BETA-SECRETASE TRANSGENIC MICE:
EFFECTS OF BACE1 AND BACE2
ON ALZHEIMER’S DISEASE PATHOGENESIS
by
MATTHEW J. CHIOCCO
Submitted in partial fulfillment of the requirements
For the degree of Doctor of Philosophy
Dissertation Advisor: Dr. Bruce T. Lamb
Department of Genetics
CASE WESTERN RESERVE UNIVERSITY
May, 2005 CASE WESTERN RESERVE UNIVERSITY
SCHOOL OF GRADUATE STUDIES
We hereby approve the dissertation of
______
candidate for the Ph.D. degree *.
(signed)______(chair of the committee)
______
______
______
______
______
(date) ______
*We also certify that written approval has been obtained for any proprietary material contained therein. Table of Contents
Chapter 1: Introduction and Research Aims ...... 10
Clinical Presentation of AD...... 10
AD Epidemiology...... 16
APP Processing and AD Genetics...... 22
Amyloid Cascade Hypothesis ...... 27
APP Secretases...... 32
BACE1: The primary β-secretase ...... 35
Regulation of β-secretase...... 39
Research Aims...... 46
Chapter 2: Altered Amyloid-β Metabolism and Deposition in Genomic-Based β- secretase Transgenic Mice...... 50
ABSTRACT...... 51
INTRODUCTION...... 52
MATERIALS AND METHODS...... 54
RESULTS ...... 61
DISCUSSION ...... 73
Chapter 3: Spatial and Temporal Control of Age-related APP Processing in
Genomic-based β-secretase Transgenic Mice...... 79
ABSTRACT...... 80
INTRODUCTION...... 81
MATERIALS AND METHODS...... 83
1 RESULTS ...... 87
DISCUSSION ...... 97
Chapter 4: Characterization of APP Processing in Genomic-based BACE2
Transgenic Animals...... 103
INTRODUCTION...... 104
MATERIALS AND METHODS...... 105
RESULTS ...... 109
DISCUSSION ...... 116
Chapter 5: Conclusions and Future Directions...... 119
Summary...... 119
Conclusions...... 121
Future Directions...... 127
Appendix...... 137
Bibliography ...... 139
2 List of Tables
Chapter 1:
Table 1-1: List of AD-associated APP mutations and the proposed
effect on APP cleavage 25
Chapter 3:
Table 3-1: Summary of alterations in brain-regional BACE1
expression and APP processing 99
3 List of Figures
Chapter 1:
Figure 1-1: APP protein processing pathway 23
Figure 1-2: Amyloid cascade hypothesis 29
Chapter 2:
Figure 2-1: Generation of genomic-based human BACE1 transgenic mice 63
Figure 2-2: Human BACE1 mRNA and protein expression and quantitation 65
Figure 2-3: Human BACE1 immunohistochemical localization 67
Figure 2-4: APP processing and Aβ metabolism in BACE1 transgenics 69
Figure 2-5: Analysis of Amyloid-β deposition in BACE1 transgenic animals 70
Figure 2-6: Amyloid-β production and BACE1 protein expression
by brain region 72
Chapter 3:
Figure 3-1: Human BACE1 expression and Aβ production in neonatal
and adult BACE1 transgenic mice 89
Figure 3-2: Human BACE1 protein expression by brain region in
neonatal and adult BACE1 transgenic mice 91
Figure 3-3: Amyloid-β metabolism by brain region in neonatal and
adult BACE1 and APP double transgenic mice 93
4 Figure 3-4: Impact of BACE1 expression on Aβ production by
brain region in adult animals 95
Figure 3-5: Amyloid-β deposition profile in aged BACE1xAPP animals 96
Chapter 4:
Figure 4-1: Generation of genomic-based human BACE2 transgenic mice
and expression of human BACE2 mRNA 111
Figure 4-2: APP processing in BACE2 transgenics 113
Figure 4-3: Analysis of Amyloid-β deposition in BACE2 transgenic animals 115
Appendix:
Figure A1: Aβ production by brain region in hemizgyous
BACE1 x homozygous APP transgenics 137
Figure A2: Profile of Aβ species in human BACE1 transgenics
by immunoprecipitation/mass spectrometry 138
5 Acknowledgements:
I would like to start by thanking my advisor, Bruce Lamb, for “showing me the ropes” and teaching me how to tackle a scientific project. From the very beginning of my rotation project in the lab, Bruce has exhibited unlimited patience through my developing years as a young scientist. He always kept an optimistic perspective when the project took some unfortunate turns as well as congratulated me whenever I made advances. He managed to keep me motivated with encouraging words that kept me focused on my goals as a student. I have been grateful for his willingness to share his life experiences as a student, providing an incredible example of how to achieve success in all my scientific endeavors.
I would also like to acknowledge all members of the laboratory who have helped me throughout my graduate training and helped to make lab an enjoyable professional environment. First, Laura Kulnane has been a superb lab mate who manages the lab with extraordinary precision, dedication, and a willingness to help out wherever possible. I am grateful to Brian Hock and Emily Lehman, my first baymate, as the senior lab mates who helped form my scientific foundation as great teachers and friends early in my graduate career. I would also like to thank Nick Varvel for his assistance with critical immunohistochemical experiments and someone who helped ease the pressures towards the end of my graduate training through both insightful scientific and light-hearted non- scientific discussions. Finally, I am especially grateful to Karen Mann for not only her perceptive comments on my writing, but also her energy and positive attitude that have been a daily source of motivation and encouragement.
6 I would like to thank my thesis committee, Mitch Drumm, Joseph Nadeau, Sanjay
Pimplikar, and Helen Salz for their advice and guidance throughout the course of my project. I am especially grateful to Helen Salz and Sanjay Pimplikar who motivated and challenged me to “tell the story” of my project.
I am also grateful to those friends at CWRU who helped graduate school life more enjoyable. I would also like to thank my family, especially my parents, for their patience and guidance over the past six years. Through their love and support, I was constantly reminded to keep striving for my goals and to always “keep smiling.” Finally, my wife,
Laura, has been the bastion of support throughout my graduate career. I am eternally grateful for her unwavering support, patience, and constant source of tranquility. She has taught me how to be strong, how to endure, and how to persevere through all the challenges over the years. For her love and for inspiring me to achieve my goals as a graduate student, I will always be thankful.
7 Beta-secretase Transgenic Mice:
Effects of BACE1 and BACE2 on Alzheimer’s Disease Pathogenesis
Abstract
by
Matthew J. Chiocco
Numerous investigations have demonstrated that genetic factors play a major role in the etiology of Alzheimer’s disease (AD). Identification of mutations in amyloid precursor protein gene (APP) on chromosome 21 has implicated the role this gene plays in early- onset (<60 years) familial AD. With the identification of the amyloid-β peptide (Aβ) as a characteristic feature of the senile plaques in the AD brain, understanding how this peptide is generated and how it contributes to the pathogenesis of the disease has become a primary focus of investigation. APP processing involves a cleavage pathway carried out by three proteins, α-, β-, and γ-secretases. The Aβ peptide is generated by the cleavage of APP by β-secretase followed by γ-secretase cleavage. Studies have shown that mutations in APP found in a Swedish pedigree with FAD leads to a significant increase in total Aβ production. Modulation of the APP processing enzymes can provide insight into the critical relationship between Aβ production and AD pathogenesis. Two genes thought to be responsible for β-secretase cleavage of APP encode the transmembrane aspartyl proteases, β-site APP cleaving enzyme-1 (BACE1) and β-site
APP cleaving enzyme-2 (BACE2). Extensive in vitro analyses have demonstrated that
BACE1 and BACE2 selectively cleave APP at the β-secretase site. Overexpression of
BACE1 in human cells expressing wild type APP or mutant APP results in an increased production of Aβ peptides. The goal of the current project has been to elucidate the role
8 BACE1 and BACE2 play in generating Aβ from APP in vivo. The overall objective has been to analyze how genomic overexpression of these genes with other AD mutations leads to alterations in APP processing and neuropathology observed in AD. While
BACE2 transgenics do not exhibit altered Aβ levels, BACE1 transgenics dramatically alter the production and deposition pattern of Aβ. These studies demonstrate how β- secretase influences APP processing, resulting in significant consequences on the development of AD pathogenesis. Further studies will provide insight into the regulation of β-secretase as a primary factor in AD neuropathology.
9 Chapter 1: Introduction and Research Aims
As one of the most severe progressively debilitating neurological diseases,
Alzheimer’s disease (AD) causes more than 100,000 deaths per year in the United States
(Lendon et al., 1997; Foundation, 2004). Identified as the most common cause of dementia, AD results in memory loss and impaired cognitive ability, along with progressive neurodegeneration to the point of incapacitation in patients. The incidence of
AD continues to expand as the people born during the population boom of the 1940s and
1950s move into their 60s and beyond (Yankner et al., 1989). This increase in AD cases is due, in large part, to the increase in life expectancy over the past several decades in the developing world. The financial strain of this disease on the health care industry is becoming enormous as the number of individuals affected with AD doubles every five years beyond the age of 65 (Association, 2004). Studies have increasingly focused on the molecular underpinnings of the disease in order to develop better treatments or even devise ways to delay onset of symptoms. Current estimates predict an intervention that delays onset of AD by five years corresponds to a 50% reduction in risk (Brookmeyer et al., 1998). Ever since 1907 with the first description of the disease by Dr. Alzheimer, many researchers have been trying to understand the complex pathogenesis of AD.
Clinical Presentation of AD
From the first clinical diagnosis of this disease in a cognitively impaired individual by Dr. Alzheimer, much research has been devoted to identifying the underlying neurological cause behind the progressive memory loss and language function, among other symptoms, observed in AD cases. In addition to the clinical
10 attributes of AD, Dr. Alzheimer identified two pathological lesions in the patient’s brain upon autopsy. These lesions, known as senile neuritic plaques and neurofibrillary tangles, were further characterized in the 1960s by Kidd and Terry (Kidd, 1963; Terry,
1963; Kidd, 1964; Terry et al., 1964) and are today defined as the classical pathological hallmarks of the disease. Additional characteristics of the AD brain soon included neuronal and synaptic loss, as well as disruptions in cortico-cortical connectivity
(Jellinger and Bancher, 1997). In particular, deficits in cholinergic transmission and associated loss of cholinergic cell bodies is common in AD patients (Selkoe, 2001;
Bossy-Wetzel et al., 2004) and are thought to contribute to the symptoms of dementia and cognitive impairments in AD. However, the diagnostic pathological features of the disease, only identified upon autopsy, are the neuritic plaques and neurofibrillary tangles
(Jellinger and Bancher, 1997), distinguishing AD from other neurodegenerative diseases.
Neuropathological Profile of AD
The precise mechanism for the progression of the pathogenesis is not yet fully understood. However, much research has focused attention on the regional pattern of pathology observed in the AD brain. Understanding the pattern of AD pathogenesis could provide insight into the neurological basis of the disease. Neurofibrillary tangles
(NFTs) are abnormal intracellular accumulations of cytoskeleton filaments, which are wound into structures known as paired helical filaments (PHFs). These filaments are composed of a hyper-phosphorylated microtubule-associated protein tau (Grundke-Iqbal et al., 1986). Tau hyper-phosphorylation leads to neurofibrillary changes classified into two types: neurofibrillary tangles or neuropil threads. Tangles are neurofibrillary changes found in the nerve cell body; neuropil threads are found in the nerve dendrite.
11 The appearance of these neurofibrillary changes follows a specific pattern in the AD brain, depending on age. First, NFTs and neuropil threads start to develop in the enthorhinal cortex of the brain, then proceed into the hippocampus and association areas of the neocortex, and finally invade the primary areas of the neocortex (Braak and Braak,
1997b, 1997a). These intraneuronal accumulations correlate more directly with cognitive impairments than senile plaques (as discussed below), but they can also be found in a wide variety of other neurodegenerative disease pathologies such as Lewy body disease,
Kuf’s disease, and progressive supranuclear palsy (Selkoe, 2001).
Conversely, neurofibrillary changes accompanied by senile neuritic plaques are unique to Alzheimer’s disease pathogenesis. Plaques are typically separated into two categories: (1) diffuse plaques and (2) fibrillar plaques. Diffuse plaques are amorphous structures immunoreactive for the primary constituent of the senile plaque, a 4-kDa peptide known as amyloid-β (Aβ). Fibrillar plaques also contain Aβ protein, but they exist primarily as dense-cored amyloid fibrils organized in a β-pleated sheet structure.
These fibrillar plaques are also known as “neuritic plaques” because of the presence of abnormal neuronal processes called dystrophic neurites surrounding the plaque structure.
It has been shown that these dystrophic neurites sometimes exhibit a structure similar to tau-positive PHFs observed in neurofibrillary tangles (Probst et al., 1989). It is generally accepted that the diffuse plaques mature into the fibrillar form as the disease progresses
(Braak and Braak, 1997b), although both types of plaques are found in end-stage AD
(Dickson and Vickers, 2001).
While plaque development does not follow as strict a regional progression as the neurofibrillary tangles, Aβ deposits can be observed in the basal neocortex, followed by
12 deposition in the frontal cortex and hippocampal formation until all areas of the cortex contain deposits at end stage AD (Braak and Braak, 1997b). Curiously, the pattern of plaque development is somewhat the inverse of that of neurofibrillary tangles. Thus, plaques and tangles co-localize in the same brain region only at end-stage AD (Braak and
Braak, 1997b), and has led to a debate over which structure is responsible for AD neurodegeneration. Nevertheless, the diagnostic criteria for AD follow strict guidelines for the presence of a required number of both plaques and tangles throughout the brain
(Jellinger and Bancher, 1997).
AD Pathology and Cognitive Impairment
A critical feature of the neuropathology profile is how, if at all, both types of pathology correlate with the degree of dementia in AD patients. In general, AD patients have an overall reduced cognitive ability. The progression of AD is characteristically marked by severely impaired learning and memory, consistent with the major symptoms of dementia. Specifically, the earliest cognitive deficits are impairments in episodic memory, the ability to recall events specific to a time and place (Welsh et al., 1992). As the disease progresses, other types of cognitive deficits are observed including attentional and executive functions, such as planning and goal-oriented tasks, as well as semantic memory and language ability (Galton et al., 2000). These cognitive abilities deteriorate further until the threshold for the onset of dementia is reached, whereby the individual experiences difficulties in normal social and occupational function. Identifying the cognitive impairments before the onset of clinical dementia has become a major priority for clinical researchers. Cases of pre-symptomatic dementia are being diagnosed as
“mild cognitive impairment” as a way to track the progression of the condition to the
13 onset and diagnosis of AD proper. With the help of brain imaging technology, clinicians are beginning to utilize positronic emission topography (PET) scans and magnetic resonance imaging (MRI) scans to identify areas of the brain affected by the pathogenesis of the disease (Nestor et al., 2004). This technology can identify patterns of brain volume loss and atrophy in pre-symptomatic cases that may serve as markers for AD cases.
Ultimately, these technologies can be utilized as tools to help pinpoint the relationship between the classic AD pathological profile and neurodegeneration.
One of the key questions for many AD researchers is whether the plaques and tangles actually initiate or even cause the neuronal loss and cognitive decline observed in
AD patients, or do the classic lesions represent “tombstones” of earlier pathological events. The correlation of both these lesions with cognitive decline has become controversial. Several studies suggest that neurofibrillary tangles correlate well with cognitive impairments (Bossy-Wetzel et al., 2004). These studies, including one of the most comprehensive longitudinal studies examining AD pathology and cognitive ability known as the “Nun Study” (Snowdon et al., 1996), have identified strong positive correlations between NFTs and increased cognitive impairment (Haroutunian et al.,
1999), memory impairment (Riley et al., 2002), and linguistic ability (Snowdon et al.,
2000). Many similar studies identified tangle density as highly correlative with a definitive clinical dementia rating. In particular, NFT density in the entorhinal cortex and hippocampus, the brain regions affected earliest by NFT pathology, is most closely associated with cognitive decline, even though tangles have been identified in non- demented subjects (Haroutunian et al., 1999).
14 However, unlike NFTs, the correlations between senile plaque deposition and cognitive ability are somewhat ambiguous. It is well known that the number of amyloid deposits do not necessarily exist in a tight relationship with the degree of dementia for
AD patients. Senile plaques have been identified in individuals without any overt cognitive impairment (Lue et al., 1999), leading to the possible conclusion that senile plaque deposition is a normal characteristic of aging. The relationship may have a positive association when the brain region exhibiting pathology is taken into account.
For example, Riley et al. identified a statistically significant association between neuritic plaques and cognitive state in both the neocortex and hippocampus (Riley et al., 2002), which are the brain regions affected earliest in the pathogenesis in the AD brain (Braak and Braak, 1997b). In addition, the numbers of insoluble plaque deposits may not be the only facet of AD pathogenesis that corresponds to the degree of dementia. The levels of soluble peptide contained in the plaques, amyloid-β, may account for the association with cognitive ability. As shown by Naslund et al., the levels of amyloid-β in the brain are highly correlative with cognitive decline (Naslund et al., 2000) and may, in fact, be a strong predictor of AD cases (Lue et al., 1999).
The most convincing evidence for the involvement of plaque pathology is the strong correlation of AD neuropathology with Down Syndrome (DS) individuals. Not only do DS patients exhibit clinical symptoms of AD as early as their 30s or 40s, but the degree of dementia corresponds with the density of plaques and tangles (Wisniewski et al., 1985). Even though NFTs correlate more positively with cognitive decline, the observations regarding senile plaque formation and amyloid-β levels may indicate that plaque accumulation is an early event in AD pathogenesis. Based on the variability of
15 AD pathology in patient populations, the development of AD pathogenesis is, nevertheless, marked by a high degree of heterogeneity, indicating genetic and environmental factors contribute to the onset and development of the disease.
AD Epidemiology
Age
Increasing age confers the highest risk for AD. The majority of cases are sporadic for which there is no known cause; however, the most consistent feature of the disease is that it primarily affects individuals over age 60. It is estimated that 4 million people are affected with AD, which is figured largely on studies that dementia affects ~ 15% of the population over 60 years of age and AD is considered to account for two-thirds of the diagnoses of dementia (~10% of the over-60 population) (Ashford, 2004). The occurrence of AD increases with age, almost doubling the incidence every 5 years
(Ashford, 2004). The classical pathological hallmarks of the disease, plaques and tangles, are detected in the brains of increasingly older individuals (Braak and Braak,
1997b). Sparse pathology can be detected as early as the third decade of life, but the plaques and tangles litter almost all areas of the brain by the eighth and ninth decade
(Braak and Braak, 1997b). Because of the risk age presents for the elderly population,
AD incidence rates become increasingly higher, with a prevalence of 1/1000 by age 62,
1/100 by age 79, and 1/10 by age 94 (Ashford, 2004). These rates further highlight age as the primary risk factor in AD.
Environment/Education/Gender
Although increasing age dramatically affects the risk of AD, the age of onset is highly variable, indicating the interaction of other risk factors. Epidemiological studies
16 have identified several environmental factors associated with the prevalence of AD.
Based on twin studies, the environment may account for ~ 50%-70% of the risk associated with AD (Gatz et al., 2005). One of the early postulated risk factors for AD was exposure to heavy metals, particularly aluminum, in drinking water (Breteler et al.,
1992). However, analysis in other studies have identified no risk for exposure to aluminum, but testing bias and the inability to properly diagnose AD could be confounding factors.
Another risk factor increasing the incidence of AD is head trauma. Several studies have identified an increased risk of AD in individuals with a history of head trauma (Breteler et al., 1992; van Duijn et al., 1992). Furthermore, individuals with a history of head trauma (primarily boxers) are diagnosed with dementia pugilistica, also known as punch-drunk syndrome, and share similar neurofibrillary tangle pathology to
AD patients (Allsop et al., 1990). It is not known whether the head trauma initiates the pathogenic events common to AD or is a consequence of early dementia, but association studies implicate a significant risk in the incidence of the disease.
One risk factor that has been met with controversy is the degree to which the level of education confers risk for AD. Studies have shown that individuals with higher education levels may reduce the relative risk of dementia and/or AD (Zhang et al., 1990;
Stern et al., 1994). It was shown that low education levels and low occupation groups exhibited higher rates of dementia (Stern et al., 1994). It is possible that higher education levels allow an individual to preserve cognitive function longer and practice cognitive skills increasing intellectual stimulation. The prevailing hypothesis is that higher education levels may also establish a minimal amount of brain activity, creating a
17 “reserve” that may reduce neuronal death due to AD (Breteler et al., 1992; Stern et al.,
1994). However, such studies may be inherently flawed in terms of testing bias, dementia diagnosis, and may even be influenced by socioeconomic status.
Another risk factor that has been implicated in increased risk for AD is gender.
Many studies have indicated women are more susceptible to AD, and suggest that two- thirds of women will be afflicted with the disease, whereas only one-third of men will contract AD before they die (Ashford, 2004). The underlying mechanisms are not known, but the gender bias may be due to the increased risk for women who are carriers of the ε4 allele of the Apolipoprotein E gene (Payami et al., 1996). In addition, sex hormone levels, particularly estrogen levels, may confer a neuroprotective effect that is lost upon the onset of menopause in women (Brinton, 2004; Li and Shen, 2005).
Longevity and increased life expectancy for women are mitigating factors for this gender susceptibility, as some studies control for longevity and report increased risk in men
(Breteler et al., 1992; Lendon et al., 1997). All of these epidemiologic studies are hampered by testing bias, inefficient diagnostic criteria, and population heterogeneity.
The inconsistencies in identifying risk factors demonstrate the complexity of AD in both clinical presentation of cognitive function and molecular pathogenesis.
Family History/Genetics
One of the most consistent risk factors, along with age, for AD is positive family history of dementia. Some of the earliest AD case-control studies have reported significantly increased AD risk for relatives of patients with dementia (Breteler et al.,
1992). Specifically, family studies have shown a substantially increased risk for first- degree relatives of individuals with AD at autopsy (Ashford, 2004) (Rubinsztein, 1997).
