J. McMillan, et al. (2006). "Rosiglitazone attenuates learning and memory deficits in Tg2576 Alzheimer mice." Exp Neurol 199(2): 265-273. Perez, E., W. Bourguet, et al. (2012). "Modulation of RXR function through ligand design." Biochim Biophys Acta 1821(1): 57-69. Picard, F. and J. Auwerx (2002). "PPAR(gamma) and glucose homeostasis." Annu Rev Nutr 22: 167-197. Podlisny, M. B., G. Lee, et al. (1987). "Gene dosage of the amyloid beta precursor protein in Alzheimer's disease." Science 238(4827): 669-671. Pratico, D. (2008). "Evidence of oxidative stress in Alzheimer's disease brain and antioxidant therapy: lights and shadows." Ann N Y Acad Sci 1147: 70-78. Pratico, D. and S. Sung (2004). "Lipid peroxidation and oxidative imbalance: early functional events in Alzheimer's disease." J Alzheimers Dis 6(2): 171-175. Pratico, D., K. Uryu, et al. (2001). "Increased lipid peroxidation precedes amyloid plaque formation in an animal model of Alzheimer amyloidosis." J Neurosci 21(12): 4183-4187. Priller, C., T. Bauer, et al. (2006). "Synapse formation and function is modulated by the amyloid precursor protein." J Neurosci 26(27): 7212-7221. Prinz, M. and J. Priller (2010). "Tickets to the brain: role of CCR2 and CX3CR1 in myeloid cell entry in the CNS." J Neuroimmunol 224(1-2): 80-84. Qaseem, A., V. Snow, et al. (2008). "Current pharmacologic treatment of dementia: a clinical practice guideline from the American College of Physicians and the American Academy of Family Physicians." Ann Intern Med 148(5): 370-378. Qin, B., L. Cartier, et al. (2005). "A key role for the microglial NADPH oxidase in APP-dependent killing of neurons." Neurobiol Aging [Epub ahead of print]. Qin, B., L. Cartier, et al. (2006). "A key role for the microglial NADPH oxidase in APP-dependent killing of neurons." Neurobiol Aging 27(11): 1577-1587. Qiu, W. Q. and M. F. Folstein (2006). "Insulin, insulin-degrading enzyme and amyloid-beta peptide in Alzheimer's disease: review and hypothesis." Neurobiol Aging 27(2): 190-198. Qiu, W. Q., D. M. Walsh, et al. (1998). "Insulin-degrading enzyme regulates extracellular levels of amyloid beta-protein by degradation." J Biol Chem 273(49): 32730-32738. Querfurth, H. W. and F. M. LaFerla (2010). "Alzheimer's disease." N Engl J Med 362(4): 329-344. Radde, R., T. Bolmont, et al. (2006). "Abeta42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology." EMBO Rep 7(9): 940-946. Ransohoff, R. M. (2007). "Microgliosis: the questions shape the answers." Nat Neurosci 10(12): 1507-1509. Reed-Geaghan, E. G., Q. W. Reed, et al. (2010). "Deletion of CD14 attenuates Alzheimer's disease pathology by influencing the brain's inflammatory milieu." J Neurosci 30(46): 15369-15373. Reed-Geaghan, E. G., J. C. Savage, et al. (2009). "CD14 and toll-like receptors 2 and 4 are required for fibrillar A{beta}-stimulated microglial activation." J Neurosci 29(38): 11982- 11992. Reifen, R. (2002). "Vitamin A as an anti-inflammatory agent." Proc Nutr Soc 61(3): 397-400.