18 Twin studies, in particular, have established the risk due to genetic factors at ~ 30%-50%
(Ashford and Mortimer, 2002; Gatz et al., 2005). Based on inheritance patterns and pedigree analysis, Familial AD (FAD) is separated into two categories: (1) early-onset
AD (EOAD) affects individuals < 65 years of age and follows an autosomal dominant mode of inheritance; (2) late-onset AD (LOAD) affects individuals > 65 years of age and follows a complex mode of inheritance.
Identifying a genetic component to the disease was initially supported by the observation of trisomy 21 Down syndrome individuals, who exhibited the clinical features of dementia as well as the classic pathological hallmarks of AD, plaques and tangles (Lai and Williams, 1989). Because these individuals typically present dementia symptoms as early as their fourth decade of life and have an extra copy of chromosome
21, investigators focused on chromosome 21 as a major potential link to an early-onset form of AD pathogenesis. By the mid-1980s, linkage analysis of several FAD pedigrees pointed to a locus on chromosome 21 significantly linked to FAD cases (St George-
Hyslop et al., 1987). The discovery of the primary constituent of the senile plaque, amyloid-β (Aβ) peptide (Glenner and Wong, 1984), helped identify and isolate the gene by finding that Aβ comes from a larger protein known as Amyloid Precursor Protein
(APP) (Goldgaber et al., 1987; Tanzi et al., 1987). Because not all the early-onset cases showed linkage to APP on chromosome 21, other genes were thought to be involved and soon identified as Presenilin1 (PSEN1) on chromosome 14 (Sherrington et al., 1995) and
Presenilin2 (PSEN2) on chromosome 1 (Levy-Lahad et al., 1995) through linkage analysis with a high prevalence of AD. Based on their dominant mode of inheritance,
19 mutations in these three genes (APP, PSEN1, and PSEN2) are thought to be causative for almost ~ 50% of early-onset FAD pedigrees (Thinakaran, 1999).
Conversely, due to the complex etiology of LOAD, loci and candidate genes, identified in association studies, cannot individually cause the disease and are simply described as AD risk factors (Rocchi et al., 2003). Some of the genes include the
Apolipoprotein E gene (ApoE) on chromosome 19, α2-Macroglobulin on chromosome
12, and Insulin degrading enzyme on chromosome 10 (Bertram et al., 2000), as well as a list of other candidate genes and loci (Rocchi et al., 2003). Furthermore, the contribution of these risk factors to the disease in late-onset cases is not well understood. The function of most of these LOAD candidate genes is unknown; particularly what role each gene may play in pathogenesis. The most prominent risk factor, ApoE4 protein, has been hypothesized to be involved in both facilitating plaque development by binding to the Aβ peptide as well as clearance of Aβ from the extracellular space before plaque formation occurs in the brain.
On the other hand, AD pathogenesis for early-onset cases is intimately linked to the accumulation of Aβ peptide. Association of the early-onset APP, PSEN1, or PSEN2 genes in FAD pedigrees is significantly associated with increased levels of amyloid-β peptide. The contribution of Aβ to AD pathogenesis is corroborated by the observation that DS individuals, carrying an extra copy of chromosome 21 (including APP), develop the clinical AD phenotypes early in life. Additionally, a recent report identified a single
DS case due to partial trisomy 21, excluding the APP gene, who exhibited no clinical and neuropathological evidence of AD (Prasher et al., 1998), suggesting dosage imbalance of
APP that leads to elevated Aβ levels may cause the disease. Thus, the mechanism for
20 AD pathogenesis involves the role the early-onset AD genes play in regulating Aβ levels, contributing to the deposition of the peptide in senile plaques.
APP
The APP gene contains 18 exons which codes for an ~ 100 kDa ubiquitously expressed protein. Due to alternative splicing, three primary isoforms can be detected with non-neuronal cells expressing the 751 amino acid (aa) and 770 aa splice forms containing a Kunitz-protease inhibitor domain, while neurons predominantly express a
695-aa isoform (Selkoe, 1998). APP is a transmembrane protein located primarily on the cell surface as well as on the endoplasmic reticulum, Golgi apparatus, and endosomes.
The physiological function of APP remains unknown, but studies have identified potential functions based on interactions with its intracellular C-terminal domain.
Interaction of the C-terminal domain with an adaptor protein, known as Fe65, can activate transcription (Cao and Sudhof, 2001), while interaction with another protein,
Tip60, induces an apoptotic pathway (Kinoshita et al., 2002). Alternatively, interactions with other proteins, such as PAT1 or REELIN, indicate that APP may play a role in neuronal trafficking and neuronal migration (Van Gassen et al., 2000).
In addition, APP may play a critical role in proper cellular survival as recent studies have identified premature death in animals overexpressing APP (Lehman et al., unpublished results) (Hsiao et al., 1995; Krezowski et al., 2004). Additionally, while
APP knockout mice are viable (Zheng et al., 1996), removal of APP along with a combination of its protein family members, APLP1 and/or APLP2, causes premature lethality (von Koch et al., 1997; Herms et al., 2004), indicating the APP protein family may play a major role in development. Although a clear function of APP has yet to be
21 identified, it is the biochemical pathway leading to Aβ generation that demonstrates its link to AD pathogenesis.
APP Processing and AD Genetics
APP is subject to cleavage by three proteases, α-,β-, and γ-secretases through two primary pathways. The first pathway involves cleavage of APP 12 residues N-terminal to the transmembrane sequence by a group of integral membrane enzymes. Termed α- secretase cleavage, this cut creates a large, ectodomain fragment (APPs-α) that is released from the cell surface and a smaller C-terminal fragment (83 aa) attached to the membrane
(called C83 or CTF-α). Subsequent cleavage of this membrane-retained fragment by a unique complex of proteins, referred to as γ-secretase, releases a 3-kDa peptide (p3) that does not form amyloid fibrils (see Figure 1-1).
The second pathway; however, involves an alternative cleavage by a different secretase, termed β-secretase, located at the N-terminus of the Aβ sequence. This cleavage releases a unique ectodomain fragment (APPs-β) and a C-terminal 99 aa membrane-retained fragment, known as C99 or CTF-β. Then, cleavage of this CTF-β by
γ-secretase results in formation of Aβ (see Figure 1-1). The length of the C-terminal end of Aβ varies depending on the specificity of γ-secretase cleavage, resulting in production of Aβ1-40 or a longer, more fibrillogenic Aβ1-42/43 (Selkoe, 1998). The longer Aβ peptide has a major role in Aβ deposition because plaque formation follows a distinct pattern with the 42-43 amino acid peptide detected first followed by the more common
Aβ1-40 (Lemere et al., 1996).
Clearly, it is the work of β-secretase together with γ-secretase that releases the 4- kDa Aβ peptides. The principal β-secretase recognition site is located at residues
22
Figure 1-1: APP protein processing pathway.
APP is drawn with the signal peptide (1-17aa), Kunitz-protease inhibitor domain (KPI) and transmembrane domain (TM) covering 700-723 aa. APP can be cleaved in one of two ways: A) α-secretase cleavage at 687 aa, releasing APPs-α and CTF-α. Cleavage by γ-secretase at 711 or 713 within TM releases the p3 peptide. B) β-secretase cleavage at 671 aa (Asp1 of Aβ), releasing APPs-β and CTF- β. Cleavage by γ-secretase at sites 711 or 713 within TM generates the 4-kDa Aβ1-40 or Aβ1-42/3 peptides. Alternative cleavage by β-secretase at 681 aa (Glu11 of Aβ), followed by γ-secretase cleavage, leads to N-terminal truncated Aβ11-40 or Aβ11-42/3 peptides.
23 671/672 of APP770. Interestingly, two Swedish families with a high incidence of early- onset AD were identified that express mutations in the two amino acids flanking the β- secretase cleavage site (aa 670/671) (Mullan et al., 1992). The APP Swedish mutation codes for two base pair transversions (G to T and A to C) at codons 670/671, resulting in changes from lysine to asparagine and methionine to leucine (denoted KM670/671NL).
In vitro cell studies and in vivo mouse studies have shown that this Swedish double mutation results in increased β-secretase cleavage and a resulting 6-8-fold increase in Aβ production compared to cells expressing wild-type APP (Citron et al., 1992; Lamb et al.,
1997). Other missense mutations have been found in early-onset FAD pedigrees (see
Table 1), including a valine to isoleucine substitution at residue 717 of APP770 (denoted
V717I) in sixteen different families (Goate et al., 1991). This mutation alters the specificity of γ-secretase cleavage, leading to higher production of the longer Aβ1-42/43 peptide (Duff et al., 1996; Jankowsky et al., 2004). In fact, most of the missense mutations identified in APP are located at or near APP cleavage sites, reinforcing the genetic relationship between Aβ production and AD (see Table 1). Since missense mutations in the APP gene account for less than 0.1% of all AD cases (Selkoe, 1999) and only 5% of all early-onset AD cases (Rubinsztein, 1997), more genes have been and will be implicated in early-onset AD.
PSEN1 and PSEN2
Two of the genes implicated in early-onset FAD almost immediately after the discovery of linkage to APP were the Presenilin genes. Since early studies linking FAD to loci on human chromosomes 14 and 1 (Levy-Lahad et al., 1995; Sherrington et al.,
1995), numerous mutations have been identified in PSEN1 and PSEN2 that are associated
24 Table 1-1: List of AD-associated APP mutations and the proposed effect on APP cleavage.
APP770 Site of Type of Population AD Type Authors Codon Mutation Substitution 670/671 β-secretase, KM NL Swedish Early-onset (Mullan et Aβ N-term al., 1992) 717 γ-secretase, Val Ile British Early-onset (Goate et 1 Aβ C-term. al., 1991) 692 α-secretase Ala Gly Flemish Early-onset, (Hendriks et CAA al., 1992) 714 γ-secretase, Thr Ala Iranian Early-onset (Pasalar et Aβ C-term al., 2002) 715 γ-secretase, Val Met Italian Early-onset (Ancolio et Aβ C-term al., 1999) 716 γ-secretase, Ile Val N/A2 Early-onset (Eckman et Aβ C-term al., 1997) 693 α-secretase Glu Gly N/A2 Early-onset, (Kamino et CAA al., 1992) 694 α-secretase Asp Asn USA: Iowa Early-onset, (Grabowski (German CAA et al., 2001) descent) 665 β-secretase? Glu Asp N/A Late-onset3 (Peacock et al., 1994) 1 This mutation was originally identified by Goate at al., 1991 and is the most common variant, but different allelic variants at codon 717 have been identified. 2 These variants were identified from a pool of mixed AD samples. 3 This is the only APP mutation associated with late-onset AD.
25 with FAD. About 120 different missense mutations have been identified in these two genes. PSEN1 and PSEN2 are causative for AD in approximately 50% of individuals with early-onset AD (Thinakaran, 1999). Identified as paralogues sharing 67% aa amino acid identity, the genes encode integral membrane proteins containing multiple transmembrane domains. The proteins are primarily localized in the endoplasmic reticulum, Golgi apparatus, and nuclear envelope until they undergo endoproteolysis, yielding amino-terminal and carboxy-terminal derivatives (Kovacs et al., 1996)
(Thinakaran et al., 1997). The function of PSENs is physiologically unknown, however, it was determined that homologues of presenilin1, first in Drosophila and C. elegans and later in mice, are involved in proteolytic cleavage of Notch1, an integral membrane protein that follows a cleavage pathway similar to APP (Levitan and Greenwald, 1995;
De Strooper et al., 1999; Struhl and Greenwald, 1999).
Given the association with early-onset FAD, many researchers have used mutant
PSEN genes to determine how they are causative for AD. More specifically, various expression studies have been utilized to analyze how the generation of the longer, amyloidogenic Aβ1-42/43 may cause pathogenesis. Aβ1-42/43 levels were increased in plasma and fibroblasts from AD patients known to have PSEN1 or PSEN2 mutations
(Scheuner et al., 1996). Cell culture studies have shown that an FAD-linked mutation in
PSEN2 from a Volga German family showed 5-10-fold increase of secreted Aβ1-42/43 compared with Aβ1-40 when transfected with APP cDNA (Tomita et al., 1997). In hopes of identifying the primary γ-secretase gene, studies using in vivo transgenic approaches helped characterize the effect the PSEN mutations had on APP processing and eventual
Aβ secretion (Duff et al., 1996; Lamb et al., 1999). These studies, in addition to recent
26 transgenic studies using FAD PSEN mutants (Jankowsky et al., 2004), demonstrate that the mutations in PSEN genes are involved in γ-secretase cleavage of APP to specifically alter Aβ1-42/43 production. Although removal of Psen1 lethality in mice results in embryonic lethality due to the protein’s role in Notch signaling (De Strooper et al., 1998), there is a five-fold reduction in Aβ from embryonic knockout fibroblast cultures. In
Psen1 conditional knockouts, mice exhibit a 40% reduction of Aβ when crossed to APP transgenic mice (Yu et al., 2001), which further indicates a role for PSEN1 in amyloidogenic processing.
Because the α- and β-secretase cleavage products accumulate in Psen1 KO mice, it is clear that PSEN1 is required for γ-secretase cleavage. However, PSEN1 or PSEN2 exhibit no native proteolytic activity. Recently, work has shown that γ-secretase activity relies on a large multimeric protein complex comprised of PSEN1 (and sometimes
PSEN2), along with Nicastrin, Aph-1, and Pen-2. This complex is also required for proper Notch signaling and downstream transcriptional activation of genes required for development and neurogenesis (Wolfe and Kopan, 2004). This dual role of the
Presenilins suggests they play a major part in normal cellular function. However, as integral players in the APP cleavage pathway, the PSEN genes are key components in understanding the relationship between Aβ and AD pathogenesis.
Amyloid Cascade Hypothesis
Mutations in each of the three genes identified by genetic studies in early-onset
AD are all linked to elevated Aβ levels in these families. The association of the FAD genes with increased Aβ levels and senile plaques have led to the hypothesis that Aβ is the primary causative agent in AD pathogenesis, known as the amyloid cascade
27 hypothesis (Figure 1-2). Investigations into the molecular mechanisms of Aβ generation and deposition have lent further support to the amyloid cascade, which is widely accepted in the field. However, focusing only on Aβ production and deposition discounts the second major pathological hallmark of Alzheimer’s disease: the neurofibrillary tangle.
Recent adaptations to the amyloid cascade hypothesis describe the role Aβ deposits may play in tangle pathology or even that neurofibrillary tangles are end-stage pathological markers not involved in AD pathogenesis (Hardy and Selkoe, 2002). This new hypothesis relies not just on the presence of the insoluble peptide in the plaque core of both AD patients and Down syndrome individuals (Masters et al., 1985), but also the observed increase in soluble levels of Aβ in these individuals, as well as a separate “high pathology” cohort of individuals who do not yet meet AD clinical diagnostic criteria (Lue et al., 1999). The biochemical detection of the soluble peptide indicates accumulation of
Aβ peptide may be a pathological predictor of cognitive decline and addresses the main criticism for the amyloid hypothesis that amyloid plaque deposition does not correlate with cognitive impairment. Also, the amyloid hypothesis lacks an explanation for the poor association between Aβ accumulation and neuronal loss in AD. Although Aβ cyto- toxicity has been observed in vitro (Yoshikawa et al., 1992; Takahashi et al., 2000), there is no clear mechanism for how Aβ causes neurodegeneration in the AD brain.
Nevertheless, the foundation for the amyloid hypothesis rests in the genetic mutations associated with AD cases. Thus, the hypothesis provides a starting point for understanding how mutations in these genes and other genes that influence Aβ production and deposition can lead to AD pathogenesis.
28
Figure 1-2: Amyloid cascade hypothesis.
Schematic diagram illustrating the factors associated with Aβ production and amyloid deposition in AD. APP can be cleaved by either α-secretase, which precludes the formation of Aβ, or by β-secretase, which leads to the generation of Aβ peptides. The foundation for this hypothesis is the presence of genetic mutations in Familial Alzheimer’s Disease pedigrees. Most notably, the Swedish FAD mutation at APP 670/671preferentially increases β-secretase cleavage and leads to elevated levels of the 99 amino acid precursor, CTF-β. Mutations at APP 717 and PSEN1 alter γ-secretase cleavage and lead to increased production of the more fibrillogenic Aβ1-42/3.
29 Aβ Transport and Clearance
Another key aspect of the amyloid cascade is the fate of the Aβ peptide once generated in the brain. Since Aβ peptides are found in non-demented healthy individuals, production of Aβ throughout the brain is thought to be a normal process, that is until the production and accumulation becomes pathogenic. The amyloid cascade hypothesis states that the increased generation of Aβ peptides caused by FAD mutations initiates AD pathogenesis and the corresponding synaptic dysfunction and neuronal loss.
Axonal Transport of APP
To draw the connection between Aβ production and neurodegeneration, researchers have tried to investigate the normal function of its precursor protein, APP, in the neuron. As an integral membrane protein, APP participates in axonal transport by the fast anterograde system (Koo et al., 1990). Furthermore, APP axonal transport, mediated by Kinesin-1 proteins associated with microtubules, may represent its physiologic function as a membrane cargo receptor through the nerve terminal (Kamal et al., 2000).
Recent studies suggest that transport of APP through the axon may, in fact, contribute to the initiation of synaptic dysfunction through the aggregation of Aβ peptides in the nerve terminal. The co-localization of APP and its processing enzymes (β-secretase and γ- secretase) in an axonal membrane compartment argues the production and deposition of
Aβ in the brain may not be a local pathogenic event (Kamal et al., 2001). Although these findings have drawn much debate regarding the localization of the Aβ-generating proteins, there is a precedent in vivo for axonal transport of the APP processing partners, particularly β- and γ-secretase, through nerve terminals in the brain (Sheng et al., 2003).
Whether such a co-localization exists in axonal membrane compartments, the possibility
30 that improper trafficking due to increased Aβ production can lead to AD pathogenesis posits an interesting connection between Aβ deposition and neurodegeneration.
Another facet of “Aβ fate” depends on the cell’s ability to remove the peptide before it accumulates to toxic levels. According to the amyloid hypothesis, the increased production of Aβ observed in FAD individuals may hinder the cell’s ability to clear the peptide as it aggregates and forms deposits in the brain. If the amyloid hypothesis is correct, then stimulating the clearance of Aβ from the brain may be as important as inhibiting Aβ production for an AD therapeutic approach. Within the brain, extracellular
Aβ peptides can be degraded by microglia cells, which are the brain’s macrophages.
Based on the accumulating number of studies, there are two other known mechanisms that facilitate Aβ clearance: (1) peptidolytic degradation and (2) receptor-mediated removal of Aβ from the brain.
Peptidolytic Degradation of Aβ
Based on in vitro binding studies, a variety of peptides have been identified that can degrade Aβ (Morelli et al., 2002). Recent studies have built strong evidence for the ability of two separate peptides, neprilysin (NEP) and insulin degrading enzyme (IDE), to hydrolyze Aβ. Interestingly, the IDE locus on chromosome 10 has been implicated in a recent genetic screen for FAD mutations (Bertram et al., 2000). As zinc metalloendopeptidases, both IDE and NEP have been shown to dramatically reduce Aβ levels in culture and reduce amyloid plaque burden in transgenic mice (Leissring et al.,
2003; Marr et al., 2003). In addition, deficiency in either of these peptides, by knockout for NEP (Iwata et al., 2001) or a known IDE deficiency for type 2 diabetes mellitus
(Farris et al., 2004), results in significant increases of Aβ peptides in mice.
31 Receptor-Mediated Aβ Clearance
Besides hydrolysis by proteases, Aβ can be removed from the brain by a completely separate mechanism, a receptor-mediated clearance pathway. Two receptor- associated proteins have been implicated in Aβ clearance, low-density lipoprotein receptor-related protein (LRP) and the receptor for advanced glycation end products
(RAGE). Binding Aβ directly, in the case of RAGE, or through a complex of other proteins including ApoE and α-2-macroglobulin, in the case of LRP, Aβ can cross the blood-brain barrier and be transported to the peripheral organs for degradation, such as the liver (Kang et al., 2000; Shibata et al., 2000; Deane et al., 2003). Using these pathways, various therapeutic strategies are being investigated for removal of Aβ peptides from the brain.
APP Secretases
While clearance of Aβ may prove to be an important therapeutic strategy, the main premise of the amyloid cascade hypothesis posits that the production of Aβ may play a more important role in AD pathogenesis. Supported by genetic studies, the hypothesis argues that the pathologic accumulation and aggregation of Aβ in FAD patients causes early-onset AD. Thus, the cleavage of APP to generate Aβ becomes a primary event in understanding AD pathogenesis. As stated above, APP processing follows two distinct pathways: α-secretase cleavage precludes Aβ production, while cleavage by β-secretase leads to potentially pathogenic Aβ. Accordingly, understanding the balance between α-secretase and β-secretase cleavage of APP can provide invaluable insights into Aβ production and AD pathogenesis.
32 Genetic mutations for early-onset AD underscore the significance of this balance between amyloidogenic and non-amyloidogenic processing of APP. These mutations influence the critical choice involved in APP processing: cleavage by α-secretase vs. β- secretase. For example, β-secretase cleavage is dramatically up-regulated in individuals expressing the APP Swedish double mutation, resulting in a significant increase in β- secretase cleavage product, CTF-β, leading to increased Aβ (Citron et al., 1992). This missense mutation flanks the β-secretase cleavage site, and thus the Asp1 of the Aβ peptide, and modulates the APP processing pathway to favor β-secretase cleavage.
Similarly, a missense mutation at codon 692 of APP770 causes a unique cerebral angiopathy, which is accompanied by increased Aβ production (Hendriks et al., 1992;
Haass et al., 1994). This mutation appears to alter the α-secretase cleavage site, creating an increased frequency of β-secretase cleavage.
Other FAD mutations, particularly the mutation at codon 717 of APP770 and the many Presenilin mutations, alter the Aβ peptide length from the shorter, more abundant
Aβ1-40 to the longer, more fibrillogenic Aβ1-42/3. These mutations alter the specificity for γ-secretase processing and have a dramatic impact on the Aβ deposition process.
Recently, substantial advances have been made in identifying the APP α- and β- secretases and characterizing their role in AD pathogenesis.
Identification of α-secretase
Several studies have identified genes involved in α-secretase cleavage. Both in vitro and in vivo studies using human cells and mouse knockout studies have directly implicated two members of a disintegrin and metalloprotease family (ADAM) as having
α-secretase activity, ADAM10 (Lammich et al., 1999) and the membrane-bound tumor
33 necrosis factor α-converting factor (TACE) (Buxbaum et al., 1998). Also, a prohormone convertase, PC7, has been identified as playing a role in α-secretase cleavage of APP in human cells (Lopez-Perez et al., 1999). Most recently, the generation of ADAM10 transgenic mice provides in vivo evidence that stimulation of the α-secretase pathway, by overexpression of ADAM10, increases α-secretase cleavage products and reduces Aβ levels (Postina et al., 2004). Furthermore, current treatments for AD, specifically acetylcholinesterase inhibitors, stimulate the α-secretase pathway by selectively increasing ADAM10 activity (Zimmermann et al., 2004), solidifying its role as a major player in α-secretase processing of APP as well as providing a mechanism for understanding AD pathogenesis.
Identification of β-secretase
Three different groups identified the gene encoding β-secretase by exclusively screening for increased Aβ levels or by analyzing cDNA libraries capable of cleaving
APP in cells (Sinha et al., 1999; Vassar et al., 1999; Yan et al., 1999). Designated β-site
APP Cleaving Enzyme-1 (BACE1, also called Asp2 or memapsin2), BACE1 is a 30.5 kb gene that codes for a 501 aa protein and contains two aspartyl protease active site motifs at aa 93-96 and 289-292, with each motif containing the highly conserved signature sequence of aspartic proteases D T/S G T/S (Vassar et al., 1999). The BACE1 gene, located on human chromosome 11q23.2, gives rise to three transcripts of 7.1, 4.4, and 2.6 kb. The protein itself is ~ 70 kDa with a single transmembrane domain, four N- glycosylation sites, and three disulfide bridges (Haniu et al., 2000). BACE1 is expressed at low levels in most peripheral tissues with robust expression in the pancreas, where it is proteolytically inactive (Bodendorf et al., 2001). BACE1 is expressed highest in the
34 brain with a regional expression pattern that parallels APP expression, with highest levels in neurocortex and hippocampus (Irizarry et al., 2001) and lowest levels in the cerebellum (Zohar et al., 2003). This differential expression pattern may have implications for Aβ production as the brain regions with the highest BACE1 expression
(cortex) exhibit the most Aβ aggregates, while the regions with the lowest BACE1 expression (cerebellum) exhibit little to no Aβ accumulation. At the cellular level,
BACE1 expression is localized to the trans-Golgi network and endosomal pathway, particularly showing co-localization with APP in endosomes (Hussain et al., 1999;
Kinoshita et al., 2002).
BACE1: The primary β-secretase
One interesting aspect of BACE1 is its proteolytic activity. When transfected in human cells expressing APP, BACE1 cleaves APP at the known β-secretase sites, Asp1 of Aβ and Glu11 within the Aβ peptide. Overexpression of BACE1 by transient transfection results in significantly increased β-secretase activity in cells expressing wild type or Swedish mutant APP (Vassar et al., 1999). In particular, Aβ levels are significantly increased in cells transfected with BACE1 and either wild-type or mutant
APP (Sinha et al., 1999). This Aβ-producing activity is primarily found in endosomal compartments where BACE1 is most highly expressed in the cell, although BACE1 can also be detected on the plasma membrane (Huse et al., 2000). It has been clearly demonstrated that BACE1 functions as the putative β-secretase in culture.
Additional evidence demonstrating the significance of BACE1 cleavage of APP in generating Aβ peptide came from Bace1 knockout mouse studies signifying BACE1 is the primary β-secretase. When Bace1 is removed in mice, Aβ peptides are completely
35 eliminated (Cai et al., 2001; Luo et al., 2001; Roberds et al., 2001). In addition, there is a concomitant increase of α-secretase-derived products, namely APPs-α and CTF-α, indicating a shift in the APP processing pathway (Cai et al., 2001; Luo et al., 2001). The
Bace1 KO mice also do not exhibit any deficiencies in motor function or in behavioral attributes, such as learning and memory (Roberds et al., 2001). In fact, when the Bace1
KO mouse is crossed to an APP FAD mutant transgenic model, the lack of Bace1 expression rescues memory deficits and electrophysiological abnormalities, normally observed in these mice (Ohno et al., 2004). These studies establish BACE1 as the primary β-secretase and demonstrate the pivotal role BACE1 may play in AD pathogenesis.
To gain insight into the role of BACE1 in the pathogenesis of AD, several groups have utilized genetic association studies in case-control populations. In almost all the screens in BACE1 coding and non-coding sequence, one unique single nucleotide polymorphism was discovered in exon 5 coding for a single amino acid substitution in codon 262 (Nicolaou et al., 2001; Nowotny et al., 2001; Clarimon et al., 2003; Gold et al., 2003; Liu et al., 2003). However, in certain populations of AD cases, no significant linkage was identified (Cruts et al., 2001; Nicolaou et al., 2001; Liu et al., 2003). Most recently, this polymorphism has been significantly associated with an AD population, particularly in carriers of APOEε4 alleles (Nicolaou et al., 2001; Nowotny et al., 2001;
Gold et al., 2003). Furthermore, a significant association in AD cases was identified for the exon 5 allele in a Chinese population in non-APOEε4 carriers (Shi et al., 2004).
While the BACE1 association studies suggest a genetic link to AD, the functional relevance for the polymorphism on AD pathology in these patients is not yet known.
36 In addition to the relatively weak genetic association studies for BACE1, there is solid evidence for a functional link to AD. Several groups have identified significant increases in β-secretase protein expression and activity in AD cases when compared to controls. This increase in activity also correlates with increases in Aβ production (Li et al., 2004) and is accompanied by a decrease in α-secretase activity (Tyler et al., 2002).
These increases in expression and activity are seen primarily in the neocortex and hippocampus, the brain regions affected earliest and most severely in AD (Fukumoto et al., 2002; Yang et al., 2003). In addition, β-secretase activity increases with age in the human brain along with elevated Aβ levels (Fukumoto et al., 2004). These studies provided in vivo evidence for the functional relevance of BACE1 in the AD brain.
BACE2: an alternative α-secretase?
The identification of BACE1 also led to the identification of a BACE paralogue,
BACE2 (also called Asp1 or memapsin-1). By GenBank database searches of ESTs,
BACE2 sequence was revealed (Saunders et al., 1999). Interestingly, BACE2 maps to human chromosome 21q22.3, within the Down syndrome critical region (Yan et al.,
1999) (Acquati et al., 2000). BACE1 and BACE2 exhibit 52% aa sequence identity and
68% similarity, sharing structural properties as unique aspartyl proteases with a single transmembrane domain (Acquati et al., 2000) (Saunders et al., 1999). However, these two genes may not be functional homologues. Unlike BACE1, BACE2 expresses two transcripts of 2.0 and 2.6 kb with highest expression in colon, kidney, and stomach
(Bennett et al., 2000). Expression in the periphery is inconsistent with the predicted β- secretase locations (i.e. in central nervous system (CNS) neurons) required for generation of Aβ, suggesting BACE2 may have a minimal role in neuronal APP cleavage.
37 While BACE1 is considered to be the functional β-secretase based on its role in
Aβ production, BACE2, on the other hand, does not exhibit putative β-secretase characteristics. Initially, in vitro studies found that BACE2 can cleave at the predicted β- secretase site (Asp1 of Aβ) and generate CTF-β and Aβ peptides (Hussain et al., 2000).
However, the primary cleavage position for BACE2 is within the Aβ sequence at
Phe19/20, which produces an α-secretase-like non-amyloidogenic peptide (Yan et al.,
2001a; Fluhrer et al., 2002). Intriguingly, the FAD APP A692G mutant flanking this preferred BACE2 cleavage position causes BACE2 to exhibit β-secretase-like activity in cells. Co-expression of BACE2 with this APP mutant in culture results in increased
CTF-β and Aβ levels (Farzan et al., 2000). Originally identified in a Flemish pedigree, individuals with this mutation at codon 692 of APP770 exhibit increased Aβ levels with a cerebral amyloid angiopathy, indicating this mutation shifts the preferred cleavage for
BACE2 toward an amyloidogenic pathway.
While BACE2 cellular expression in the trans-Golgi network matches the site of highest BACE1 activity, its proteolytic activity, nonetheless, precludes the generation of
Aβ. When BACE1 and BACE2 are co-expressed in cells with APP, secretion of CTF-α cleavage products is higher and Aβ levels are reduced compared to cells expressing only
BACE1 and APP (Fluhrer et al., 2002), suggesting BACE2 may function as an alternative
α-secretase. Based on its principal cleavage position, BACE2 antagonizes the effect of
APP processing by BACE1 (Basi et al., 2003) and may actually play a protective role in
AD pathogenesis by precluding the formation of Aβ. However, the expression pattern in the periphery argues that BACE2 may not regulate β-secretase cleavage in the brain.
38 Regulation of β-secretase
Although BACE2 may not alter BACE1 activity under normal conditions, the regulation of β-secretase cleavage represents an area of extensive investigation as a means to develop an AD therapeutic tool to regulate or inhibit Aβ production. BACE1 is regulated at the level of (1) transcription/alternative splicing, (2) post-translation, and (3) proteolytic activity. In addition to its regulation, several groups have identified other native BACE1 substrates as well as several unique BACE1 cleavage sites.
Understanding how BACE1 is regulated and its substrate specificity will provide further insight into its influence on AD pathogenesis.
Alternative Splicing and Transcriptional Regulation
BACE1 regulation at the level of transcription starts with the gene’s alternative splicing. When BACE1 was cloned, Northern blot analysis revealed the expression of three unique transcripts of different sizes in multiple tissues (Vassar et al., 1999), indicating alternative splicing of the gene product. Besides the brain, BACE1 is expressed at robust levels in the pancreas, but exhibits no proteolytic activity (Sinha et al., 1999). By reverse transcription-PCR, it was determined that a BACE1 variant is expressed in the pancreas. This variant expresses a differentially spliced isoform of
BACE1 that lacks a 132-bp fragment of exon 3 (Bodendorf et al., 2001). The spliced isoform results in a truncated peptide with a portion of its catalytic residues removed, explaining its inability to cleave APP (Bodendorf et al., 2001). Although this represents a unique inhibition of activity, the mechanism for this differential splicing pattern is not known.
39 A second form of BACE1 regulation is transcription factor binding. Mapping of the BACE1 promoter revealed a putative binding site for the general transcription factor,
Sp1 (Christensen et al., 2004). These studies demonstrated that Sp1 binding not only up- regulates BACE1 transcription, but also increases Aβ production (Christensen et al.,
2004). Transcriptional regulation by Sp1 binding plays a significant role in BACE1 protein expression and activity, which in turn dramatically affects β-secretase cleavage.
Post-Translational Processing
BACE1 exhibits common post-translational modifications, based on its protein sequence motifs and maturation through the cell. The BACE1 protein is synthesized with a pro-peptide on its N-terminus that is targeted for removal upon reaching the Golgi
(Capell et al., 2000). Removal of the 24 amino acid pro-peptide is primarily carried out by a pro-protein convertase, furin (Benjannet et al., 2001). The BACE1 protein also undergoes complex N-glycosylation at four sites and contains three disulfide bonds
(Haniu et al., 2000). These post-translational modifications help to explain the protein’s increase in size from an immature 50 kDa peptide to a mature 65 kDa protein as it is processed from the ER to the Golgi apparatus to the cell surface. While these modifications are necessary for the maturation and proper folding of BACE1, they are not essential for BACE1 activity. Upon mutation of the disulfide bonds in the catalytic domain, BACE1 is still able to process APP and generate β-secretase cleavage products
(Fischer et al., 2002). Furthermore, removal of the pro-peptide is not necessary for proper β-secretase activity (Benjannet et al., 2001). Based on a series of activity assays and detection of APP processing products, it was determined that these post-translational modifications help to target BACE1 protein to the trans-Golgi network and early
40 endosomal compartments, where it is most active (Huse et al., 2002). However, variants of BACE1 protein without the pro-peptide domain or disulfide bonds are present in the
ER, where BACE1 is still able to cleave APP and generate Aβ (Benjannet et al., 2001;
Fischer et al., 2002).
BACE1 Turnover
Another form of BACE1 regulation is its endoproteolysis and degradation.
BACE1 undergoes endoproteolysis, generating stable 37kDa N- and C-terminal fragments, which still remain associated in a complex stabilized by disulfide bonds (Huse et al., 2003). However, this endoproteolysis primarily occurs in peripheral tissues such as pancreas, liver, and muscle while the holoprotein remains intact in the brain (Huse et al.,
2003). The turnover of BACE1 has significant implications on its activity. BACE1 is known to have a long half-life, ~ t1/2 of 16 hours (Huse et al., 2000). Based on its localization with ubiquitin proteins, it was determined that the ubiquitin-proteasome pathway mediates the turnover of BACE1 in the cell (Qing et al., 2004). Blocking the ubiquitin-proteasomal pathway inhibits the degradation of BACE1 and as a result, significantly elevates the production of Aβ (Qing et al., 2004). Because of its effect on
Aβ production, the turnover of BACE1 may also have, in varying degrees, an impact on
AD pathogenesis.
Activity-Dependent Regulation
While there are certain minimal requirements for proper BACE1 function, specific cell biological events and protein binding partners are known to influence β- secretase cleavage of APP to generate Aβ. First, the unique structure of the BACE1 protein itself confers proper β-secretase function. Unlike most other aspartyl proteases,
41 BACE1 contains a type-I integral membrane domain. The presence of the transmembrane domain is critical for proper β-secretase function in vivo (Yan et al.,
2001b). Without the transmembrane domain, BACE1 is unable to be properly targeted to the late Golgi and endosomal compartments, where Aβ is predominantly generated (Xu et al., 1997). Like APP, trafficking of BACE1 significantly affects its activity, primarily by its spatial association with its substrates. As described above, APP participates in a complex transport system in neurons. Within polarized cells like neurons, cell-surface proteins are sequestered to either the basolateral portion of the plasma membrane, in the case of APP with its basolateral sorting signal, or to the apical plasma membrane for
BACE1 (Capell et al., 2002). When BACE1 is predominantly targeted to the apical surface, β-secretase cleavage and Aβ generation is significantly reduced (Capell et al.,
2002). Although BACE overexpression preferentially stimulates cleavage before this sorting pathway occurs, proper targeting of BACE1 within the cell regulates β-secretase activity and, accordingly, Aβ production.
Another form of BACE1 activity inhibition results from its interaction with Aβ- associated molecules, heparan sulfate and its structural analogue heparin. In addition to their interaction with Aβ, heparan sulfate proteoglycans (HSPGs) bind the active site of
BACE1 and inhibit β-secretase cleavage (Scholefield et al., 2003). Co-localizing with
BACE1 in the Golgi complex, heparan sulfate is thought to sequester APP from BACE1 by attaching an oliogosaccharide chain and prevent access to the β-secretase cleavage site. BACE1 activity inhibition also results from interaction with its binding partners, known as reticulon proteins (He et al., 2004). Identified as natural binding partners in
42 mammalian cells, reticulon proteins, namely neuronal-specific RTN3, block the access of
BACE1 to APP, resulting in reduced Aβ levels (He et al., 2004).
Conversely, the interaction of BACE1 with other proteins promotes Aβ production. It has been shown that a membrane lipid, ceramide, stimulates APP processing and Aβ generation (Puglielli et al., 2003). Known to be a second messenger and backbone of all complex sphingolipids, ceramide increases the steady-state levels of
BACE1 and prevents its degradation (Puglielli et al., 2003). Interestingly, ceramide levels are increased three-fold in the AD brain (Han et al., 2002). As a primary regulator of β-secretase function, ceramide may play a significant role in AD pathogenesis.
Because of its contribution to lipid structure, ceramide levels may also have a direct impact on another known risk factor for AD. Individuals with elevated cholesterol levels are at increased risk for AD (Ashford, 2004). Recent studies indicate cholesterol influences APP processing and Aβ generation. Increasing dietary cholesterol in APP transgenic mice results in significantly increased Aβ levels (Refolo et al., 2000).
Furthermore, BACE1 activity is enhanced by the addition of cholesterol (Riddell et al.,
2001). The mechanism for this phenomenon that has been proposed is the role lipid rafts may play in regulating APP processing. Lipid rafts are areas of the plasma membrane enriched in sphingolipids and cholesterol. The complex organization of these rafts can sequester specific membrane proteins within the fluid environment of the membrane’s liquid bilayer. Along with APP and PSEN1, BACE1 co-localizes with these lipid raft domains, targeting APP and its processing partners to lipid rafts resulting in enriched Aβ levels (Cordy et al., 2003; Ehehalt et al., 2003). These observations indicate the cellular
43 microenvironment influences amyloidogenic processing of APP by regulating the association of BACE1 and APP within the plasma membrane.
Other BACE1 Substrates
While most studies have focused on the role of BACE1 in the proteolysis of APP into Aβ, BACE1 is likely involved in the proteolysis of additional proteins. Because
Bace1 knockout mice do not exhibit any morphological or physiological abnormalities, it was thought that Aβ is the only substrate for BACE1. However, two different substrates have been identified, a Golgi-resident sialyltransferase (Kitazume et al., 2001) and P- selectin glycoprotein ligand-1 (Lichtenthaler et al., 2003). The regulation of sialyltransferases may be important in specific pathological conditions including inflammation. The cleavage of sialylα2,6galactose, ST6Gal 1, by BACE1 may be important for proper metabolism and secretion of glycoconjugates (Kitazume et al.,
2001). The other known substrate, P-selectin glycoprotein ligand-1 (PSGL-1), is involved in the inflammatory response in brain and peripheral tissues as well as the process of leukocyte recruitment, indicating BACE1 may play a role in a native immunological response (Lichtenthaler et al., 2003). While these studies demonstrate that BACE1 may have alternative functions, little is known about these processes beyond the cleavage by BACE1 of these substrates.
Alternative BACE1 Cleavage Specificity
BACE1 exhibits differential cleavage activity for its major substrate, APP. Co- expressing BACE1 and APP results in the generation of the 99 aa CTF-β as well as a smaller C-terminal fragment, consistent with a cleavage at Glu11 of Aβ (Vassar et al.,
1999). BACE1 cleavage followed by γ-secretase cleavage results in N-terminally
44 truncated Aβ peptides, namely Aβ11-40 or Aβ11-42/3 (see Figure 1-1). Upon removal of mouse Bace1, neither Aβ1-x nor Aβ11-x can be detected (Cai et al., 2001), indicating
BACE1 is the primary secretase for generation of both full-length and N-terminal truncated peptides. Further studies in the Bace1 knockout demonstrated potential species specificity for the generation of N-terminally truncated Aβ peptides. Expression of human APP in wild-type cells results in only a murine form of Aβ11-40, while only the expression of human BACE1 and APP in Bace1 negative cells results in a human form of
Aβ11-40 (Cai et al., 2001). It was postulated that human BACE1 can only cleave human
APP and mouse Bace1 can only cleave mouse App at Glu11 of Aβ. However, it was later demonstrated by mass spectrometry and fluorescence energy transfer (FRET) that human APP transgenic mice exhibit similar levels of both mouse and human N- terminally truncated Aβ peptides (Pype et al., 2003; Yang et al., 2004). These results suggest that although the murine and human Aβ sequence differ at the Aβ11-x cleavage position, the affinity for BACE1 cleavage may not differ based on species.
Nevertheless, the generation of truncated Aβ peptides may be important for understanding AD pathogenesis. Profiles of AD brains indicate that these truncated Aβ variants, particularly Aβ11-42, are prominent species in the amyloid plaques, while Aβ1-
40 peptides are present in both AD brain and normal elderly brains (Naslund et al., 1994).
In vitro studies indicate the N-terminally truncated Aβ peptides are more neurotoxic and aggregate into clumps more easily than full-length species (Pike et al., 1995), indicating these peptides are critical for amyloid deposition. Since so much attention has been devoted to the full-length Aβ1-40 or Aβ1-42/3 peptides, it is not known what role the truncated Aβ peptides play in AD pathogenesis. Cell culture studies have demonstrated
45 that BACE1 overexpression with Swedish mutant APP expression significantly increases the production of Aβ11-40 and Aβ11-42/3, finding that BACE1 can use full-length APP as well as CTF-β as a substrate (Liu et al., 2002). It was further demonstrated that these
N-terminal Aβ variants are present inside the cell, prior to secretion of Aβ, and are preferentially generated in the trans-Golgi network, the site of highest BACE1 activity
(Lee et al., 2003). The accumulation of intracellular truncated Aβ peptides may mark the pathologic event involved in secretion of Aβ and amyloid deposition.
In addition to N-terminal truncations, BACE1 cleavage may also generate C- terminal Aβ peptides. Mass spectrometry analysis of AD brain, particularly amyloid plaques, showed the presence of Aβ peptides terminating at residues 34 and 38 of the Aβ sequence (Pype et al., 2003). Recent studies suggest that BACE1 cleavage may generate
Aβ1-34 and Aβ1-38 in the AD brain. It has been shown that generation of these peptides depends on γ-secretase cleavage, particularly PSEN1 expression (Fluhrer et al., 2003; Shi et al., 2003). Again, it is not known what role these peptides play in AD pathogenesis, but these studies demonstrate the unique cleavage specificity for BACE1. While the studies of β-secretase have demonstrated the role BACE1 plays in APP processing and
Aβ production, little is known about the in vivo role of BACE1 on Aβ deposition and AD pathogenesis.
Research Aims
Utilizing in vivo analysis is undoubtedly one of the most powerful approaches to elucidating a biological phenomenon. Most importantly, transgenic studies can provide an experimental model for the phenomena, in this case, Alzheimer’s disease. The studies outlined in the current thesis have been to generate and characterize BACE transgenic
46 animals to gain better insight into the role these genes play in APP processing and, ultimately, develop a model for AD pathogenesis.