186

Reiman, E. M., J. B. Langbaum, et al. (2011). "Alzheimer's Prevention Initiative: a plan to accelerate the evaluation of presymptomatic treatments." J Alzheimers Dis 26 Suppl 3: 321-329. Reines, S. A., G. A. Block, et al. (2004). "Rofecoxib: no effect on Alzheimer's disease in a 1-year, randomized, blinded, controlled study." Neurology 62(1): 66-71. Renno, T., M. Krakowski, et al. (1995). "TNF-alpha expression by resident microglia and infiltrating leukocytes in the central nervous system of mice with experimental allergic encephalomyelitis. Regulation by Th1 cytokines." J Immunol 154(2): 944-953. Riddell, D. R., H. Zhou, et al. (2008). "Impact of apolipoprotein E (ApoE) polymorphism on brain ApoE levels." J Neurosci 28(45): 11445-11453. Riddell, D. R., H. Zhou, et al. (2007). "The LXR agonist TO901317 selectively lowers hippocampal Abeta42 and improves memory in the Tg2576 mouse model of Alzheimer's disease." Mol Cell Neurosci 34(4): 621-628. Rinne, J. O., D. J. Brooks, et al. (2010). "11C-PiB PET assessment of change in fibrillar amyloid- beta load in patients with Alzheimer's disease treated with bapineuzumab: a phase 2, double-blind, placebo-controlled, ascending-dose study." Lancet Neurol 9(4): 363-372. Rizvi, N. A., J. L. Marshall, et al. (1999). "A Phase I study of LGD1069 in adults with advanced cancer." Clin Cancer Res 5(7): 1658-1664. Rolyan, H., A. C. Feike, et al. (2011). "Amyloid-beta protein modulates the perivascular clearance of neuronal apolipoprotein E in mouse models of Alzheimer's disease." J Neural Transm 118(5): 699-712. Roses, A. D. and A. M. Saunders (1994). "APOE is a major susceptibility gene for Alzheimer's disease." Curr Opin Biotechnol 5(6): 663-667. Saijo, K. and C. K. Glass (2011). "Microglial cell origin and phenotypes in health and disease." Nat Rev Immunol 11(11): 775-787. Sasaki, A., M. Shoji, et al. (2002). "Amyloid cored plaques in Tg2576 transgenic mice are characterized by giant plaques, slightly activated microglia, and the lack of paired helical filament-typed, dystrophic neurites." Virchows Arch 441(4): 358-367. Sayre, L. M., D. A. Zelasko, et al. (1997). "4-Hydroxynonenal-derived advanced lipid peroxidation end products are increased in Alzheimer's disease." J Neurochem 68(5): 2092-2097. Schmittgen, T. D. and K. J. Livak (2008). "Analyzing real-time PCR data by the comparative C(T) method." Nat Protoc 3(6): 1101-1108. Schneider, J. A., Z. Arvanitakis, et al. (2009). "The neuropathology of probable Alzheimer disease and mild cognitive impairment." Ann Neurol 66(2): 200-208. Selkoe, D. J. (2000). "Toward a comprehensive theory for Alzheimer's disease. Hypothesis: Alzheimer's disease is caused by the cerebral accumulation and cytotoxicity of amyloid beta-protein." Ann N Y Acad Sci 924: 17-25. Selkoe, D. J. (2001). "Clearing the brain's amyloid cobwebs." Neuron 32(2): 177-180. Selkoe, D. J. (2008). "Soluble oligomers of the amyloid beta-protein impair synaptic plasticity and behavior." Behav Brain Res 192(1): 106-113. Selkoe, D. J. (2011). "Alzheimer's disease." Cold Spring Harb Perspect Biol 3(7). Seo, J. B., H. M. Moon, et al. (2004). "Activated liver X receptors stimulate adipocyte differentiation through induction of peroxisome proliferator-activated receptor gamma expression." Mol Cell Biol 24(8): 3430-3444. Serghides, L. and K. C. Kain (2001). "Peroxisome proliferator-activated receptor gamma-retinoid X receptor agonists increase CD36-dependent phagocytosis of Plasmodium falciparum- parasitized erythrocytes and decrease malaria-induced TNF-alpha secretion by monocytes/macrophages." J Immunol 166(11): 6742-6748.

187

Seshadri, S., P. A. Wolf, et al. (1997). "Lifetime risk of dementia and Alzheimer's disease. The impact of mortality on risk estimates in the Framingham Study." Neurology 49(6): 1498- 1504. Shankar, G. M., S. Li, et al. (2008). "Amyloid-beta protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory." Nat Med 14(8): 837-842. Shibata, M., S. Yamada, et al. (2000). "Clearance of Alzheimer's amyloid-ss(1-40) peptide from brain by LDL receptor-related protein-1 at the blood-brain barrier." J Clin Invest 106(12): 1489-1499. Shimohama, S., H. Tanino, et al. (2000). "Activation of NADPH oxidase in Alzheimer's disease brains." Biochem Biophys Res Commun 273(1): 5-9. Shirotani, K., S. Tsubuki, et al. (2001). "Neprilysin degrades both amyloid beta peptides 1-40 and 1-42 most rapidly and efficiently among thiorphan- and phosphoramidon-sensitive endopeptidases." J Biol Chem 276(24): 21895-21901. Simard, A. R. and S. Rivest (2004). "Bone marrow stem cells have the ability to populate the entire central nervous system into fully differentiated parenchymal microglia." FASEB J 18(9): 998-1000. Simard, A. R., D. Soulet, et al. (2006). "Bone marrow-derived microglia play a critical role in restricting senile plaque formation in Alzheimer's disease." Neuron 49(4): 489-502. Simons, M., P. Keller, et al. (1998). "Cholesterol depletion inhibits the generation of beta- amyloid in hippocampal neurons." Proc Natl Acad Sci U S A 95(11): 6460-6464. Sisodia, S. S. and D. L. Price (1995). "Role of the beta-amyloid protein in Alzheimer's disease." FASEB J 9(5): 366-370. Sisodia, S. S. and P. H. St George-Hyslop (2002). "gamma-Secretase, Notch, Abeta and Alzheimer's disease: where do the presenilins fit in?" Nat Rev Neurosci 3(4): 281-290. Smith, M. A., P. L. Richey Harris, et al. (1997). "Widespread peroxynitrite-mediated damage in Alzheimer's disease." J Neurosci 17(8): 2653-2657. Sonnen, J. A., E. B. Larson, et al. (2009). "Free radical damage to cerebral cortex in Alzheimer's disease, microvascular brain injury, and smoking." Ann Neurol 65(2): 226-229. Sperling, R. A., P. S. Aisen, et al. (2011). "Toward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging-Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease." Alzheimers Dement 7(3): 280-292. Stadtman, E. R. and R. L. Levine (2000). "Protein oxidation." Ann N Y Acad Sci 899: 191-208. Stahl, S. M. (2000). "The new cholinesterase inhibitors for Alzheimer's disease, Part 1: their similarities are different." J Clin Psychiatry 61(10): 710-711. Stalder, A. K., F. Ermini, et al. (2005). "Invasion of hematopoietic cells into the brain of amyloid precursor protein transgenic mice." J Neurosci 25(48): 11125-11132. Stalder, M., A. Phinney, et al. (1999). "Association of microglia with amyloid plaques in brains of APP23 transgenic mice." Am J Pathol 154(6): 1673-1684. Standridge, J. B. (2006). "Vicious cycles within the neuropathophysiologic mechanisms of Alzheimer's disease." Curr Alzheimer Res 3(2): 95-108. Stewart, W. F., C. Kawas, et al. (1997). "Risk of Alzheimer's disease and duration of NSAID use." Neurology 48(3): 626-632. Strittmatter, W. J., A. M. Saunders, et al. (1993). "Apolipoprotein E: high-avidity binding to beta- amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease." Proc Natl Acad Sci U S A 90(5): 1977-1981. Strum, J. C., R. Shehee, et al. (2007). "Rosiglitazone induces mitochondrial biogenesis in mouse brain." J Alzheimers Dis 11(1): 45-51.