Genomic-based BACE Transgenics
This thesis will describe the genomic-based approach to develop β-secretase transgenic mice containing the human BACE1 and/or BACE2 genes. There have been many studies investigating transgenic mouse models for AD (Guenette and Tanzi, 1999) and most of these studies have utilized cDNA-based approaches for the human genes involved in AD. In particular, Hsiao et al. developed a transgenic mouse line containing the cDNA coding for the human APP695 with the Swedish double mutation linked with a hamster prion protein promoter and inserted into a cosmid vector (Hsiao et al., 1996). As one of the most widely known transgenic AD models (called Tg2576), the APP transgenic mice develop age-dependent memory impairments, significant increases in
Aβ1-42/3, and markedly increased amyloid plaques (Hsiao et al., 1996). While this transgenic approach provides a good model for AD, this approach has some disadvantages.
As with any transgenic approach, certain assumptions must be taken into account.
These assumptions are dependent on the transgene’s spatial and temporal expression, copy number, the presence of upstream and downstream regulatory elements, and chromosomal environment. The above cDNA-based approach utilizes a heterologous promoter, such as the hamster prion protein promoter, fused to the already-processed
APP cDNA, which can be expressed in specific cell types dictated by the prion promoter.
To generate a genetic model for AD pathogenesis, it is certainly more appropriate to generate a transgenic under the control of native transcriptional and chromosomal
47 regulatory elements. Towards these efforts, my studies in the current thesis describe how
I have been able to generate mice carrying the entire genomic copies of BACE1 and
BACE2, respectively, carried in bacterial artificial chromosome (BAC) clones. This approach makes fewer assumptions regarding the pathogenesis of the disease because the genes are under the control of their native regulatory elements (i.e. promoters, enhancers), are relatively low in copy number, and contain their native transcriptional and translational machinery due to the presence of the genomic sequence. In addition, this genomic-based approach ensures proper spatial and temporal expression allowing accurate assessment of the gene’s regulation in the pathogenic process. Certainly, this approach is a more suitable model for AD where little is currently understood regarding the temporal and spatial control of pathogenesis. This is the case for BACE1, particularly because very little is known about the in vivo regulation of BACE1 expression and activity as well as the regulation of APP processing throughout the lifespan. To this end, this approach has entailed mapping and characterization of the BACE1 and BACE2 genomic loci, identification of genomic clones containing each of the BACE genes, and purification of the genomic clones for introduction into the mouse.
Influence of β-secretase on APP Processing
By crossing BACE transgenics to mutant APP transgenics, I have sought to address several questions involving the role of the BACE genes in APP processing and
AD-related pathogenesis. What effect does expression of the human BACE1 or BACE2 genes have on APP processing? I have found that modulation of the β-secretase genes can dramatically alter the APP processing pathway in different brain regions and across all ages, from young to old. Analyzing the expression and regulation of the BACE genes
48 over the mouse lifespan, how does BACE influence the production of Aβ as animals age?
Through these studies, I have been able to determine the role BACE1 plays in altering Aβ generation (Chapters 2 & 3), while genomic expression of BACE2 does not influence production of Aβ (Chapter 4). These studies provide insight into the in vivo role the
BACE genes play in the regulation of the APP processing pathway.
Influence of β-secretase on AD Pathogenesis
This thesis also describes how β-secretase dictates Aβ production and deposition.
In particular, I have determined how BACE overexpression correlates with the formation of plaques and AD-like neuropathology in mice. Specifically, I have sought to understand how β-secretase expression modulates the APP processing pathway with respect to regional amyloid deposition. Knowing the development of AD pathology in humans, does human BACE expression influence a pathology profile that models the progression of pathology in human AD? Thus, by generating β-secretase transgenics, I have sought to develop an in vivo model examining the effects of specific alterations in
APP processing on AD-related phenotypes to reveal the complex pathogenesis of
Alzheimer’s disease.
49 Chapter 2: Altered Amyloid-β Metabolism and Deposition in Genomic-Based β- secretase Transgenic Mice
Authors: Matthew J. Chiocco1, Laura Shapiro Kulnane1, Linda Younkin2, Steve
Younkin2, Geneviève Evin3 and Bruce T. Lamb1
1Department of Genetics, Case Western Reserve University and University Hospitals of
Cleveland, Cleveland, OH 44106, USA
2Mayo Clinic Jacksonville, Jacksonville, FL 32224, USA
3Department of Pathology, University of Melbourne, Parkville 3010, Australia.
Reference: Chiocco, M.J., Kulnane, L.S., Younkin, L., Younkin, S., Evin G., and Lamb,
B.T., (2004) “Altered Amyloid-β Metabolism and Deposition in Genomic-Based β- secretase Transgenic Mice, “ Journal of Biological Chemistry 279, 52535-52542.
Acknowledgements: We thank S. Gandy for the generous gift of the 369 antibody and R.
Yan for the generous gift of B279 antibody. We would also like to thank Nicholas
Varvel (CWRU) for his help in preparing cryosections for immunohistochemistry.
50 ABSTRACT
Amyloid-β (Aβ), the primary component of the senile plaques found in Alzheimer’s disease (AD), is generated by the rate-limiting cleavage of Amyloid Precursor Protein
(APP) by β-secretase followed by γ-secretase cleavage. Identification of the primary β- secretase gene, BACE1, provides a unique opportunity to examine the role this unique aspartyl protease plays in altering Aβ metabolism and deposition that occurs in AD. The current experiments seek to examine how modulating β-secretase expression and activity alters APP processing and Aβ metabolism in vivo. Genomic-based BACE1 transgenic mice were generated that overexpress human BACE1 mRNA and protein. The highest- expressing BACE1 transgenic line was mated to transgenic mice containing human APP transgenes. Our biochemical and histochemical studies demonstrate that mice overexpressing both BACE1 and APP show specific alterations in APP processing and age-dependent Aβ deposition. We observed elevated levels of Aβ isoforms as well as significant increases of Aβ deposits in these double transgenic animals. In particular, the double transgenics exhibited a unique cortical deposition profile, which is consistent with a significant increase of BACE1 expression in the cortex relative to other brain regions.
Elevated BACE1 expression coupled with increased deposition provides functional evidence for β-secretase as a primary effector in regional amyloid deposition in the AD brain. Our studies demonstrate, for the first time, that modulation of BACE1 activity may play a significant role in AD pathogenesis in vivo.
51 INTRODUCTION
Alzheimer’s disease (AD) is a neurodegenerative disease characterized clinically by progressive cognitive impairment (Tanzi, 1999) and neuropathologically by the presence of senile neuritic plaques and neurofibrillary tangles within the brain. The primary constituent of the senile plaques is amyloid-β (Aβ) (Glenner and Wong, 1984), a peptide of 39-42 amino acids derived from the amyloid precursor protein (APP). Aβ deposition proceeds in a characteristic pattern within the brain with the appearance of plaques first in the basal neocortex, followed by deposition in the frontal cortex and hippocampal formation until all areas of the cortex contain deposits at end stage AD
(Braak and Braak, 1997b).
The generation of Aβ from APP involves three proteases with distinct activities, termed α-, β-, and γ-secretase. APP cleavage follows two pathways: cleavage by α- secretase generates C-terminal fragment-α (CTF-α), precluding the formation of Aβ upon subsequent γ-secretase cleavage. Alternatively, cleavage by β-secretase at Asp1 or
Glu11 of the Aβ sequence (Vassar et al., 1999) generates a unique C-terminal membrane- retained fragment, known as CTF-β. Subsequent cleavage of CTF-β by γ-secretase results in formation of Aβ (for review, see Selkoe, 1999). The Swedish FAD double mutation of APP (Mullan et al., 1992) appears to shift this cleavage pathway to favor processing by β-secretase, leading to a significant increase in Aβ production (Citron et al., 1992; Lamb et al., 1997).
The first APP secretase gene identified was that encoding β-secretase (BACE1)
(Hussain et al., 1999; Sinha et al., 1999; Vassar et al., 1999; Yan et al., 1999). BACE1, located on human chromosome 11q23.3, encodes a unique aspartyl protease with a single
52 transmembrane domain (Vassar et al., 1999). Overexpression of BACE1 in culture shows significantly increased β-secretase activity and Aβ levels in cells co-expressing
APP (Sinha et al., 1999; Vassar et al., 1999). Several studies have shown that BACE1 protein and activity levels are significantly increased in the AD (Fukumoto et al., 2002;
Sun et al., 2002; Yang et al., 2003), particularly the neocortex and hippocampus
(Fukumoto et al., 2002). BACE1 is the primary β-secretase involved in generation of Aβ in vivo, as shown by knockout experiments in mice (Cai et al., 2001; Luo et al., 2001;
Roberds et al., 2001). These studies are significant in considering therapeutics for Aβ elimination, but it is still unclear how alterations in BACE1 levels and activity may play a role in AD pathology in vivo.
To gain insight into how the modulation of BACE1 activity affects AD pathogenesis, we have utilized genomic-based transgenic methods to model AD pathology in vivo. Here we report the characterization of a human BACE1 transgenic line that produces human BACE1 mRNA and protein in mice. In addition, we show that animals expressing human BACE1 and human mutant APP have altered production of
APP C-terminal fragments and increased levels of Aβ peptides. Finally, we report that human BACE1 and human APP transgenic animals exhibit an altered brain-regional pattern of Aβ production and deposition, which reflects the relative levels of BACE1 protein in these brain regions. These studies demonstrate that overexpression of human
BACE1 alters the APP processing pathway and directly impacts the regional pattern of
Aβ deposition. Our findings suggest that modulation of human BACE1 expression and activity may play a significant role in AD pathogenesis.
53 MATERIALS AND METHODS
Animals and Genetic Crosses. All animals were handled according to official guidelines
(IACUC). Animals were bred on a mixed C57BL/6J x SJL background. BACE1 transgenic animals were crossed to APP transgenics (Tg2576), which contain the
Swedish mutant human APP cDNA (kindly provided by K. Hsiao, U. of Minnesota).
Progeny of this cross were sacrificed at 2-3 months of age for biochemical analysis and at
12 months of age for Aβ immunohistochemistry and age-dependent biochemical analyses.
Human BACE1 Genomic Analysis. Human BACE1 was mapped to human chromosome
11q23 (Vassar et al., 1999; Yan et al., 1999). Sequence-tagged sites (STSs) and gene markers on 11q23.3 were used to identify the precise genomic localization of human
BACE1. The human BACE1 cDNA (GenBank Accession #AF190725) was used to search the human genome by NCBI Blast (http://www.ncbi.nlm.nih.gov/BLAST) to identify genomic bacterial artificial chromosome (BAC) clones containing the full-length human BACE1 gene. In addition, BAC-end sequence (available from http://www.tigr.org/tdb/humgen/bac_end_search/bac_end_intro.shtml) and the published genome assembly (http://genome.cse.ucsc.edu/) were valuable resources in determining locus coverage for each clone. Sequencing was performed by Cleveland Genomics
(Cleveland, OH) if no end sequence was available. Genomic clones identified in silico were further verified by PCR with STSs A005A12, WI-7101, D11S1340, D11S939, and
BAC vector arm primers.
Isolation and Purification of BAC Clones. BAC clones 794Ill (Genbank Accession
#AP000761), 677N11 (Genbank Accession #AC020997), and 407N16 were obtained
54 from the human RPCI-11 library. Once BAC clones containing full-length BACE1 were identified, they were purified using the Clontech Nucleobond column (Palo Alto, CA).
To obtain a higher grade of ultra-pure DNA suitable for pronuclear microinjection (Yang et al., 1997), purified BAC DNA was passed through a CL4B Sepharose column
(Amersham Pharmacia Biotech, Uppsala, Sweden). The column was equilibrated with injection buffer (10mM Tris-HCl, pH 7.5; 0.1mM EDTA; and 100mM NaCl) and the
DNA collected in 12 elution fractions. The appropriate fraction containing the BAC was diluted to a concentration ranging between 0.6 ng/µl – 1.0 ng/µl.
Generation of BACE1 Transgenic Mice. Transgenic mice were generated by direct microinjection of BAC DNA into the pronuclei of fertilized mouse eggs. This method was performed as described (Hogan et al., 1994). The embryos injected were F2 progeny of a C57BL6/SJL F1 cross and were surgically transferred into the oviducts of a pseudopregnant CD-1 female. Founder mice transgenic for BAC clones corresponding to the human BACE1 locus were identified by PCR from mouse tail DNA. Genotyping markers to identify founders were the STS A005A12, which is located ~4 kilobases downstream of BACE1 exon 9, and a custom primer, which is located 10 kilobases upstream of BACE1 exon 1, with the sequence 5’- TGGAGAGTAATTTGCAATGCC-3’ and 3’- TTTGAATCCAAGGTTTTGCC-5’.
Southern Blot Hybridization. Genomic DNA was prepared from a human BACE1 PCR- positive transgenic mouse tail and a nontransgenic mouse tail using a standard salting out procedure (Miller et al., 1988). BAC 794Ill was purified as described above. Southern blots containing BamHI-digested genomic tail DNA (10µg) and BAC DNA (~ 100 ng) were prepared as described (Sambrook, 2001). Both 32P-labeled human BACE1 cDNA
55 and 32P-labeled human repetitive Alu element probes were used. The 1.1 kbp human
BACE1 cDNA specific to the 3’end of BACE1 (IMAGE clone 490377, GenBank
#AA136283) was obtained from Research Genetics (Invitrogen Life Technologies,
Carlsbad, CA). The 284 bp Alu element utilized was described previously (Lamb et al.,
o 1993). Hybridization was done at 65 with Church and Gilbert Solution (500mM NaPO4,
1mM EDTA, 1% bovine serum albumin, 7% SDS). Blots were washed with 0.2X SSC and 0.5% SDS at 65o.
RT-PCR Analysis. Total RNA was isolated from transgenic and nontransgenic mouse brain, pancreas, kidney, liver, colon, spleen, lung, testes, and ear by polytron homogenization of tissue in TRIzol Reagent (Invitrogen Life Technologies, Carlsbad,
CA). Each RNA preparation (2 µg) was reverse transcribed in the presence of 50 pmol random hexamer primers (Invitrogen Life Technologies, Carlsbad, CA) and Reverse
TranscriptaseTM (Invitrogen Life Technologies, Carlsbad, CA). After reverse transcription, 1/20 of the RT reaction was incubated in a PCR reaction with 50 pmol of primers with sequence 5’-GGAGGGAGCATGATCATTGG-3’ and 5’-
ACAGTCGTCTTGGGACGTGG-3’ for 20 cycles at an annealing temperature of 600.
These primers recognize both mouse and human Bace1/BACE1 transcripts and amplify a
446 bp product that spans exons 4-8 of the gene. PCR products were purified and then digested with BglII restriction enzyme, which digests the mouse Bace1 product uniquely into two fragments (324 and 122 bp in size), leaving human BACE1 product uncleaved.
For quantitation of total human BACE1 transcripts relative to mouse Bace1 transcripts in brain tissue, included in the PCR reaction was 50 ρmol exon8 primer, which was 5’ end-labeled with γ[32P] ATP and T4 polynucleotide kinase. After
56 determining the linear range of the PCR reaction between RT products and BACE1 PCR product, the last cycle (cycle 20) was spiked with γ[32P]ATP-BACE1-exon8 primer.
Resulting BglII PCR products were fractionated on 2% agarose gels, stained with ethidium bromide, photographed, and then dried and exposed to phosphorimager.
Tissue Preparation and Biochemical Analysis. Protein was isolated from 2-month old transgenic and nontransgenic mouse brain tissues by polytron homogenization in 1%
CHAPS solution in phosphate-buffered saline with 1x protease inhibitors [pepstatin, 1
µM; leupeptin, 4.5 µg/ml; aprotinin, 30 µg/ml; phenylmethanesulfonyl flouride (PMSF),
1 mM]. For analysis of BACE1 protein expression, twenty micrograms of total protein were resolved on 8% Tris-Glycine gels (Invitrogen Life Technologies, Carlsbad, CA).
Two antibodies were used to analyze BACE1 protein expression: (1) α-BACE—3599, a rabbit polyclonal antibody raised against amino acids 46-136 of human BACE1 (Cai et al., 2001) (kindly provided by P. Wong, PhD., Johns Hopkins U.) and (2) 00/06, a rabbit polyclonal antibody raised against amino acids 485-501 of human BACE1 (Holsinger et al., 2002).
For analysis of APP C-terminal fragments, twenty micrograms of total protein were resolved on 4-12% Bis-Tris gradient gels (Invitrogen Life Technologies, Carlsbad,
CA). Brain tissue extracts were blotted with 369, a polyclonal antibody raised against the
C-terminus of APP (kindly provided by S. Gandy, MD, PhD., Thomas Jefferson U.). All protein blots were probed with either anti-rabbit HRP-conjugated (Amersham
Biosciences, Piscataway, NJ) or peroxidase-labeled Protein A (KPL, Inc., Gaithersburg,
MD) secondary antibodies and detected by chemiluminescence (ECL, Amersham
Biosciences, Piscataway, NJ).
57 For analysis of soluble and insoluble protein pools, brain regions (cortex, hippocampus, cerebellum, olfactory bulb) from 12-month old animals were homogenized in 0.1M carbonate/ 50mM NaCl (pH 11.5) buffer with protease inhibitors (20 µg/mL aprotinin, 10 µg/mL leupeptin) as previously described (DeMattos et al., 2002). After homogenization, samples were centrifuged at 14,000 rpm for 20 minutes at 4oC. The supernatant was transferred to another tube and neutralized to pH 7.4 with 1M Tris, pH
6.8. The pellet was further homogenized in 5 M Guanidine-HCl in 50 mM Tris-HCl, pH
8.0 as previously described (Johnson-Wood et al., 1997).
Western Blot Quantitation. BACE1 protein expression was quantitated by comparing the amount of BACE1 protein in transgenic and nontransgenic brain tissue extracts relative to a standard curve of Bace1 expression. After incubation with ECL, the signal was captured with a fluorescence imager (Flour-S-Max Imaging Machine, Bio-Rad, Hercules,
CA) for each sample in triplicate and used to determine the fold increase of human
BACE1 protein expression. APP C-terminal fragments were quantitated by comparing the relative ratio of CTF-β to the total amount of C-terminal fragments generated for each animal. After exposing blots to film, the films were scanned and APP C-terminal fragment bands were quantitated by image densitometry using ImageQuant 1.20 software
(Molecular Dynamics).
Sandwich ELISA for Aβ. Brain extracts from 2-month old animals were analyzed for levels of Aβ peptides as previously described (Suzuki et al., 1994). Sandwich ELISAs were used with antibodies detecting the different species of Aβ ending in 40 aa.
Specifically, BNT-77/BA-27 was used to capture and detect Aβ peptides x-40. Brain region extracts from 12-month old animals (described above) were analyzed for levels of
58 Aβ peptides using Aβ1-40 ELISA (Biosource International, Camarillo, CA). Dilution ranges were determined for carbonate (soluble) and guanidine (insoluble) extracts and standards were serially diluted in the same sample dilution buffer containing protease inhibitors. The values were read using fluorometric plate reader Wallac 1420 multilabel counter (Perkin/Elmer, Wellesly, MA). Aβ1-40 concentrations were determined based on sample values relative to the serially diluted standards representing a standard curve of known Aβ1-40 concentration. Each sample was analyzed in triplicate and sample values are expressed as pmole Aβ40/gram brain tissue weight.
Immunohistochemistry. For Aβ immunostaining, half-brains were immersion fixed in
10% formalin and embedded in paraffin. The methods for Aβ immunostaining were performed as previously described by Kulnane and Lamb (Kulnane and Lamb, 2001).
Brain sections were stained with 6E10 (Senetek, Napa, CA), a mouse monoclonal antibody that is directed against Aβ1-17. For BACE1 immunostaining, animals were anesthetized with Avertin (0.02 cc/gm of body weight) and transcardially perfused with ice-cold phosphate buffer (0.1 M, pH 7.4) followed by 4% paraformaldehyde. After perfusion, whole brains were immersion-fixed in 4% paraformaldehyde for > 24h and then cryo-protected in 30% sucrose. Sagittal sections were cut on a sliding microtome and stored at –80oC. Brain sections were treated with avidin/biotin blocking kit (Vector
Laboratories, Burlingame, CA) and blocked in 10% normal goat serum (Sigma-Aldrich
Co., St. Louis, MO). Sections were stained with B279 rabbit polyclonal antibody (kindly provided by R. Yan, Ph.D., Cleveland Clinic Foundation), directed towards aa 295-310 of human BACE (Yan et al., 2001b), which shares identical sequence with mouse Bace1.
All sections were incubated in biotinylated goat anti-rabbit secondary antibody (Vector
59 Laboratories, Burlingame, CA). Sections were prepared according to standard methods using Vectastain ABC Kit (Vector Laboratories, Burlingame, CA) with the chromogen
3,3-diaminobenzidine tetrahydrochloride (DAB) and counterstained with hematoxylin.
For BACE1 immunofluorescence, cryosections were stained with B279 primary antibody
O/N at 4o C. For double immunofluorescence, cryosections were stained with B279
BACE1 antibody and mouse monoclonal antibody anti-NeuN MAB377 (Chemicon
International, Temecula, CA). After several washes, Alexa 488-conjugated goat anti- rabbit (for BACE1 signal) and Alexa 546-conjugated goat anti-mouse (for NeuN signal) secondary antibodies were added to the incubation buffer (Molecular Probes, Eugene,
OR) and mounted with Vecta-Shield Hard Set containing Dapi counterstain (Vector
Laboratories, Burlingame, CA).
Aβ Plaque Quantitation. Digital images of brain sections were captured with Qimaging
Micropublisher and Qcapture v2.64 (Burnaby, British Columbia Canada). Amyloid deposition was assessed in brain regions ranging from primary motor cortex, primary somatosensory cortex, primary and secondary visual cortex, and hippocampus. For quantitative analysis of Aβ deposition, five sections from each animal were analyzed with Image-Pro Plus v4.5 (Mediacybernetics Inc., Silver Spring, MD). The five sections spanned mouse brain Bregma locations from lateral 0.84mm to lateral 2.40mm (Paxinos,
2001). Plaque counts were determined following specific criteria for an immunopositive plaque, namely color, percent area, roundness, length, and width. Length and width criteria were fixed at minimal restrictions of > 5 µm per plaque. These analyses were performed blinded to genotype.
60 Statistical analysis. Two-tailed t tests were utilized for statistical analyses when variances for each sample population were equal. Two-tailed t test with Welch’s correction was used to adjust for unequal variances between sample populations. One- way ANOVA was utilized to compare BACE1 expression across brain regions. All data were analyzed by GraphPad Prism 2.0 statistical software (GraphPad Software, Inc., San
Diego, CA, 1998).