188

Sun, Y., J. Wang, et al. (2008). "Found in inflammatory zone 1 induces angiogenesis in murine models of asthma." Lung 186(6): 375-380. Suon, S., J. Zhao, et al. (2010). "Systemic treatment with liver X receptor agonists raises apolipoprotein E, cholesterol, and amyloid-beta peptides in the cerebral spinal fluid of rats." Mol Neurodegener 5: 44. Szekely, C. A. and P. P. Zandi (2010). "Non-steroidal anti-inflammatory drugs and Alzheimer's disease: the epidemiological evidence." CNS Neurol Disord Drug Targets 9(2): 132-139. Szeles, L., S. Poliska, et al. (2010). "Research resource: transcriptome profiling of genes regulated by RXR and its permissive and nonpermissive partners in differentiating monocyte- derived dendritic cells." Mol Endocrinol 24(11): 2218-2231. Tamagno, E., P. Bardini, et al. (2002). "Oxidative stress increases expression and activity of BACE in NT2 neurons." Neurobiol Dis 10(3): 279-288. Tan, J., T. Town, et al. (2000). "CD45 opposes beta-amyloid peptide-induced microglial activation via inhibition of p44/42 mitogen-activated protein kinase." J Neurosci 20(20): 7587- 7594. Tan, J., T. Town, et al. (2000). "CD45 inhibits CD40L-induced microglial activation via negative regulation of the Src/p44/42 MAPK pathway." J Biol Chem 275(47): 37224-37231. Tanzi, R. E., J. L. Haines, et al. (1988). "Genetic linkage map of human chromosome 21." Genomics 3(2): 129-136. Tanzi, R. E., R. D. Moir, et al. (2004). "Clearance of Alzheimer's Abeta peptide: the many roads to perdition." Neuron 43(5): 605-608. Terwel, D., K. R. Steffensen, et al. (2011). "Critical role of astroglial apolipoprotein E and liver X receptor-alpha expression for microglial Abeta phagocytosis." J Neurosci 31(19): 7049- 7059. Toledo, E. M. and N. C. Inestrosa (2010). "Activation of Wnt signaling by lithium and rosiglitazone reduced spatial memory impairment and neurodegeneration in brains of an APPswe/PSEN1DeltaE9 mouse model of Alzheimer's disease." Mol Psychiatry 15(3): 272-285, 228. Tontonoz, P. and D. J. Mangelsdorf (2003). "Liver X receptor signaling pathways in cardiovascular disease." Mol Endocrinol 17(6): 985-993. Town, T., Y. Laouar, et al. (2008). "Blocking TGF-beta-Smad2/3 innate immune signaling mitigates Alzheimer-like pathology." Nat Med 14(6): 681-687. Treuter, E. and N. Venteclef (2011). "Transcriptional control of metabolic and inflammatory pathways by nuclear receptor SUMOylation." Biochim Biophys Acta 1812(8): 909-918. Trinh, N. H., J. Hoblyn, et al. (2003). "Efficacy of cholinesterase inhibitors in the treatment of neuropsychiatric symptoms and functional impairment in Alzheimer disease: a meta- analysis." JAMA 289(2): 210-216. Tzeng, S. F. and J. P. Wu (1999). "Responses of microglia and neural progenitors to mechanical brain injury." Neuroreport 10(11): 2287-2292. Uchida, K. and E. R. Stadtman (1992). "Modification of histidine residues in proteins by reaction with 4-hydroxynonenal." Proc Natl Acad Sci U S A 89(10): 4544-4548. Valledor, A. F. (2005). "The innate immune response under the control of the LXR pathway." Immunobiology 210(2-4): 127-132. van Duijn, C. M., T. A. Tanja, et al. (1992). "Head trauma and the risk of Alzheimer's disease." Am J Epidemiol 135(7): 775-782. Vanmierlo, T., K. Rutten, et al. (2011). "Liver X receptor activation restores memory in aged AD mice without reducing amyloid." Neurobiol Aging 32(7): 1262-1272.