RESULTS
Identification of Human BACE1 Genomic Clones
Based on the mapping of BACE1 previously reported (Saunders et al., 1999), we were able to detect the position of the BACE1 locus on human chromosome 11q23.3
(Figure 2-1A). Examination of the genomic and published protein sequence indicated that BACE1 spans nine exons, extending across ~ 30,541 base pairs. Using the human genomic sequence available and known nearby genes on chromosome 11, we were able to identify several genomic clones with extensive coverage of the BACE1 genomic locus.
We used STS-mapping techniques with four sequence-tagged sites (STSs) to pinpoint the extent of genomic clone coverage for BACE1. Further analysis of the BAC Fingerprint database (http://genome.wustl.edu/projects/human) and the genome survey sequence
(GSS @ NCBI) database revealed three BAC clones that localized to the BACE1 gene region, BACs 794Ill (Genbank Accession #AP000761), 677N11 (Genbank Accession
#AC020997), and 407N16 (Figure 2-1A).
BAC 794Ill contains ~ 45 kilobases upstream of the ATG start site for BACE1 and ~ 99 kilobases downstream of the BACE1 polyA tail. BACs 677N11 and 407N16
61 contains ~ 40 kilobases upstream of the 5’end of BACE1. These BACs also contain ~ 38 kilobases downstream of the 3’end of BACE1 (shown in Figure 2-1A). Thus, these three clones, which cover the entire BACE1 locus, were determined suitable to generate
BACE1 transgenic mice.
Expression and Activity of Human BACE1 in Mice
Upon pronuclear microinjection of each BACE1 BAC clone, we recovered at least one founder mouse for each clone (Figure 2-1A). Each of these transgenic lines was analyzed for construct integration, copy number, and BACE1 expression levels. As our previous studies indicated that genomic-based transgenes exhibited copy number dependence (Lamb et al., 1993), we chose to focus on transgenic line 794Ill-3 in the current study as this line exhibited the highest copy number and level of BACE1 expression, as described below. To confirm the integrity of the BAC in PCR-positive animals for line 794Ill-3, Southern blot hybridization with the human Alu repeat probe revealed that 794Ill-3 transgenics exhibit an identical human-specific Alu profile to that of the purified BAC clone (Figure 2-1C). Furthermore, Southern blot hybridization with a human BACE1 cDNA specific for the 3’ end of BACE1 revealed the expected 12.4kbp band in animals transgenic for BAC794Ill when digested with BamHI (Figure 2-1B).
These results confirm BAC 794Ill has integrated intact in this transgenic line.
The level of human BACE1 mRNA expression in transgenic line 794Ill-3 was analyzed using an RT-PCR assay with conserved oligonucleotides that amplify both mouse and human BACE1 transcripts across exons 4-8. The resulting products were digested with a restriction enzyme that cleaves the mouse products and leaves the human products undigested. This RT-PCR analysis revealed the co-expression of mouse and
62
Figure 2-1: Generation of genomic-based human BACE1 transgenic mice.
A, (top) Genomic map of the BACE1 locus with the relative position of the BACE1 gene (as represented by the STS A005A12), nearby genes and STSs, and corresponding BAC clones on chromosome 11q23.3. (bottom) Human BACE1 transgenics recovered by pronuclear microinjection. Founders were recovered for three different BAC clones shown at right. Plus (+) signs indicate PCR-positive animals for the corresponding STSs and primers at the chromosome 11 locus. B and C, Genomic DNA from BACE1 transgenic (Tg), nontransgenic littermate (NT), human genomic (Hu), and/or 794Ill BAC DNA was digested with BamHI, run on agarose gel, transferred to nylon membrane and hybridized with a 32P-labeled human BACE1 cDNA (B) and human Alu repetitive element (C). On the left are shown the sizes of molecular weight markers in kbp.
63 human Bace1/BACE1 in multiple tissues including brain, pancreas, lung, liver, colon, and spleen (Figure 2-2A).
To quantify the level of mRNA expression in transgenic line 794Ill-3, γ-32P- labeled oligonucleotides were added in the linear range of the PCR reaction for the final cycle. The amount of undigested human BACE1 products relative to digested mouse
Bace1 products were based on densitometry upon exposure to a phosphorimager. After
PCR amplification of the BACE1/Bace1 cDNA and restriction enzyme digestion, we determined that human BACE1 mRNA in brain is expressed ~ 4.7-fold higher than mouse
Bace1 (two-tailed t test, p value = 0.017) (Figure 2-2B).
To determine the level of human BACE1 protein levels in transgenic line 794Ill-
3, Western blot analysis of brain extracts revealed higher protein expression in BACE1 transgenic animals compared to nontransgenic littermates, as shown by the dark smear corresponding to the ~70kD human BACE1 protein in transgenic brain tissue extracts for both a C-terminal antibody (Figure 2-2C) and an N-terminal antibody (data not shown).
To quantify the level of human BACE1 protein overexpression, we used the polyclonal antibody specific for the C-terminus of BACE1, which recognizes both mouse, and human BACE1 (Holsinger et al., 2002). Quantification was performed by comparing
BACE1 transgenic and nontransgenic brain extracts to a standard curve of nontransgenic protein levels using a fluorescence imager. We determined BACE1 transgenic mice express approximately 2-fold higher BACE1 protein compared to nontransgenic controls
(two-tailed t test, p value < 0.0001) (Figure 2-2D).
To characterize the regional localization of BACE1 expression in the brain, immunohistochemical staining with B279 antibody (kindly provided by R. Yan) indicates
64
Figure 2-2: Human BACE1 mRNA and protein expression and quantitation.
A, RT-PCR analysis with conserved primers corresponding to sequences identical between mouse and human Bace1/BACE1 mRNAs was performed on reverse-transcribed RNA from brain, kidney, liver, heart, colon, spleen, lung, testes, and ear from nontransgenic (NT) and human BACE1 transgenic (Tg) mice. The resulting 446 bp products were digested with BglII to generate 324 and 122 bp mBace fragments, while the human products (hBACE) remain undigested. Left, approximate sizes in bp. B, Relative levels of human BACE1 mRNA: mouse Bace1 mRNA in brain. Human BACE1 mRNA is expressed 4.7-fold higher than mouse Bace1 mRNA (two-tailed t test, p value = 0.017, error bars represent S.E.). C, Western blot analysis of mouse and human BACE1 in 1% CHAPS brain extracts. Twenty micrograms of total protein from eight week-old nontransgenic control (Ctl), BACE1 transgenic (Tg), and mouse Bace1 KO mice were run on 8% Tris-glycine gels, transferred to polyvinylidene difluoride membrane, and blotted with C-terminal antibody BACE-00/6. On the left for each gel are sizes of molecular weight markers in kilodaltons. D, BACE1 protein expression was quantitated by comparing the amount of BACE1 protein in transgenic (n=9) and nontransgenic (n=9) brain tissue extracts relative to a standard curve of Bace1 expression using a fluorescence imager for capture of chemiluminescent signal. Human BACE1 is expressed ~2-fold higher than control (two-tailed t test, p value < 0.0001).
65 the pattern of BACE1 expression is primarily in the mitral cell layer of the olfactory bulb
(Figure 2-3A and 2-3B) and cortical layers (Figure 2-3E and 2-3F) with low-level staining in the hippocampal neurons (Figure 2-3C and 2-3D) and Purkinje cells of the cerebellum (data not shown). Double immunofluorescence with anti-mouse NeuN antibody and BACE1 B279 antibody confirms co-localization of human BACE1 in mouse neurons (data not shown). Since this antibody recognizes both mouse Bace1 and human BACE1, DAB chromogen staining and immunofluorescent staining display increases of BACE1 expression in the transgenic animal (Figure 2-3B, D, F) compared to the nontransgenic control (Figure 2-3A, C, E). The neurons exhibiting the highest expression are present in the outer layers of the primary and secondary motor cortices
(Paxinos, 2001), displaying punctate cellular staining (Figure 2-3G and 2-3H), consistent with subcellular localization to Golgi compartments and early endosomes (Yan et al., 2001a; Kinoshita et al., 2003).
Human BACE1 and APP Processing
To determine the impact of BACE1 overexpression on APP processing, we crossed the human BACE1 transgenics to human APP transgenic line Tg2576. These
APP transgenics express ~ 6-fold higher human APP protein than that of endogenous mouse App and exhibit significantly higher levels of both Aβ1-40 and Aβ1-42(3) peptides (Hsiao et al., 1996). We first analyzed the levels of APP processing products, namely CTF-β and CTF-α. Brain extracts from 2-month-old progeny of BACE1 crossed to Tg2576 were analyzed by Western blot (Figure 2-4A). The relative ratio of CTF-β to the total amount of C-terminal fragments generated for each animal was determined with each extract run in triplicate. We determined that animals transgenic for both BACE1 and
66
Figure 2-3: Human BACE1 immunohistochemical localization.
Brains from 2-month old BACE1 transgenic (B, D, F) and nontransgenic (A, C, E) animals were cryo-protected in 30% sucrose after perfusion with 4% paraformaldehyde. Sagittal sections were analyzed by staining with polyclonal antibody B279 followed by staining with biotinylated anti-goat secondary antibody (A-F) or Alexa488-conjugated secondary antibody (G and H). Positive immunoreactivity is present in the olfactory bulb (A and B), hippocampus (C and D), and frontal cortex (E and F). Scale bar in F and H = 200 µm.
67 Tg2576 express significantly higher C-terminal fragment ratios compared to Tg2576 transgenics alone (two-tailed t test, p value = 3.00x10-4) (Figure 2-4B).
To examine the impact of BACE1 overexpression on Aβ metabolism, the levels of Aβ peptide in whole brain extract was measured by ELISA (Figure 2-4C). Animals transgenic for both BACE1 and Tg2576 at two months of age exhibit significantly higher levels of Aβ1-40 (two-tailed t test, p value = 4.00x10-3).
Human BACE1 and Aβ Deposition
To determine the effect of BACE1 overexpression on Aβ deposition and AD-like neuropathology, the amount of Aβ deposition in brain sections was measured by immunostaining with Aβ antibodies. The APP transgenics, Tg2576, develop Aβ deposits at approximately 9-12 months of age (Hsiao et al., 1996). Immunostaining with an Aβ- specific antibody at six and ten months of age, we observe no differences in Aβ deposition and plaque profile between the double transgenics compared to APP transgenics alone (data not shown). However, at 12 months of age, animals transgenic for both BACE1 and the Swedish mutant APP exhibit an altered Aβ plaque profile compared to mutant APP transgenics alone (Figure 2-5). More specifically, the double transgenics exhibit a significantly higher number of immunoreactive Aβ deposits in the frontal cortex compared to the single transgenics (two-tailed t test with Welch’s correction, p value = 0.0252) (Figure 2-5G). However, while the double transgenics exhibit a unique profile of Aβ deposition in the frontal cortex (Figure 2-5A and 2-5B) and olfactory bulb (Figure 2-5C and 2-5D), there are only slight increases in hippocampal Aβ deposition (Figure 2-5E and 2-5F) between the two groups (two-tailed t test, p value = 0.126; data not shown).
68
Figure 2-4: APP processing and Aβ metabolism in BACE1 transgenics.
A, Western blot analysis of APP holoprotein (top) and APP C-terminal fragments (bottom) in 1% CHAPS brain extracts. Twenty micrograms of total protein from ~2 month-old progeny of BACE1 transgenics crossed to Tg2576 (APP cDNA) transgenic animals were run on 4-12% Bis-tris gradient gels, transferred to polyvinylidene difluoride membrane, and blotted with APP C-terminal polyclonal antibody 369. On the left are sizes of molecular weight makers in kilodaltons. B, Quantitation of relative band intensities of APP C-terminal fragments using chemiluminescence with each brain extract analyzed in triplicate. The ratio of CTF-β: (CTF-β+CTF-α) was compared in animals transgenic for both BACE1 and Tg2576 (n = 6) to animals transgenic for Tg2576 alone (n= 7). Double transgenics had significantly higher APP CTF ratios by two-tailed t test, p value = 0.0003. C, Brain extracts from ~ 2-month old animals were analyzed by ELISA. BACE1/Tg2576 animals (n= 5) had significantly higher levels of Aβ1-40 compared to Tg2576 alone (n= 5) by two-tailed t test, p value = 0.0040.
69 Figure 2-5: Analysis of Amyloid-β deposition in BACE1 transgenic animals.
Brains from 12-month old animals were fixed in 10% formalin, embedded in paraffin, and sectioned on a sagittal plane (10 µm thick). Scattered sections were analyzed by staining using standard protocols (Kulnane et al., 2001) with mAb 6E10, which detects amino acids 1-17 in the Aβ region. Brain sections of animals transgenic for both BACE1 and Tg2576 in cortex (A), olfactory bulb (C), and hippocampus (E). Brain sections of animals transgenic for Tg2576 alone in cortex (B), olfactory bulb (D), and hippocampus (F). A similar pattern of Aβ immunostaining was observed in all double BACE1/Tg2576 transgenics and single Tg2576 transgenics. Scale bar in F = 200 µm. G, Immunopositive plaques were counted in sections (5 sections/animal) from each animal transgenic for BACE1 and Tg2576 (n= 6) as well as Tg2576 alone (n= 6) using Image-Pro Plus (Mediacybernetics, Silver Spring, MD) to identify Aβ deposits in each brain region. The average plaque number is significantly higher in cortical brain regions of the double transgenics compared to single transgenics (**two-tailed t test with Welch’s correction, p value = 0.0252, error bars represent SEM), while hippocampal deposits are not statistically different between both groups.
70 To further explore the impact of BACE1 expression on Aβ deposition in brain regions, the levels of Aβ peptide for both soluble and insoluble protein pools in four specific brain regions were measured by ELISA at twelve months of age (Figure 2-6).
Brain region extracts were prepared using a sequential extraction procedure previously described (DeMattos et al., 2002). In soluble (carbonate-extracted) protein pools, animals transgenic for both BACE1 and Tg2576 exhibit significantly higher levels of
Aβ1-40 in the olfactory bulb (p value = 1.63x10-2) and cortex (p value = 3.40x10-3) relative to single Tg2576 transgenics (Figure 2-6A). In insoluble (guanidine-extracted) protein pools, double transgenics exhibit significantly higher levels of Aβ1-40 in the olfactory bulb (p value = 1.57x10-2), hippocampus (p value = 4.00x10-4), and strikingly higher levels in the cortex (p value = 1.60x10-3) compared to Tg2576 transgenics alone in each of these brain regions (Figure 2-6B).
To examine whether the level of human BACE1 expression in different brain regions might be responsible for the altered pattern of Aβ deposition, we analyzed the amount of human BACE1 protein expression by Western blot. To quantify the relative amount of BACE1 protein expression in brain regions at 12 months of age, we compared brain region extracts from BACE1 and Tg2576 transgenic mice to a standard curve of transgenic protein amounts. Similar to the pattern observed from BACE1 immunohistochemistry at young ages (Figure 2-3), we determined that olfactory bulb
BACE1 expression in the aged transgenic mice is significantly higher than BACE1 expression in the cortex, hippocampus, and cerebellum (One-way ANOVA, p value <
0.001) (Figure 2-6C).
71
Figure 2-6: Amyloid-β production and BACE1 protein expression by brain region.
A and B, Brain region extracts from 12 month old progeny of BACE1 transgenics crossed to Tg2576 transgenics were made using a sequential extraction procedure (see Methods) and analyzed by ELISA. A, Soluble Aβ1-40 levels by brain region. BACE1/Tg2576 animals had significantly higher levels of Aβ in olfactory bulb (#: two-tailed t test, p value = 0.0163) and cortex (*: two-tailed t test, p value = 0.0034) compared to Tg2576 animals. B, Insoluble Aβ1-40 levels by brain region. BACE1/Tg2576 animals had significantly higher levels of Aβ in olfactory bulb (**: two-tailed t test, p value = 0.0157), hippocampus (: two-tailed t test, p value = 0.0004), and cortex (*: two-tailed t test, p value = 0.0016) compared to Tg2576 animals. C, BACE1 protein expression was quantitated using Western blot analysis by comparing the amount of BACE1 protein in cerebellum, olfactory bulb, hippocampus, and cortex tissue extracts (n=7 for each brain region) relative to a standard curve of BACE1 expression using a fluorescence imager to capture chemiluminescent signal. Human BACE1 expressed in olfactory bulb is significantly higher than cortex, hippocampus, and cerebellum (ANOVA, p value < 0.001), while cortical BACE1 expression is significantly higher than hippocampus (ANOVA, p value < 0.05) and cerebellum (ANOVA, p value < 0.001).
72 Cortical BACE1 expression is significantly elevated compared to hippocampal and cerebellar BACE1 expression (ANOVA, p value < 0.05 and p value < 0.001, respectively). Thus, the increase in Aβ production and deposition within the olfactory bulb and cortical layers of double transgenics is correlated with elevated levels of
BACE1 expression in the olfactory-cortical brain regions.
DISCUSSION
The identification of BACE1 has paved the way for a better understanding of the function of β-secretase, a protein long to be considered one of the major players involved in the development of Alzheimer’s disease neuropathology. Previous findings in cell culture have demonstrated how BACE1 alters the APP processing pathway, resulting in significantly increased levels of Aβ peptides (Sinha et al., 1999; Vassar et al., 1999; Yan et al., 1999). Complete elimination of Aβ peptides in Bace1 knockout mice indicates
BACE1 is the primary β-secretase involved in generation of Aβ (Cai et al., 2001; Luo et al., 2001; Roberds et al., 2001), demonstrating the significance of using β-secretase as a potential therapeutic target for Alzheimer’s disease.
The genetic significance for BACE1 in human AD remains elusive. While several studies have reported no significant linkage of BACE1 with Alzheimer’s disease
(Nicolaou et al., 2001; Liu et al., 2003), only weak associations have been identified for several polymorphisms in the BACE1 gene (Nowotny et al., 2001). More recently, Gold et al. (Gold et al., 2003) report a significant association of a BACE1 polymorphism for late-onset Alzheimer disease in Apolipoprotein ε4 carriers. Another group (Clarimon et al., 2003) found a significant association in exon 5 of BACE1 in AD subjects, even in
73 individuals not carrying the ε4 allele of apolipoprotein E gene. While a clear genetic association of BACE1 with Alzheimer’s disease has yet to be confirmed, BACE1 is hypothesized to play a major role in AD pathogenesis.
Transgenic studies will provide insight into the role of BACE1 in AD pathology.
Utilizing a cDNA-based approach, two transgenic models have been developed for
BACE1, finding that overexpression of BACE1 in neurons increases the steady-state levels of Aβ peptides (Bodendorf et al., 2002) and can accelerate amyloid deposition
(Mohajeri et al., 2004). This approach makes several assumptions by using a heterologous promoter, which drives expression in a specific temporal and spatial pattern.
In the current study, we have utilized a genomic-based approach to generate human
BACE1 transgenics with genomic clones corresponding to the entire human BACE1 gene.
A genomic-based approach makes fewer assumptions in vivo because BACE1 is under the control of its native regulatory elements and contains its native transcriptional and translational machinery due to the presence of genomic sequence. This approach ensures proper spatial and temporal expression allowing accurate assessment of BACE1’s role in
Aβ production and AD pathology.
Here, we demonstrate that genomic-based human BACE1 transgenic animals express elevated levels of human BACE1 mRNA and protein. We observe that BACE1 expression is localized primarily to olfactory bulb neurons, cortical layers, and hippocampal formation, matching the regional distribution of APP (Marcinkiewicz and
Seidah, 2000; Irizarry et al., 2001). When crossed to human mutant APP transgenics, the
BACE1 and APP double transgenics exhibit increased levels of APP C-terminal fragments and corresponding Aβ peptides as early as two months of age, indicating that
74 modulation of BACE1 activity alters the APP processing pathway in vivo. These results are consistent with the other BACE1 transgenics previously described (Bodendorf et al.,
2002; Mohajeri et al., 2004). However, unlike Bodendorf et al., we were unable to observe any alterations in endogenous mouse App processing or wild-type human APP processing (data not shown).
Most notably, we report, for the first time, that human BACE1 and human APP transgenic animals exhibit an altered Aβ plaque deposition profile. We do not observe accelerated amyloid pathology as reported by Mohajeri et al.; rather we observe dramatic increases of amyloid deposition at the previously reported age of 12 months (Hsiao et al.,
1996). Not only are Aβ deposits significantly increased relative to controls, but the human BACE1 and human APP double transgenics also exhibit a unique brain-regional deposition profile. We report increased cortical Aβ deposits upon BACE1 overexpression, while hippocampal deposits are not significantly elevated. This unique deposition pattern is corrobated by the profile of Aβ production across brain regions in the aged APP and BACE1 double transgenics. In particular, soluble and insoluble Aβ is dramatically increased in the cortex and olfactory bulb, while hippocampal Aβ is increased only in the insoluble fractions, presumably as amyloid deposits begin to accumulate in the hippocampus. Increasing evidence suggests that the equilibrium between soluble and insoluble Aβ peptides in the brain reflects the progression of AD- like pathogenesis (Kawarabayashi et al., 2001). Based on our studies, human BACE1 overexpression appears to trigger a switch from elevated hippocampal Aβ production and deposition normally observed in mutant APP transgenics (Tg2576) (Hsiao et al., 1996) to increased cortical Aβ production and deposition, suggesting BACE1 overexpression
75 might control the brain-regional deposition pattern and AD-like neuropathology. Our studies provide functional evidence for the potential significance of BACE1 in AD pathogenesis.
Based on clinical human studies, the development of AD neuropathology progresses in a distinct pattern that correlates with age (Braak and Braak, 1997b). This pattern is characterized by initial amyloid deposition in neocortical regions of the AD brain followed by spread of deposition to the hippocampus and all cortical areas (Braak and Braak, 1997b). Numerous factors likely control the progression of amyloid deposition within the AD brain.
While it remains unclear how brain-regional amyloid deposition occurs within the
AD brain, several studies suggest that one of the controlling factors could be the expression and activity of the major players involved in APP processing and Aβ generation. Based on studies by Lehman et al. (Lehman et al., 2003), the abundance of
APP holoprotein and the APP processing products varies by brain region in the Tg2576
APP transgenics. Both CTF-β, the precursor to Aβ, as well as Aβ itself are higher in the hippocampus relative to cortex, olfactory bulb, and cerebellum in these animals (Lehman et al., 2003). Furthermore, the level of mRNA expression for BACE1, the primary β- secretase, in Tg2576 transgenics is brain-region specific, with highest levels in the hippocampus, dentate gyrus, and cerebellar molecular layer (Irizarry et al., 2001). These findings differ, however, with the studies of Zohar et al. (Zohar et al., 2003) in which the levels of BACE1 mRNA are higher in the frontal cortex relative to hippocampus in normal rat and human brain.
76 In our human BACE1 and mutant APP double transgenics, the unique cortical- specific deposition profile is reflective of elevated BACE1 expression in the olfactory bulb and cortex relative to other brain regions, particularly cerebellum and hippocampus.
The relationship of up-regulated BACE1 protein expression and AD is consistent with previous findings, which report increased neocortical BACE1 expression and activity in the AD brain (Fukumoto et al., 2002; Holsinger et al., 2002) and is correlated with Aβ load in AD (Li et al., 2004). Particularly noteworthy is the high level of BACE1 expression in the olfactory bulb accompanied by significantly elevated levels of soluble and insoluble Aβ peptides. Olfactory deficits have long thought to be an early marker for cognitive decline in AD patient populations (Doty et al., 1987; Serby et al., 1991; Serby et al., 1996). In recent years, several groups have identified AD pathologic changes in olfactory structures in early stages of AD (Kovacs et al., 2001; Christen-Zaech et al.,
2003). Furthermore, these pathologic changes in the olfactory system are associated with the increasing severity of cortical pathology (Kovacs et al., 2001; Christen-Zaech et al.,
2003). Based on olfactory bulb anatomy, it is well known that olfactory neurons project to the entorhinal cortex, which is a component of the hippocampal formation, and indirectly to the frontal cortex (Kandel et al., 2000). This is particularly the case in the rodent brain in which small areas of the cortex receive afferents from the olfactory bulb
(Kovacs, 2004). In addition, recent evidence that BACE1 and APP are axonally transported through nerve terminals in the brain indicates that Aβ production and deposition is a dynamic process (Kamal et al., 2001; Sheng et al., 2003), resulting in the selective disruption of neuronal connections in which cortico-cortical circuits are most vulnerable (Delatour et al., 2004). These findings together with the neurobiology of the
77 olfactory system could help to explain our studies in which the brain regions with highest
BACE1 expression, the olfactory bulb and cortex, affect such a dramatic elevation in Aβ production and deposition.
Our findings indicate that the level of expression and activity of the major player in Aβ generation, BACE1, may be a major factor in the regional pattern of Aβ deposition. These studies demonstrate that BACE1 alters the APP processing pathway and directly impacts the development of AD pathology in vivo. Further studies will examine both the expression of BACE1 as well as the production of Aβ in specific brain regions for the human BACE1 and APP double transgenics across a variety of ages from young to old. These experiments are critical to further understand the relationship between Aβ metabolism and the deposition of Aβ throughout the brain.
78 Chapter 3: Spatial and Temporal Control of Age-related APP Processing in
Genomic-based β-secretase Transgenic Mice
Authors: Matthew J. Chiocco1 and Bruce T. Lamb1
1Department of Genetics, Case Western Reserve University and University Hospitals of
Cleveland, Cleveland, OH 44106, USA
Reference: Will be submitted for publication
Acknowledgements: We thank S. Gandy for the generous gift of the 369 antibody, G.
Evin for the generous gift of the 00/6 antibody, and R. Yan for the generous gift of B279 antibody. We would also like to thank Nicholas Varvel (CWRU) for his help in preparing cryosections for immunohistochemistry.
79 ABSTRACT
Genetic mutations associated with Alzheimer’s disease (AD) in the Amyloid Precursor
Protein (APP) gene specifically alter the production of the APP processing product, amyloid-β (Aβ) peptide, generated by β- and γ-secretases. The accumulation and deposition of Aβ is hypothesized to cause AD pathogenesis, leading to the debilitating neurological deficits observed in AD patients. However, it is unclear how processing of
APP to generate Aβ corresponds with the age-dependent pattern of brain-regional neurodegeneration common in AD. We have previously shown that overexpression of
BACE1, the primary β-secretase gene, in mice expressing an AD mutant form of APP leads to significantly elevated regional Aβ levels, which coincide with the regional pattern of Aβ deposition. In the current study, we have used our genomic-based β- secretase transgenic mice to determine how BACE1 regulates the spatial and temporal pattern of Aβ production throughout post-natal development. Specifically, we observed unique differences in the brain-regional expression pattern between neonatal and adult
BACE1 transgenic mice. These alterations in the BACE1 expression profile directly corresponds with age-related differences in regional Aβ production and deposition.
These studies indicate that modulation of BACE1 expression leads to dramatic alterations in APP processing and AD-like neuropathology. Furthermore, our studies provide further evidence that BACE1 plays a major role in the regulation of the APP processing pathway, influencing the age-dependent onset of AD pathogenesis.
80 INTRODUCTION
Amyloid-β (Aβ) peptides, the primary constituent of senile plaques in
Alzheimer’s disease (AD), are generated from the sequential cleavages of Amyloid
Precursor Protein (APP) by β- and γ-secretases. The generation and deposition of Aβ peptides in the AD brain is hypothesized to be the causative phenomenon in AD pathogenesis. This hypothesis, known as the amyloid cascade, is corroborated genetically by the presence of familial AD mutations in the APP gene (Goate et al., 1991;
Hendriks et al., 1992; Mullan et al., 1992) that shift the processing of the APP holoprotein towards β-secretase cleavage, resulting in the increase of Aβ levels accompanied by a decrease of the non-amyloidogenic and non-pathogenic p3 peptide generated by the other APP holoprotein secretase, known as α-secretase (Citron et al.,
1992; Haass et al., 1994). The accumulation and aggregation of Aβ progresses in an age- related brain-regional pattern (Braak and Braak, 1997b), but it is unknown how this pattern of age-dependent Aβ production and deposition is regulated. Thus, the spatial and temporal control of APP expression and its processing in the brain may have significant implications on Aβ production and deposition.
Expression studies have shown that APP expression in the rodent brain is increased in early embryonic and neonatal animals relative to adult animals, where specific brain regions, particularly the hippocampus and cortex, exhibit the highest expression (Loffler and Huber, 1992; Marcinkiewicz and Seidah, 2000). Moreover, the coordinated expression of α-secretase, β-secretase, and APP throughout the brain
(Marcinkiewicz and Seidah, 2000) indicates the significance for strict regulation of APP
81 processing. Thus, the competition for the APP holoprotein substrate by β- and α- secretases is a critical step in Aβ production and, potentially, the pathogenesis of AD.
Modulation of the activity of these APP secretases will provide insight into the competitive balance of the APP processing pathway. With the identification of BACE1
(Sinha et al., 1999; Vassar et al., 1999; Yan et al., 1999), it was determined that β- secretase cleavage can be, in fact, the rate-limiting event for Aβ production. In addition, overexpression of BACE1 in APP transgenic mice results in increased β-secretase cleavage products particularly increased Aβ levels (Bodendorf et al., 2002; Chiocco et al., 2004; Mohajeri et al., 2004) as well as altered amyloid deposition pathology (Chiocco et al., 2004; Mohajeri et al., 2004; Willem et al., 2004). Specifically, we have previously reported an altered brain-regional deposition profile in our double BACE1xAPP transgenic mice. Intriguingly, the brain regions exhibiting the most severe amyloid-β pathology correspond with both significantly elevated Aβ levels and highest BACE1 protein expression (Chiocco et al., 2004). Our previous studies implicate BACE1 as a regulator of brain-regional Aβ production and deposition in vivo.
To further understand how BACE1 regulates APP processing in vivo, we have analyzed our genomic-based BACE1 transgenic mice for age-dependent and regional alterations in Aβ production. Our previous studies, comparing the cDNA-based Tg2576 and our genomic-based R1.40 transgenic, have demonstrated that brain-regional processing of the human Swedish mutant APP transgene differs by transgene type
(Lehman et al., 2003). Furthermore, as we have previously shown, characterizing human
APP expression and processing under the control of its native regulatory elements
82 removes biases involving the spatial and temporal regulation of the APP transgene (Lamb et al., 1997).
Here, we report the unique profile of BACE1 expression and brain-regional Aβ production in early post-natal and adult animals. We observe altered brain-regional levels of BACE expression and APP expression, which may control the production of Aβ by brain region at both post-natal ages and adult ages. Although Aβ levels are higher in young animals, the brain-regional and age-dependent increase of Aβ in adult animals is due to elevated regional BACE1 expression, particularly in the cortex. Most notably, increased human BACE1 expression corresponds with regional alterations in amyloid deposition. These findings provide key insights into the regulation of the APP processing pathway and provide evidence for the role of BACE1 as a primary regulator of age- related regional Aβ production throughout the mammalian lifespan.
MATERIALS AND METHODS
Animals and Genetic Crosses. All animals were handled according to official guidelines
(IACUC). Animals were bred on a C57BL/6J background. BACE1 BAC transgenic animals (Chiocco et al., 2004), were crossed to either hemizygous or homozygous R1.40
YAC APP transgenic animals (Lamb et al., 1997). R1.40 APP transgene homozygosity was determined by fluorescence in situ hybridization, as described previously (Kulnane and Lamb, 2001). Progeny of these crosses were sacrificed at post-natal day 7 and 60 for biochemical and immunohistochemical analysis. For Aβ deposition analysis, animals were aged 13 and 18 months before being sacrificed.
83 Tissue Procurement. For post-natal ages day 7 and 60, whole brains were removed and immediately dissected along the midline. For brain region studies, brains were dissected on a filter soaked with phosphate buffer saline (PBS) while chilling on wet ice. Each brain was dissected into the following regions: olfactory bulb, cerebellum, hippocampus, and cortex. Each hemi-brain or brain region was immediately frozen on dry ice and transferred to -80oC freezer until use for biochemical analysis described below. For immunohistochemical studies, animals were anesthetized with Avertin (0.02 cc/gm of body weight) and transcardially perfused with ice-cold phosphate buffer (0.1 M, pH 7.4) followed by 4% paraformaldehyde. After perfusion, whole brains were immersion-fixed in 4% paraformaldehyde for > 24h and then cryo-protected by immersion in a sucrose gradient of 10%, 20%, and 30% sucrose. Sagittal sections were cut on a sliding microtome and stored at –80oC.
Protein Extraction. Protein was isolated from frozen transgenic hemi-brain or brain region tissues by polytron homogenization using a PowerGen 125 homogenizer (Fisher,
Pittsburgh, PA) with 5x95mm or 7x95mm generator probes, depending on buffer volume needed for tissue homogenization. For Western blot analysis, tissues were homogenized in 1% CHAPS solution in PBS with 1x protease inhibitors [pepstatin, 1 µM; leupeptin,
4.5 µg/ml; aprotinin, 30 µg/ml; 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), 1 mM] as previously described (Lehman et al., 2003). Total protein concentration was determined by the BCA assay (Pierce, Rockford, IL). For Aβ ELISA analysis, tissues were homogenized in 5 M Guanidine-HCl in 50 mM Tris-HCl, pH 8.0 as previously described (Johnson-Wood et al., 1997).
84 Western blot analysis and quantitation. Total proteins were separated by electrophoresis and transferred onto Immobilon-P transfer membranes (Millipore, Billerica, MA) for antibody detection. All protein blots were probed with anti-rabbit HRP-conjugated
(Amersham Biosciences, Piscataway, NJ) secondary antibody and detected by chemiluminescence (ECL, Amersham Biosciences, Piscataway, NJ). For analysis of
BACE1 protein expression, equal amounts of total protein from brain tissue extracts were resolved on 8% Tris-Glycine gels (Invitrogen Life Technologies, Carlsbad, CA) and detected with rabbit polyclonal antibody 00/6 (kindly provided by G. Evin, Ph.D.,
University of Melbourne). The relative level of human BACE1 protein expression was determined by interpolating signal intensity captured by a fluorescence imager (Flour-S-
Max Imaging Machine, Bio-Rad, Hercules, CA) relative to a standard curve of nontransgenic Bace1 expression, as described previously (Chiocco et al., 2004). For analysis of APP C-terminal fragments, twenty micrograms of total protein were resolved on 4-12% Bis-Tris gradient gels (Invitrogen Life Technologies, Carlsbad, CA). Brain tissue extracts were blotted with 369, a polyclonal antibody raised against the C-terminus of APP (kindly provided by S. Gandy, MD, PhD., Thomas Jefferson U.). APP C- terminal fragments were quantified by comparing the relative ratio of CTF-β to the total amount of C-terminal fragments generated for each animal, as described previously
(Lehman et al., 2003; Chiocco et al., 2004).
Sandwich ELISAs for Aβ. Brain extracts (either hemi-brain or brain region) animals were analyzed for levels of Aβ peptides. Protein extracts were analyzed for levels of Aβ peptides using Aβ1-40 ELISA (Biosource International, Camarillo, CA), as previously described (Chiocco et al., 2004). The values were read using fluorometric plate reader
85 Wallac 1420 multilabel counter (Perkin/Elmer, Wellesly, MA). Aβ1-40 concentrations were determined based on sample values relative to the serially diluted standards representing a standard curve of known Aβ1-40 concentration. Each sample was analyzed in triplicate and sample values are expressed as pmole Aβ1-40/gram brain tissue weight.
Immunohistochemistry. Brain sections were treated with avidin/biotin blocking kit
(Vector Laboratories, Burlingame, CA) and blocked in 10% normal goat serum (Sigma-
Aldrich Co., St. Louis, MO). For BACE1 detection, sections were stained with B279 rabbit polyclonal antibody (kindly provided by R. Yan, Ph.D., Cleveland Clinic
Foundation), directed towards aa 295-310 of human BACE (Yan et al., 2001b), which shares identical sequence with mouse Bace1. For APP detection, sections were stained with 6E10 (Senetek, Napa, CA), a mouse monoclonal antibody that is directed against
Aβ1-17. For BACE1 signal, sections were incubated in biotinylated goat anti-rabbit secondary antibody or biotinylated goat anti-mouse secondary antibody (Vector
Laboratories, Burlingame, CA) for APP signal. For Aβ immunostaining, half-brains were immersion fixed in 10% formalin and embedded in paraffin. The methods for Aβ immunostaining were performed as previously described by Kulnane and Lamb (Kulnane and Lamb, 2001) and stained with mAb 6E10. Sections were prepared according to standard methods using Vectastain ABC Kit (Vector Laboratories, Burlingame, CA) with the chromogen 3,3-diaminobenzidine tetrahydrochloride (DAB) and counterstained with hematoxylin. For immunofluorescence, cryosections were stained with B279 BACE1 antibody. After several washes, Alexa 488-conjugated goat anti-rabbit (for BACE1 signal) secondary antibody was added to the incubation buffer (Molecular Probes,
86 Eugene, OR) and mounted with Vecta-Shield Hard Set containing Dapi counterstain
(Vector Laboratories, Burlingame, CA).
Statistical analysis. Two-tailed t tests were utilized for statistical analyses when variances for each sample population were equal. Two-tailed t test with Welch’s correction was used to adjust for unequal variances between sample populations. One- way ANOVA was utilized to compare BACE1 expression across brain regions. In these cases, Tukey’s post hoc test was utilized to assess multiple comparisons across several sample populations. All data were analyzed by GraphPad Prism 4.0 statistical software
(GraphPad Software, Inc., San Diego, CA, 1998).
RESULTS
Human BACE1 and APP Processing by Whole Brain: Neonatal vs. Adult Mice
Knowing human BACE1 expression alters APP processing and Aβ production in the cDNA-based Swedish mutant APP transgenic (Tg2576) model (Chiocco et al., 2004), we wanted to determine how human BACE1 expression alters APP processing in a genomic-based APP transgenic model, the YAC APP R1.40 line (Lamb et al., 1997).
The YAC APP transgenic line expresses the Swedish mutant human APP ~ 2-fold higher than endogenous mouse App under the control of its native regulatory elements (Lamb et al., 1997). The genomic-based R1.40 APP line makes fewer assumptions about the spatial and temporal regulation of APP expression and processing. Thus, we crossed our human BACE1 BAC transgenic line to the YAC APP R1.40 transgenic line to characterize the regulation of APP processing over the course of the mouse lifespan.
87 Based on evidence that BACE1 gene expression is elevated in embryonic and post-natal animals (Marcinkiewicz and Seidah, 2000), we compared young post-natal and adult mouse brain for differences in BACE1 protein expression and Aβ production. By
Western blot, we confirmed human BACE1 protein expression is elevated in post-natal day 7 (P7) mice compared to adult mice (Figure 3-1A). We determined human BACE1 expression is more than two-fold higher in P7 transgenic animals compared to post-natal day 60 (P60) adult transgenics (Figure 3-1B) (two-tailed t test, p value < 0.0001).
To examine the impact of human BACE1 overexpression on Aβ metabolism in these double APPxBACE1 transgenics, the levels of Aβ peptide in whole brain extracts were measured by ELISA (Figure 3-1C). We found that Aβ1-40 levels were significantly higher in P7 BACE1 and APP double transgenics compared to P60 BACE1 and APP double transgenics (two-tailed t test, p value < 0.0001). Thus, Aβ production in whole brain extracts is significantly altered in neonatal mice.
BACE1 Protein Expression by Brain Region
We have previously documented that brain regional differences in BACE1 expression was correlated with increased Aβ production and deposition (Chiocco et al.,
2004). To further examine the profile of BACE1 expression by brain region and its relationship to APP processing, we analyzed protein expression in early post-natal and adult BACE1 transgenic extracts by Western blot. The pattern of BACE1 protein expression differs in P7 transgenic mice (Figure 3-2A) from that of adult transgenic mice
(Figure 3-2C). In the young P7 transgenics, BACE1 expression in the cortex and hippocampus is higher than expression in the cerebellum (ANOVA, p value < 0.01)
(Figure 3-2B). In the P60 adult transgenics, however, BACE1 expression in the
88
Figure 3-1: Human BACE1 expression and Aβ production in neonatal and adult BACE1 transgenic mice.
A, Western blot analysis of mouse and human BACE1 in 1% CHAPS brain extracts. Twenty micrograms of total protein from adult BACE1 transgenic in lane 1 (TG), adult nontransgenic control in lane 2 (NT), one-week old BACE1 transgenic in lane 3 (TG), and mouse Bace1 KO in lane 4 (mBACE KO) mice were run on 8% Tris-glycine gels, transferred to polyvinylidene difluoride membrane, and blotted with C-terminal antibody BACE-00/6. On the left for each gel are sizes of molecular weight markers in kilodaltons. B, BACE1 protein expression was quantified by comparing the amount of BACE1 protein in post-natal day 7 (P7) BACE1 transgenics (n=11) and post-natal day 60 (P60) BACE1 transgenics (n=10) brain tissue extracts relative to a standard curve of Bace1 expression using a fluorescence imager for capture of chemiluminescent signal. Human BACE1 is expressed ~2-fold higher in P7 transgenic animals compared to P60 transgenic animals (two-tailed t test with Welch’s correction, p value < 0.0001). C, Brain extracts from BACE1/R1.40 YAC APP double transgenic animals were analyzed by Aβ1- 40 ELISA. P7 double transgenic animals (n= 10) had significantly higher levels of Aβ1- 40 compared to P60 double transgenic animals (n= 9) (two-tailed t test with Welch’s correction, p value < 0.0001).
89 hippocampus is the lowest-expressing brain region, while expression in the cortex remains elevated. Cortical BACE1 expression is significantly higher compared to cerebellum (ANOVA, p value < 0.05) and hippocampus (ANOVA, p value < 0.001), while olfactory bulb expression is only significantly higher than hippocampus (ANOVA, p value < 0.001) (Figure 3-2D).
To confirm this altered regional pattern of human BACE1 expression observed by
Western blot, we analyzed the immunohistochemical profile of BACE1 expression in these early post-natal and adult animals. Fixed brain sections from young post-natal mice exhibited more prominent BACE1 staining than that observed in sections from adult mice
(Figure 3-2E-H). Surprisingly, the P7 transgenics exhibited a unique nuclear sub- cellular staining pattern (Figure 3-2E, F), while the profile of BACE1 staining in the adult mice (Figure 3-2G, H) revealed punctate sub-cellular staining consistent with endosomal markers as seen in previous studies (Yan et al., 2001b; Chiocco et al., 2004).
Additional studies are required to understand this unique nuclear staining pattern in the neo-natal brain. Nonetheless, consistent with our Western blot data, cortical and hippocampal cells exhibited the strongest BACE1 staining in the sections of early post- natal mice, while the most prominent staining is present in the cortical layers in sections from adult mice.
APP Processing by Brain Region: Neonatal vs. Adult Mice
To examine the potential consequences of altered regional BACE1 expression on
APP processing, we analyzed APP expression by Western blot (Figure 3-3A). First, based on relative intensity, we observed increased levels of the 100 kDa holoprotein expression at young ages compared to APP levels at adult stages (Figure 3-3A, right).
90
Figure 3-2: Human BACE1 protein expression by brain region in neonatal and adult BACE1 transgenic mice.
A and C, Western blot analysis of human BACE1 in 1% CHAPS extracts from olfactory bulb, cerebellum, hippocampus, and cortex brain regions in P7 BACE1 transgenics (A) and P60 BACE1 transgenics (C) were run on 8% Tris-glycine gels, transferred to polyvinylidene difluoride membrane, and blotted with C-terminal antibody BACE-00/6. B and D, BACE1 protein expression was quantitated using Western blot analysis by comparing the amount of BACE1 protein relative to a standard curve of Bace1 expression using a fluorescence imager to capture chemiluminescent signal. B, For P7 BACE1 transgenics (n= 12), human BACE1 expression is higher in the cortex and hippocampus than in the cerebellum (ANOVA, p value < 0.01). D, For the P60 BACE1 transgenics (n= 10), olfactory bulb BACE1 expression is higher than hippocampus (ANOVA, p value < 0.001); expression in the cortex is higher than in cerebellum (ANOVA, p value < 0.05) and hippocampus (ANOVA, p value < 0.001). E-H, Brains from P7 BACE1 transgenic (E, F) and P60 BACE1 transgenic (G, H) animals were cryo- protected in 30% sucrose after fixing in 4% paraformaldehyde. Sagittal sections were analyzed by staining with polyclonal antibody B279 followed by staining with Alexa488- conjugated secondary antibody. Positive sub-cellular staining is present in the cortex (E and G) and hippocampus (F and H). Scale bar in H = 25 µm.