189

Varga, T., Z. Czimmerer, et al. (2011). "PPARs are a unique set of fatty acid regulated transcription factors controlling both lipid metabolism and inflammation." Biochim Biophys Acta 1812(8): 1007-1022. Vassar, R., B. D. Bennett, et al. (1999). "Beta-secretase cleavage of Alzheimer's amyloid precursor protein by the transmembrane aspartic protease BACE." Science 286(5440): 735-741. Vlad, S. C., D. R. Miller, et al. (2008). "Protective effects of NSAIDs on the development of Alzheimer disease." Neurology 70: 1672-1677. von Bernhardi, R., J. Tichauer, et al. (2011). "Proliferating culture of aged microglia for the study of neurodegenerative diseases." J Neurosci Methods 202(1): 65-69. Wadosky, K. M. and M. S. Willis (2012). "The story so far: post-translational regulation of peroxisome proliferator-activated receptors by ubiquitination and SUMOylation." Am J Physiol Heart Circ Physiol 302(3): H515-526. Wahrle, S. E., H. Jiang, et al. (2005). "Deletion of Abca1 increases Abeta deposition in the PDAPP transgenic mouse model of Alzheimer disease." J Biol Chem 280(52): 43236-43242. Wahrle, S. E., H. Jiang, et al. (2008). "Overexpression of ABCA1 reduces amyloid deposition in the PDAPP mouse model of Alzheimer disease." J Clin Invest 118(2): 671-682. Wahrle, S. E., H. Jiang, et al. (2004). "ABCA1 is required for normal central nervous system ApoE levels and for lipidation of astrocyte-secreted apoE." J Biol Chem 279(39): 40987-40993. Walsh, D. M., I. Klyubin, et al. (2002). "Amyloid-beta oligomers: their production, toxicity and therapeutic inhibition." Biochem Soc Trans 30(4): 552-557. Wang, K., S. Chen, et al. (2008). "Retinoids induce cytochrome P450 3A4 through RXR/VDR- mediated pathway." Biochem Pharmacol 75(11): 2204-2213. Wang, L., G. U. Schuster, et al. (2002). "Liver X receptors in the central nervous system: from lipid homeostasis to neuronal degeneration." Proc Natl Acad Sci U S A 99(21): 13878- 13883. Watson, G. S., B. A. Cholerton, et al. (2005). "Preserved cognition in patients with early Alzheimer disease and amnestic mild cognitive impairment during treatment with rosiglitazone: a preliminary study." Am J Geriatr Psychiatry 13(11): 950-958. Weggen, S., J. L. Eriksen, et al. (2001). "A subset of NSAIDs lower amyloidogenic Abeta42 independently of cyclooxygenase activity." Nature 414(6860): 212-216. Wegiel, J., H. Imaki, et al. (2003). "Origin and turnover of microglial cells in fibrillar plaques of APPsw transgenic mice." Acta Neuropathol 105(4): 393-402. Wegiel, J., K. C. Wang, et al. (2001). "The role of microglial cells and astrocytes in fibrillar plaque evolution in transgenic APP(SW) mice." Neurobiol Aging 22(1): 49-61. Wesson, D. W., A. H. Borkowski, et al. (2011). "Sensory Network Dysfunction, Behavioral Impairments, and Their Reversibility in an Alzheimer's beta-Amyloidosis Mouse Model." J Neurosci 31(44): 15962-15971. Wesson, D. W., E. Levy, et al. (2010). "Olfactory Dysfunction Correlates with Amyloid-beta Burden in an Alzheimer's Disease Mouse Model." Journal of Neuroscience 30(2): 505- 514. Wilcock, D. M., M. N. Gordon, et al. (2001). "Number of Abeta inoculations in APP+PS1 transgenic mice influences antibody titers, microglial activation, and congophilic plaque levels." DNA Cell Biol 20(11): 731-736. Wilkinson, B., J. Koenigsknecht-Talboo, et al. (2006). "Fibrillar beta-amyloid-stimulated intracellular signaling cascades require Vav for induction of respiratory burst and phagocytosis in monocytes and microglia." J Biol Chem 281(30): 20842-20850.