91 Although APP expression is generally elevated, we observed noticeably increased APP levels by brain region, particularly in the cortex and hippocampus, at this stage. Elevated
APP holoprotein levels are accompanied by high levels of CTF-β, the precursor to Aβ
(Figure 3-3A, left) in the same brain regions. Consistent with in situ hybridization studies (Marcinkiewicz and Seidah, 2000), these protein expression studies indicate that, along with increased APP holoprotein expression, APP processing is altered in early post-natal mice.
To examine the impact of human BACE1 and APP expression on Aβ metabolism by brain region in early post-natal and adult mice, the levels of Aβ peptide in brain region extracts were analyzed by ELISA (Figure 3-3, B-E). Similar to the altered regional pattern of APP expression by Western blot, we observed a selective increase of Aβ levels by brain region, particularly in the hippocampus, in the young BACE1 and APP double transgenics compared to adult BACE1 and APP double transgenics (two-tailed t test, p value = 0.018). Increased regional APP expression levels in young animals corresponds with increased Aβ levels, indicating the significance of APP expression in the APP processing pathway. While these studies demonstrate the impact of BACE1 and APP expression on Aβ production levels, they have not addressed the age-related phenomenon of Aβ metabolism commonly observed in AD.
Brain Regional and Age-Dependent Aβ Production
To examine age-dependent alterations in APP processing due to BACE1 expression, we compared Aβ levels in animals expressing both APP and BACE1 versus animals expressing APP alone. First, we analyzed Aβ1-40 levels in whole brain extracts by ELISA. Double transgenics for BACE1 and APP exhibit no significant differences in
92
Figure 3-3: Amyloid-β metabolism by brain region in neonatal and adult BACE1 and APP double transgenic mice.
A, Western blot analysis of APP holoprotein and APP C-terminal fragments in 1% CHAPS brain region extracts from P7 and P60 BACE1/R1.40 YAC APP double transgenic animals. Twenty micrograms of total protein from olfactory bulb (OB), cerebellum (CB), hippocampus (Hipp), and cortex (Ctx) were run on 4-12% Bis-tris gradient gels, transferred to polyvinylidene difluoride membrane, and blotted with APP C-terminal polyclonal antibody 369. Shown are APP holoprotein and C-terminal fragments at one exposure timepoint (left) and APP holoprotein only at a shorter exposure time (right). On the left of each gel are sizes of molecular weight makers in kilodaltons. B-E, Brain region extracts were analyzed by Aβ1-40 ELISA. P7 double (n= 12) transgenics were compared to P60 double transgenics (n= 12) for total Aβ1-40 levels in olfactory bulb (B), cerebellum (C), hippocampus (D), and cortex (E). Hippocampal Aβ1-40 is higher in P7 double transgenic animals compared to P60 double transgenics (two-tailed t test with Welch’s correction, p value = 0.018).
93 whole brain for Aβ production compared to APP transgenics alone (data not shown).
Knowing regional BACE1 expression significantly alters Aβ levels (Chiocco et al.,
2004), we investigated whether Aβ production differed by brain region in 2-month old mice. Adult mice expressing human APP and human BACE1 did not exhibit any differences of Aβ1-40 levels compared to age-matched APP transgenics alone in brain region extracts for olfactory bulb, cerebellum, or hippocampus (Figure 3-4 A-C).
However, double transgenics expressing human APP and human BACE1 did exhibit significantly higher levels of Aβ1-40 compared to age-matched APP transgenics alone in cortex, an increase of ~ 25% (Figure 3-4D) (two-tailed t test, p value < 0.0001). These results indicate significantly elevated BACE1 protein expression in the cortex may directly influence cortical Aβ production in adult animals but not in younger animals.
Age-Dependent Regional Alterations of Aβ Deposition
This age-related regional increase in Aβ1-40 levels is particularly striking when
BACE1xAPP double transgenic animals are aged further. To examine the impact of altered regional BACE1 expression and Aβ production on Aβ deposition, we analyzed the Aβ immunohistochemical profile of BACE1xAPP double transgenic animals compared to age-matched single APP transgenics alone. Consistent with increased cortical BACE1 expression and Aβ production, Aβ deposits only developed in the cortex of double transgenics at both 13 months and 18 months (Figure 3-5A and 3-5C) whereas age-matched APP transgenic controls did not develop any alterations in Aβ deposition
(Figure 3-5B and 3-5D). Furthermore, although Aβ pathology is strictly limited to the cortex, deposits spread throughout the primary and secondary motor cortices (Paxinos,
2001) from 13 months to the 18-month time-point, indicating an age-related expansion of
94
Figure 3-4: Impact of BACE1 expression on Aβ production by brain region in adult animals.
Aβ1-40 ELISA analysis of brain region extracts from P60 BACE1/R1.40 YAC APP double transgenic mice (n= 12) and age-matched R1.40 YAC APP single transgenic mice (n= 12). Total Aβ1-40 levels were analyzed in olfactory bulb (A), cerebellum (B), hippocampus (C), and cortex (D). BACE1/R1.40 transgenic mice have significantly higher Aβ1-40 levels in the cortex (D) compared to R1.40 transgenics alone (two-tailed t test with Welch’s correction, p value < 0.0001).
95
Figure 3-5: Amyloid-β deposition profile in aged BACE1xAPP animals.
Brains from 13-month (A, B) and 18-month (C, D) old animals were fixed in 10% formalin, embedded in paraffin, and sectioned on a sagittal plane (10 µm thick). Scattered sections were analyzed by staining using standard protocols (Kulnane et al., 2001) with mAb 6E10, which detects amino acids 1-17 in the Aβ region. Shown are whole brain scans (upper) and frontal cortex brain sections (lower) of animals transgenic for both BACE1 and homozygous R1.40 YAC APP at 13 months (A) and 18 months (C). Shown are whole brain scans (upper) and frontal cortex brain sections (lower) of animals transgenic for homozygous R1.40 YAC APP alone at 13 months (B) and 18 months (D). Scale bar in D = 100 µm.
96 amyloid deposition. This altered deposition profile indicates age-related differences in brain regional BACE1 expression influences the pattern and progression of AD-like neuropathology.
DISCUSSION
According to the amyloid hypothesis, the generation of Aβ from APP is the primary pathogenic event in Alzheimer’s disease. The correlation of genetic mutations in early-onset AD pedigrees, particularly APP and the PSEN genes, with increased Aβ levels indicates amyloidogenic processing of APP predominates in AD patients (Goate et al., 1991; Hendriks et al., 1992; Mullan et al., 1992; Levy-Lahad et al., 1995). These and other mutations (Rocchi et al., 2003) that result in elevated Aβ production, carried out by
β-secretase and γ-secretase, provide insight into how modulation of APP processing may cause AD pathogenesis. However, it is still unclear how the spatial and temporal patterns of APP processing correlate with the age-related profile of Aβ production and neuropathology observed in AD. Along these lines, increased regional activity of BACE1
(Fukumoto et al., 2002), as the primary β-secretase gene, in the AD brain indicates
BACE1 may play a major role in regulating the pattern of Aβ production and, ultimately,
AD pathogenesis.
In hopes of understanding the influence of BACE1 expression on APP processing, we report that the profile of brain-regional expression of human BACE1 in our β- secretase transgenic mice alters the temporal and spatial pattern of Aβ production.
Specifically, we observe alterations in brain regional BACE1 protein expression as the animals age from early post-natal stages to adulthood. Interestingly, increased
97 hippocampal BACE1 and APP expression are significantly elevated in young animals compared to adult animals. However, the regional pattern of BACE1 expression, specifically in the cortex, selectively influences regional Aβ production in adult animals, which ultimately results in distinct patterns of AD-like neuropathology. Our studies provide evidence that BACE1 regulates age-related Aβ production and deposition.
Understanding how the regulation of the APP processing pathway influences Aβ production has become a crucial area of investigation. As with most biochemical pathways, the expression and activity of the enzyme are as critical as the availability of the enzyme’s substrate with respect to the amount of product generated. In the APP processing pathway, the relative expression of APP coupled with the expression and activity of the APP secretases will likely dictate the amount of Aβ generated. Similar to previous studies (Loffler and Huber, 1992; Marcinkiewicz and Seidah, 2000; Basha et al.,
2005), we observe increased APP protein expression in early post-natal animals relative to adult animals. It has been suggested that increased APP levels in young animals is due to its potential role in the developing brain (Lahiri et al., 2002). However, it is not known why BACE1 protein levels are also more than 2-fold higher in younger animals. The increased BACE1 and APP protein levels lead to significant increases in Aβ levels at this young age compared to adult animals.
In our studies, elevated Aβ levels in animals co-expressing genomic transgenes of human BACE1 and human APP are primarily a result of an age-related regional distribution of both APP and BACE1 protein (refer to Table 1). Intriguingly, increased
APP holoprotein levels in young animals is accompanied by selective brain-regional increases in APP processing products, namely the Aβ precursor peptide, CTF-β.
98 Table 3-1: Summary of alterations in brain-regional BACE1 expression and APP processing.
Plus (+) signs indicate relative levels of expression or protein processing products. P7 corresponds to post-natal day 7 animals; P60 corresponds to post-natal day 60 animals.
99 Along these lines, cortical and hippocampal Aβ peptides are almost 2-fold higher than
Aβ levels in cerebellum and olfactory bulb. Although BACE1 expression is significantly higher in the cortex and hippocampus at this age, only Aβ levels in the hippocampus are increased relative to adult animals. One of the possible explanations for this is the only brain region in adult animals expressing human BACE1 with increased Aβ levels is the cortex, where BACE1 protein expression is the highest. Our studies suggest that as APP expression reduces with age, BACE1 can significantly influence APP processing, possibly indicating that coordinated expression levels of BACE1 and APP alters the APP processing pathway. Although APP processing by brain region in adult animals of our
YAC APP transgenic model has been analyzed previously (Lehman et al., 2003), these studies demonstrate the functional consequences of β-secretase overexpression on regional Aβ production at different ages.
Degradation and clearance of Aβ from the brain is another possible mechanism by which Aβ levels are regulated. Impairments in the Aβ clearance pathway leading to accumulation and aggregation of Aβ deposits may initiate AD pathogenesis. Clearance of Aβ follows two major pathways: proteolytic degradation and receptor-mediated transport from the brain (Tanzi et al., 2004). Among many proteases, recent studies have built strong evidence for the ability of two metalloproteases, neprilysin (NEP) and insulin degrading enzyme (IDE), to hydrolyze Aβ and reduce amyloid plaque burden in mice
(Iwata et al., 2001; Leissring et al., 2003; Marr et al., 2003; Farris et al., 2004).
Similarly, two receptor-associated proteins, low-density lipoprotein receptor-related protein (LRP) and the receptor for advanced glycation end products (RAGE), are instrumental in Aβ transport from the brain to the periphery for degradation (Kang et al.,
100 2000; Shibata et al., 2000; Deane et al., 2003; Mohajeri et al., 2004). Our findings also lead to interesting questions about Aβ turnover. Since we do not observe Aβ plaque accumulation in young animals, it might be possible that Aβ clearance is more efficient at this age. The changes we observe in the regional distribution of Aβ levels as BACE1 transgenic animals age indicates Aβ degradative activity may also differ by brain region.
Analyzing age-related differences in activity of known Aβ clearance genes by brain region would likely provide insight into the balance between Aβ clearance versus Aβ accumulation and deposition.
Finally, the age-related brain-regional differences in Aβ production we observe is particularly significant due to the altered regional deposition profile in the aged
BACE1xAPP double transgenics. As we have previously shown in another APP transgenic model (Chiocco et al., 2004), brain-regional BACE1 expression in aged, depositing animals corresponds with significant alterations in the regional deposition profile. The cortex-specific deposition profile in the BACE1 and YAC APP double transgenic animals indicates BACE1 regional expression influences the onset and site of
Aβ deposition. These studies suggest that age-dependent brain-regional fluctuations in
BACE1 and APP protein expression ultimately have a dramatic impact on regional Aβ deposition (Table 1, shaded boxes). These studies lead to interesting questions about
BACE1 expression and regulation. For example, do expression patterns change immediately prior to and/or subsequent to deposition? How do BACE1 expression patterns alter the biochemical structure of Aβ peptides so as to influence aggregation?
Besides regional expression patterns, further studies could provide insight into
BACE1 activity. Recent studies have identified a family of proteins involved in
101 modulating BACE1 activity, known as reticulon proteins (He et al., 2004). These proteins, particularly Reticulon3 (RTN3), have been shown to bind Aβ directly and inhibit Aβ activity (He et al., 2004). Further studies could explore the expression profile of the reticulon proteins in our BACE1 transgenic animals. These studies could provide insight into the relationship between BACE1 activity and age-related regional Aβ production and deposition profiles.
In summary, we have characterized unique patterns of Aβ production in our
BACE1xAPP double transgenic animals. Although Aβ levels are generally increased in young animals due to elevated levels of both BACE1 and APP, the pattern of BACE1 overexpression alone significantly impacts brain regional Aβ production in an age- dependent manner. These studies demonstrate how BACE1 and APP expression modulate age-related Aβ metabolism and provide invaluable insight into the regulation of the APP processing pathway. Further studies examining the temporal and spatial patterns of BACE1 activity regulation and Aβ production might offer a crucial link underlying the mechanism of BACE1’s role in Aβ deposition and AD pathogenesis.
102 Chapter 4: Characterization of APP Processing in Genomic-based BACE2
Transgenic Animals
Authors: Matthew J. Chiocco and Bruce T. Lamb
103 INTRODUCTION
Alzheimer’s disease (AD) neuropathology is characterized by two pathological hallmarks, amyloid-β deposits and neurofibrillary tangles. Genetic studies suggest the generation and deposition of amyloid-β (Aβ) peptide in the brain is the critical link to AD pathogenesis. Amyloid-β is generated by the sequential cleavage of Amyloid Precursor
Protein (APP) by β-secretase and γ-secretase. The identification of the β-secretase gene,
BACE1, paved the way for the discovery of its homologue, BACE2. The gene was initially identified in a cDNA homology search with BACE1 (Saunders et al., 1999).
Also known as Asp1 or memapsin 1, BACE2 maps to human chromosome 21q22.3 within the Down syndrome critical region (Yan et al., 1999) (Acquati et al., 2000).
BACE1 and BACE2 exhibit 52% aa sequence identity and 68% similarity, sharing structural properties as unique aspartyl proteases with single transmembrane domains (Acquati et al., 2000) (Saunders et al., 1999). Because of its sequence identity, it was postulated that BACE2 is a functional homologue to BACE1 as a β-secretase.
However, data from cell culture studies suggested that BACE2 overexpression reduces the secretion of Aβ (Yan et al., 2001a). With highest expression in peripheral tissues,
BACE2 does not exhibit abundant neuronal expression expected for a putative β- secretase. Further studies demonstrated that the primary cleavage position for BACE2 resides within the Aβ peptide at residues Phe19/20, resulting in the production of a non- amyloidogenic peptide (Farzan et al., 2000). Intriguingly, this cleavage position coincides with the early-onset familial AD Flemish mutation (A692G), which is associated with a unique form of AD commonly resulting in clinical dementia symptoms accompanied by cerebral amyloid angiopathy (CAA) (Hendriks et al., 1992). Thus,
104 BACE2 may be genetically associated with AD, particularly involving the pathogenesis of vascular amyloid.
However, BACE2 antagonizes the effect of BACE1 cleavage when the two are co-expressed in cells with APP (Basi et al., 2003). BACE1 cleavage products, CTF-β and Aβ, are reduced when BACE2 is present. This surprising finding indicates BACE2 may function as an alternative α-secretase, competing for the APP substrate. These findings have led to the hypothesis that BACE2 may play a protective role in reducing
Aβ secretion and, ultimately, AD pathogenesis. To test this hypothesis, I have generated genomic-based BACE2 transgenics that express the entire human BACE2 gene. By crossing these animals to APP mutant transgenics, I have analyzed the effect of BACE2 on APP processing and Aβ production. Unlike previous in vitro and cell culture studies, our studies demonstrate that genomic expression of human BACE2 in the mouse does not alter APP processing in the brain.
MATERIALS AND METHODS
Animals and Genetic Crosses. All animals were handled according to official guidelines
(IACUC). Animals were bred on a mixed C57BL/6J x SJL background. BACE2 transgenic animals were crossed to APP transgenics (Tg2576), which contain the
Swedish mutant human APP cDNA (kindly provided by K. Hsiao Ashe, U. of
Minnesota). Progeny of these crosses were sacrificed at 2 months of age for expression and biochemical analysis.
Human BACE2 Genomic Analysis. Human BACE2 was mapped to human chromosome
21q22.3 (Acquati et al., 2000). Similar to the methods used for BACE1 genomic analysis
105 (Chiocco et al., 2004), sequence-tagged sites (STSs) and gene markers were used to identify the precise genomic localization of human BACE2 on chromosome 21. Since chromosome 21 sequence was near completion, identifying genomic clones spanning
BACE2 was somewhat easier than BACE1 genomic analysis. Genomic sequence contigs were used to search the human genome by NCBI Blast
(http://www.ncbi.nlm.nih.gov/BLAST) and identify genomic bacterial artificial chromosome (BAC) clones containing the full-length human BACE2 gene. In addition,
BAC-end sequence (available from http://www.tigr.org/tdb/humgen/bac_end_search/bac_end_intro.shtml) and the published genome assembly (http://genome.cse.ucsc.edu/) verified the locus coverage for each clone. Genomic clones identified in silico were further verified by PCR with STSs
D21S53, D21S355, D21S266, and BAC vector arm primers.
Isolation and Purification of BAC Clones. BAC clones 1107H21 from the RPCI-11 library, 2509H19 and 2590L4 from the CITBI-EI library were obtained. Once BAC clones containing full-length BACE2 were identified, they were purified using the
Clontech Nucleobond column (Palo Alto, CA) and passed through a CL4B Sepharose column (Amersham Pharmacia Biotech, Uppsala, Sweden) and collected in 12 elution fractions. The appropriate fraction containing the BAC was diluted to a concentration ranging between 0.6 ng/µl – 1.0 ng/µl, suitable for pronuclear micro-injection (Yang et al., 1997).
Generation of BACE2 Transgenic Mice. Transgenic mice were generated by direct microinjection of BAC DNA into the pronuclei of fertilized mouse eggs. This method was performed as described (Hogan et al., 1994). The embryos injected were F2 progeny
106 of a C57BL6/SJL F1 cross and were surgically transferred into the oviducts of a pseudopregnant CD-1 female. Founder mice transgenic for BAC clones corresponding to the human BACE2 locus were identified by PCR from mouse tail DNA. Genotyping markers to identify founders were the STS D21S355, which is located upstream of
BACE2 exon 8, and a custom primer, which is located one kilobase upstream of BACE2 exon 1, with the sequence 5’- GGCGTCTATGGCTAAACGAG-3’ and 3’-
TGCAGGTCAGATGCTACTGG-5’.
RT-PCR Analysis. Total RNA was isolated from transgenic and nontransgenic mouse brain, kidney, liver, colon, spleen, and small intestine by polytron homogenization of tissue in TRIzol Reagent (Invitrogen Life Technologies, Carlsbad, CA). Each RNA preparation (2 µg) was reverse transcribed in the presence of 50 pmol random hexamer primers (Invitrogen Life Technologies, Carlsbad, CA) and Reverse TranscriptaseTM
(Invitrogen Life Technologies, Carlsbad, CA). After reverse transcription, 1/20 of the RT reaction was incubated in a PCR reaction with 50 pmol of primers with sequence 5’-
CTGGACTGCAGAGAGTATAAC-3’ and 3’-GCATAGGACACAATCCACA-5’ for 27 cycles at an annealing temperature of 600. These primers recognize both mouse and human Bace2/BACE2 transcripts and amplify a 553 bp product that spans exons 5-9 of the gene. PCR products were purified and then digested with KpnI restriction enzyme, which digests the mouse Bace2 product uniquely into two fragments (449 and 104 bp in size), leaving human BACE2 product uncleaved.
Tissue Preparation and Biochemical Analysis. Protein was isolated from 2-month old transgenic and nontransgenic mouse brain tissues by polytron homogenization in TNES
107 solution with 1% SDS with 1x protease inhibitors [pepstatin, 50 µg/ml; leupeptin, 50
µg/ml; aprotinin, 10 µg/ml; 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), 1 mM].
For analysis of APP C-terminal fragments, twenty-five micrograms of total protein were resolved on 4-12% Bis-Tris gradient gels (Invitrogen Life Technologies,
Carlsbad, CA). Brain tissue extracts were blotted with 369, a polyclonal antibody raised against the C-terminus of APP (kindly provided by S. Gandy, MD, PhD., Thomas
Jefferson U.). All protein blots were probed with anti-rabbit HRP-conjugated
(Amersham Biosciences, Piscataway, NJ) secondary antibody and detected by chemiluminescence (ECL, Amersham Biosciences, Piscataway, NJ).
For analysis of total Aβ1-40 levels, hemi-brains were homogenized in 5M
Guanidine-HCl in 50 mM Tris-HCl, pH 8.0 as previously described (Johnson-Wood et al., 1997). Brain extracts from 2-month old animals were analyzed for levels of Aβ peptides using Aβ1-40 ELISA (Biosource International, Camarillo, CA). The values were read using fluorometric plate reader Wallac 1420 multilabel counter (Perkin/Elmer,
Wellesly, MA). Aβ1-40 concentrations were determined based on sample values relative to the serially diluted standards representing a standard curve of known Aβ1-40 concentration. Each sample was analyzed in triplicate and sample values are expressed as pmole Aβ1-40/gram brain tissue weight.