190

Wilkinson, B. L., P. E. Cramer, et al. (2012). "Ibuprofen attenuates oxidative damage through NOX2 inhibition in Alzheimer's disease." Neurobiol Aging 33(1): 197 e121-132. Wilson, D. A., M. Kadohisa, et al. (2006). "Cortical contributions to olfaction: plasticity and perception." Semin Cell Dev Biol 17(4): 462-470. Wisniewski, K. E., A. J. Dalton, et al. (1985). "Alzheimer's disease in Down's syndrome: clinicopathologic studies." Neurology 35(7): 957-961. Wolf, P. A. (2012). "Contributions of the Framingham Heart Study to Stroke and Dementia Epidemiologic Research at 60 Years." Arch Neurol. Wood, J. G., S. S. Mirra, et al. (1986). "Neurofibrillary tangles of Alzheimer disease share antigenic determinants with the axonal microtubule-associated protein tau (tau)." Proc Natl Acad Sci U S A 83(11): 4040-4043. Wu, G. and S. M. Morris, Jr. (1998). "Arginine metabolism: nitric oxide and beyond." Biochem J 336 ( Pt 1): 1-17. Wyss-Coray, T. (2006). "Inflammation in Alzheimer disease: driving force, bystander or beneficial response?" Nat Med 12(9): 1005-1015. Wyss-Coray, T., J. D. Loike, et al. (2003). "Adult mouse astrocytes degrade amyloid-beta in vitro and in situ." Nat Med 9(4): 453-457. Yamamori, T., O. Inanami, et al. (2000). "Roles of p38 MAPK, PKC and PI3-K in the signaling pathways of NADPH oxidase activation and phagocytosis in bovine polymorphonuclear leukocytes." FEBS Lett 467(2-3): 253-258. Yamamoto, M., T. Kiyota, et al. (2008). "Cytokine-mediated inhibition of fibrillar amyloid-beta peptide degradation by human mononuclear phagocytes." J Immunol 181(6): 3877- 3886. Yan, P., A. W. Bero, et al. (2009). "Characterizing the appearance and growth of amyloid plaques in APP/PS1 mice." J Neurosci 29(34): 10706-10714. Yan, Q., J. Zhang, et al. (2003). "Anti-inflammatory drug therapy alters beta-amyloid processing and deposition in an animal model of Alzheimer's disease." J Neurosci 23(20): 7504- 7509. Yang, H., K. Chen, et al. (2010). "Vitamin A deficiency results in dysregulation of lipid efflux pathway in rat kidney." Pediatr Nephrol 25(8): 1435-1444. Zandi, P. P., J. C. Anthony, et al. (2002). "Reduced incidence of AD with NSAID but not H2 receptor antagonists: the Cache County Study." Neurology 59(6): 880-886. Zelcer, N., N. Khanlou, et al. (2007). "Attenuation of neuroinflammation and Alzheimer's disease pathology by liver x receptors." Proc Natl Acad Sci U S A 104(25): 10601-10606. Zelcer, N. and P. Tontonoz (2006). "Liver X receptors as integrators of metabolic and inflammatory signaling." J Clin Invest 116(3): 607-614. Zempel, H. and E. M. Mandelkow (2011). "Linking Amyloid-beta and Tau: Amyloid-beta Induced Synaptic Dysfunction via Local Wreckage of the Neuronal Cytoskeleton." Neurodegener Dis. Zhang, Q., J. Pan, et al. (2011). "Aerosolized bexarotene inhibits lung tumorigenesis without increasing plasma triglyceride and cholesterol levels in mice." Cancer Prev Res (Phila) 4(2): 270-276. Zhang, W., M. Bai, et al. (2012). "Early memory deficits precede plaque deposition in APPswe/PS1dE9 mice: Involvement of oxidative stress and cholinergic dysfunction." Free Radic Biol Med. Zhang, Y. Q. and K. D. Sarge (2008). "Sumoylation of amyloid precursor protein negatively regulates Abeta aggregate levels." Biochem Biophys Res Commun 374(4): 673-678.

191

Zhou, L., L. H. Shen, et al. (2010). "Retinoid X receptor agonists inhibit phorbol-12-myristate-13- acetate (PMA)-induced differentiation of monocytic THP-1 cells into macrophages." Mol Cell Biochem 335(1-2): 283-289. Zhou, Y., Y. Su, et al. (2003). "Nonsteroidal anti-inflammatory drugs can lower amyloidogenic Abeta42 by inhibiting Rho." Science 302(5648): 1215-1217. Zhu, Y., H. Hou, et al. (2011). "CD45 deficiency drives amyloid-beta peptide oligomers and neuronal loss in Alzheimer's disease mice." J Neurosci 31(4): 1355-1365. Zhu, Y., E. Nwabuisi-Heath, et al. (2012). "APOE genotype alters glial activation and loss of synaptic markers in mice." Glia 60(4): 559-569.

192