Aβ Immunohistochemistry. Twelve-month and eighteen-month old progeny of
BACE2xTg2576 transgenics were sacrificed and hemi-brains were fixed in 10% formalin. Formalin-fixed brains were embedded in paraffin and sliced into 10 µm sections. Brain sections were stained for Aβ using monoclonal antibody 6E10 using methods previously described (Kulnane and Lamb, 2001).
108 Statistical analysis. Two-tailed t tests were utilized for statistical analyses when variances for each sample population were equal. Two-tailed t test with Welch’s correction was used to adjust for unequal variances between sample populations.
RESULTS
Identification of Human BACE2 Genomic Clones
Screening the TIGR BAC-end sequence and GSS databases revealed approximately 7-8 clones that corresponded to the region containing BACE2. The DNA sequence contig spanning this region is represented by two GenBank entries (Accession nos. AL163284 and AL163285). Three clones were determined to cover the entire gene and flanking sequences: BAC 1107H21 from the RPCI-11 library and BACs 2509H19 and 2590L4, both from the CITBI-EI library (Figure 4-1A). BAC 1107H21 contains
18.2 kb upstream of the first exon of BACE-2 and 57.6 kb of downstream sequence.
BACs 2509H19 and 2590L4 contain similar upstream and downstream sequence. Since these three clones extend across the BACE2 region with almost identical coverage, one clone was chosen, 1107H21, to generate BACE2 transgenic mice. In addition, the region corresponding to the BACE2 locus is relatively gene-poor, according to the most recent genome assembly (http://genome.cse.ucsc.edu/), indicating confounding effects from nearby genes should not occur.
Expression and Activity of Human BACE2 BAC Clone
Upon pronuclear microinjection, we recovered two founder mice for BAC clone
1107H21. The founder animals were verified by PCR with primers specific for the
BACE2 gene as well as BAC clone 1107H21 vector arm primers (data not shown). To
109 ensure proper germline transmission, the founder animals were backcrossed to the
C57BL6/J line. The BAC transgene was faithfully transmitted to ~ 50% of the progeny upon subsequent backcross generations.
To analyze expression of human BACE2, we utilized an RT-PCR assay that distinguishes mouse Bace2 from human BACE2 similar to the assay previously described for BACE1 (Chiocco et al., 2004). Specifically, the assay uses conserved mouse and human oligonucleotides that span exons 5-9 to amplify both mouse and human forms of the BACE2 gene, and subsequent digestion with a restriction enzyme that cleaves mouse
Bace2 and leaves human BACE2 uncleaved. Expression of human BACE2 and mouse
Bace2 can be detected in multiple tissues including brain, small intestine, colon, and spleen (Figure 4-1B). Using this assay, we qualitatively assess the level of mRNA expression for the human transgene across multiple tissues based on the relative intensity of uncleaved human products. Expression of human BACE2 mRNA across these tissues varies, with highest expression in the small intestine and spleen and lowest expression in the brain and colon.
Since the PCR primers span predicted splice sites, this assay also detects the presence of a human-specific alternative BACE2 isoform. This isoform, according to the most recent genome assembly and predicted protein sequence, has the entire exon 7 spliced out in-frame while the other eight exons are intact. Amplifying this isoform results in a truncated ~ 400 bp fragment, indicated by an arrowhead (Figure 4-1B). Little is known about the expression pattern of this truncated isoform, although both aspartyl protease active sites are intact in this variant.
110
Figure 4-1: Generation of genomic-based human BACE2 transgenic mice and expression of human BACE2 mRNA.
A, (top) Genomic map of the BACE2 locus with the relative position of the BACE2 gene (as represented by the STSS D21S53 and D21S355), nearby genes and STSs, and corresponding BAC clones on chromosome 21q22.3. (bottom) Human BACE2 transgenics recovered by pronuclear microinjection. Two founders were recovered for BAC clone 1107H21. Plus (+) signs indicate PCR-positive animals for the corresponding STSs and primers at the chromosome 21 locus. B, RT-PCR analysis with conserved primers corresponding to sequences identical between mouse and human Bace2/BACE2 mRNAs was performed on reverse-transcribed RNA from brain, small intestine, colon, spleen from nontransgenic (NT) and human BACE2 transgenic (TG) mice. The resulting 553 bp products were digested with Kpn1 to generate 449 and 104 bp mBace fragments, while the human products (hBACE) remain undigested. Arrowhead points to the ~ 400 bp product detected in the assay representing the alternative isoform expressed in human BACE2 transgenic tissue (brain, colon, small intestine). Left, approximate sizes in bp.
111 Human BACE2 and APP Processing
To assess the impact of human BACE2 expression on APP processing, we crossed the BACE2 transgenic animals to two different APP transgenic models of AD,
Tg2576 transgenics and R1.40 transgenics. The cDNA-based Tg2576 transgenic expresses the Swedish mutant APP controlled by a hamster prion promoter at six-fold higher levels of human APP relative to endogenous mouse App (Hsiao et al., 1996).
Alternatively, the genomic-based R1.40 transgenic expresses the Swedish mutant APP under the control of its native promoter and regulatory elements at two-fold increased levels relative to mouse App (Lamb et al., 1993). Both of these transgenic animals exhibit increased APP processing products, particularly CTF-β and Aβ, and develop Aβ deposits in an age-dependent manner (Hsiao et al., 1996; Lamb et al., 1997). Because these two transgenic models exhibit altered patterns of APP processing (Lehman et al.,
2003), we characterized the effect of BACE2 overexpression on APP processing in both transgenic lines.
To examine alterations in APP processing products, we compared APP expression in the double transgenic BACE2xAPP animals to single transgenic APP animals by
Western blot. APP processing is not altered, based on the relative intensity of CTF-β compared to CTF-α in either the double transgenic APP and BACE2 (Figure 4-2A, lanes
1 and 3) or the APP transgenic alone (Figure 4-2A, lane 2).
Human BACE2 and Aβ Metabolism
To assess the impact of BACE2 expression on Aβ production, brain extracts from
2-month old animals were analyzed by Aβ1-40 ELISA. We determined that Aβ1-40 levels are not significantly altered in the progeny of either BACE2xAPP transgenic
112
Figure 4-2: APP processing in BACE2 transgenics.
A, Western blot analysis of APP holoprotein (top) and APP C-terminal fragments (bottom) in 1% CHAPS brain extracts. Twenty micrograms of total protein from ~2 month-old progeny of BACE2 transgenics crossed to R1.40 (APP YAC) transgenic animals were run on 4-12% Bis-tris gradient gels, transferred to polyvinylidene difluoride membrane, and blotted with APP C-terminal polyclonal antibody 369. On the left are sizes of molecular weight makers in kilodaltons. B, Brain extracts from ~ 2-month old animals were analyzed by Aβ 1-40 ELISA. There were no significant differences of Aβ levels in the progeny of R1.40 YAC APP x BACE2 animals (n= 12) (left) or in the progeny of Tg2576 APP cDNA transgenics x BACE2 animals (n= 6) (right).
113 crosses (Figure 4-2B). Since APP is expressed almost 3-fold higher in the cDNA-based
Tg2576 APP transgenics relative to the genomic-based R1.40 YAC APP transgenics
(Lamb et al., 1993; Hsiao et al., 1996), Aβ levels are elevated in the Tg2576 animals, reflecting the differences between APP transgene expression. Yet, there are no significant differences between the Tg2576/BACE2 double transgenics and the age- matched Tg2576 transgenics alone. Consistent with our observations for APP processing, expression of BACE2 does not alter Aβ production.
Human BACE2 and Aβ Deposition
To examine age-dependent alterations of Aβ deposition in BACE2 transgenics, we analyzed brain sections of aged progeny from the BACE2xTg2576 cross using immunohistochemical methods. By immunostaining with an Aβ-specific antibody, we did not observe any differences in the pattern and number of deposits in the double
BACE2xAPP transgenics compared to age-matched single APP transgenic animals
(Figure 4-3). There was no delay of plaque development when we analyzed
BACE2/APP double transgenic and APP control animals at 12 months of age (Figure 4-
3A and 4-3B), when plaque pathology usually develops (Hsiao et al., 1996). In addition, we did not observe a reduction of plaque pathology in the BACE2-expressing animals compared to control animals at eighteen months (Figure 4-3C and 4-3D) when Aβ deposits virtually cover the entire brain. Thus, we determined that Aβ plaque deposition is not altered in aged animals expressing BACE2 and Swedish mutant APP, further indicating BACE2 expression does not alter APP processing and AD-like pathology.
114
Figure 4-3: Analysis of Amyloid-β deposition in BACE2 transgenic animals.
Whole brain scans and hippocampal sections from 12-month old (A, B) and 18-month old (C, D) animals were fixed in 10% formalin, embedded in paraffin, and sectioned on a sagittal plane (10 µm thick). Scattered sections were analyzed by staining using standard protocols (Kulnane et al., 2001) with mAb 6E10, which detects amino acids 1-17 in the Aβ region. The pattern of amyloid-β deposition does not differ between BACE2/Tg2576 double transgenics (A, C) and Tg2576 transgenics alone (B, D).
115 DISCUSSION
The function of BACE2 and its role in APP processing remains somewhat elusive. Identified as a homologue to BACE1, BACE2 was proposed to be a functional orthologue to the primary β-secretase gene (Saunders et al., 1999; Acquati et al., 2000).
Sharing high sequence similarity and aspartyl protease structural motifs, BACE2 was considered a complementary β-secretase to BACE1 in generating Aβ peptides from APP.
However, soon after its discovery, the expression pattern of BACE2 was revealed to be inconsistent with that of a putative β-secretase. Furthermore, its primary cleavage position resides within the Aβ sequence, precluding the generation of amyloidogenic peptides. It is hypothesized that BACE2 may play a protective role in AD pathogenesis by reducing the accumulation of toxic Aβ levels.
To understand the role of BACE2 in APP processing, we sought to generate genomic-based BACE2 transgenics that express the entire human BACE2 gene. Here, we report the successful generation of BACE2 transgenics with a complete genomic copy of the BACE2 gene on human chromosome 21. Expression analysis by RT-PCR reveals mRNA expression is highest in peripheral tissues such as small intestine and spleen, while expression in brain is lower. Crossing these BACE2 transgenic animals to APP transgenic animals, we determine that BACE2 expression does not alter APP processing and Aβ production.
Based on its potential protective role in precluding Aβ formation, we expected to observe reductions of Aβ levels in BACE2 transgenic animals co-expressing human APP harboring familial AD mutations. However, initial studies indicate Aβ levels are not altered in animals expressing both BACE2 and human mutant APP. We postulate that
116 one of the primary reasons for this is the low expression of the human BACE2 transgene in the brain relative to peripheral tissues such as spleen and small intestine. This observation is consistent with previous studies finding that BACE2 expression is highest in peripheral tissues with minimal expression in the brain (Bennett et al., 2000).
Although BACE2 may act as an alternative α-secretase in cells (Yan et al., 2001a), the native expression of BACE2 does not coincide with neuronal APP expression, precluding the effect BACE2 may play in reducing Aβ levels.
Based on these results, the genomic expression of human BACE2 does not alter
APP processing in AD transgenic mouse models. However, BACE2 may have a genetic basis in familial AD cases. Individuals with a mutation at codon 692 of APP develop a unique angiopathy known as cerebral amyloid angiopathy (CAA), which is characterized by cerebral hemorrhage due to vascular amyloid deposition (Hendriks et al., 1992). The disease pathology is also marked by a lack of dystrophic neurites and neurofibrillary tangles. Known as the Flemish mutation, this mutation substitutes alanine for glycine and lies adjacent to the preferred cleavage position for BACE2. Cell culture studies have found that co-expression of BACE2 with APP harboring the Flemish mutation significantly increases the production of CTF-β and Aβ (Farzan et al., 2000). These studies indicate that this familial AD mutation shifts the BACE2 cleavage site to favor amyloidogenic production, showing the role BACE2 may play in the pathogenesis of this particular form of AD. Although BACE2 expression may not alter neuronal Aβ production, its expression in the context of specific AD mutations may alter Aβ levels in the vasculature and contribute to a form of AD pathogenesis. To investigate vascular Aβ,
117 further studies involving Thioflavin immunostaining are necessary in examining these
BACE2 transgenics for alterations of vascular amyloid.
As a protective agent or as a primary player in amyloid angiopathy, BACE2 may play a significant role in APP processing. Our initial studies demonstrate the genomic form of human BACE2 expression does not significantly alter APP processing and amyloid deposition. It will be necessary to characterize BACE2 protein expression and further analyze the type of Aβ peptides generated in these transgenic mice. Alternatively, studies examining ectopic BACE2 neuronal expression might help determine the role
BACE2 may play as an alternative α-secretase. These investigations might elicit a potentially therapeutic link for BACE2 in precluding Aβ production and possibly preventing AD pathogenesis.
118 Chapter 5: Conclusions and Future Directions
Summary
The neurological basis of Alzheimer’s disease pathology has puzzled researchers for almost 100 years since the disease was originally described. Studying AD is difficult for several reasons: (1) onset of the disease is highly variable, (2) an extremely heterogeneous array of factors (genetic and environmental) has been implicated in the cause of the disease, and (3) the disease pathology is inconsistent among affected AD individuals. Beyond the complex pathogenesis of AD, characterizing the disease incidence is made more difficult because individuals cannot be officially diagnosed until post mortem. In essence, therapeutic intervention becomes increasingly demanding without knowing a treatment population.
In the past twenty years, however, genetic studies have provided insight into both identifying candidate AD cases as well as the molecular underpinnings of the complex disease. Most notably, these studies have helped to identify the potential causes behind the development of the two pathological AD hallmarks: neurofibrillary tangles and senile amyloid plaques. By characterizing this pathogenesis, these lesions are found throughout the aging brain in distinct patterns. While tangles are more coincident with cognitive impairments, senile plaques follow a brain-regional pattern consistent with increasing age
(Braak and Braak, 1997b). Although disputes have arisen concerning which, if either, of these two pathological features causes AD, genetic studies have built overwhelming evidence for a prevailing hypothesis in the field regarding AD pathogenesis. The amyloid cascade hypothesis, drawing from the preponderance of families with early- onset AD associated with increased Aβ levels, states that the generation of the 4 kDa Aβ
119 peptide from APP is the cause of the neurodegeneration and, ultimately, the cognitive impairments observed in AD patients. Thus, the production and regulation of Aβ generation has become a primary focus to develop strategies for therapeutic intervention involving the elimination of Aβ from the brain.
Identification of the secretases involved in APP cleavage has led to significant advances towards the characterization of the APP processing pathway. Considered the
Aβ rate-limiting event (Sinha et al., 1999), β-secretase cleavage initiates the Aβ arm of the APP processing pathway. Complete removal of the gene encoding β-secretase in mice results eliminates all Aβ peptides (Cai et al., 2001; Luo et al., 2001; Roberds et al.,
2001). With the identification of BACE1 as the primary β-secretase gene, it became apparent how the aspartyl protease generates Aβ from APP. However, it has been unclear how BACE1 and its homologue, BACE2, influence APP processing and Aβ deposition in vivo. This thesis describes an in vivo approach to characterize the impact of human genomic BACE overexpression on age-related APP processing, Aβ production, and Aβ deposition. These studies have shown that genomic BACE1 expression has distinct brain-regional patterns of expression that, in turn, influences both protein levels and β-secretase activity. This regional expression significantly influences APP processing and Aβ production across the lifespan, resulting in distinct Aβ deposition patterns. In addition, these studies have shown that genomic expression of BACE2 does not significantly alter the APP processing pathway. The studies in this current work demonstrate that BACE1 is a primary regulator of the APP processing pathway and ultimately impacts the development of AD-like neuropathology.
120 Conclusions
Genomic Expression of Human β-secretase Genes
Alzheimer’s disease pathogenesis is characterized by a high degree of heterogeneity among affected populations. The classic neuropathological features of the disease, amyloid plaques and NFTs, do not necessarily correspond with severity of the disease (Haroutunian et al., 1999; Dickson and Vickers, 2001). Increasing age remains the only consistent factor that correlates with risk for the disease. Although genetic factors have helped to classify AD incidence, considerable variability still exists in defining disease onset (Ashford, 2004). Thus, to characterize a disease for which very little is known about the cause of pathogenesis, it is important to utilize approaches that remove as many assumptions as possible regarding development and regulation of genetic factors involved in neuropathology.
In the current work, we took advantage of a genomic-based approach to generate
β-secretase transgenic mice. Since the genes are controlled by their respective native regulatory elements, this approach removes any bias regarding the gene’s expression and regulation. Using public genomic resources, we have characterized the genomic loci for each of the β-secretase genes, namely BACE1 on human chromosome 11 and BACE2 on human chromosome 21. We identified genomic clones containing BACE1 or BACE2 and generated transgenic mice using a pronuclear microinjection strategy (Hogan et al.,
1994). Expression analysis of BACE1 or BACE2 transgenic mice follows patterns consistent with previous studies (Vassar et al., 1999; Bennett et al., 2000). While BACE1 expression is ubiquitous, mRNA expression is highest in the brain (Figure 2-2).
121 Genomic BACE2 expression, however, results in higher peripheral tissue expression with lowest expression in the brain (Figure 4-1).
In recent years, several BACE1 transgenic mice have been generated using a cDNA-based approach that also uses a heterologous promoter to control cell-specific or tissue-specific expression (Bodendorf et al., 2002; Mohajeri et al., 2004; Willem et al.,
2004; Lee et al., 2005). This approach makes specific assumptions about the spatial and temporal expression and regulation of BACE1 throughout the mouse lifespan. Using the genomic-based approach, we have re-capitulated the endogenous temporal and spatial expression patterns of the BACE1 and BACE2 genes, and gained further insight into the gene’s effect on the spatial and temporal patterns of AD pathogenesis.
Spatial and Temporal Control of Human BACE1 Expression
Studies of mice expressing human BACE1 under endogenous regulation have led to significant findings regarding the spatial and temporal expression patterns of the primary β-secretase gene. As described in Chapter 3, we determined that BACE1 is developmentally regulated, with greater than 2-fold increased protein expression at early post-natal stages compared to adult animals, suggesting that BACE1 may play a major role in brain or CNS development. Several studies have identified other substrates
(besides APP) for BACE1, such as a sialyltransferase and a P-selectin glycoprotein
(Kitazume et al., 2001; Lichtenthaler et al., 2003). The aspartyl protease’s role in development may be particularly interesting, considering that its major substrate, APP, has been implicated in mammalian development of the CNS (Lahiri et al., 2002), particularly neuronal adhesion and cell survival during embryonic development (Herms et al., 2004).
122 Most notably, BACE1 brain regional expression is modulated over the course of the animal’s lifespan. From early post-natal stages to aged animals, we observed a fluctuating pattern of BACE1 brain-regional expression (Figure 2-6 and Figure 3-2).
Particularly noteworthy is the change in expression patterns in the hippocampus as the
BACE1 transgenic animals age. While protein expression in the hippocampus is among the highest-expressing brain region at early stages, BACE1 is selectively decreased in the hippocampus in an age-dependent manner. Similarly, olfactory bulb expression is modulated from post-natal stages to adulthood with intermediate expression in young animals to remarkably high expression in aged animals. Unlike the hippocampus and olfactory bulb, relative levels of BACE1 expression in the cerebellum and cortex remain constant throughout the animal lifespan. While the cerebellum consistently exhibits the lowest expression, cortical expression is among the highest-expressing regions for
BACE1 protein levels.
These observations indicate the fluctuating patterns of BACE1 expression are selectively modulated, depending on developmental stage and brain region. These selective regional alterations for human BACE1 expression lead to interesting questions about BACE1 regulation during development. For example, what influences the changes in BACE1 expression as the animal ages? Besides the influence on APP processing
(discussed below), what is the purpose of modifying regional BACE1 expression? These shifting expression patterns offer a glimpse into the regulation of this unique transmembrane aspartyl protease and may potentially lead to further insights into BACE1 function.
123 Influence of the β-secretase genes on APP Processing
The studies in this current thesis were designed to analyze the impact of human β- secretase expression on APP processing in vivo. In particular, we attempted to describe how exogenous expression of BACE1 alters Aβ production in transgenic mice. Although genomic expression of human BACE2 does not alter APP processing (Chapter 4), our studies indicate human BACE1 expression significantly influences Aβ production. These studies describe how slight modulations of β-secretase expression can have dramatic biological consequences on the APP processing pathway. The impact on Aβ production is reflective of the developmental and regional regulation of BACE1 protein. The spatial and temporal expression of BACE1 directly alters the pattern of Aβ production in two ways, (1) by age and (2) by brain region. These alterations in Aβ production have a corresponding impact on Aβ deposition, providing an insight into the role of BACE1 in the APP processing pathway and AD pathogenesis.
Alterations in Temporal and Spatial Aβ Production
Production of Aβ in young animals is significantly increased compared to adult animals (Figure 3-1). This increase of steady-state Aβ levels is possibly reflective of both increased BACE1 and APP expression at young ages (Figure 3-1, 3-3).
Interestingly, increasing the dosage of APP in animals expressing BACE1 does not alter regional Aβ production (Appendix, Figure A1). These observations indicate that the appropriate balance of BACE1 and APP expression is necessary to alter the APP processing pathway.
However, the most striking observation is the influence of regional BACE1 expression on Aβ production in adult animals, both pre- and post-deposition. It has been
124 previously established that APP levels are lower in adult animals (Loffler and Huber,
1992; Basha et al., 2005). Increased cortical Aβ could primarily be a result of increased
BACE1 expression in the cortex in adult animals. This is particularly evident in aged animals where Aβ levels are highest in olfactory bulb and cortex, the two brain regions with the highest BACE1 expression (Figure 2-6). Although increased less than 2-fold,
BACE1 protein expression in the cortex leads to greater than 2-fold elevated Aβ levels in the cortex. Intriguingly, previous studies have determined that Aβ levels in this APP transgenic line (Tg2576) are highest in the hippocampus (Lehman et al., 2003). Thus, regional BACE1 expression levels alone directly influence age-related Aβ production, indicating how significant a role BACE1 plays in the APP processing pathway during aging.
Alterations in Regional Aβ Deposition
Our studies have also shown that BACE1 influences the Aβ neuropathological profile of APP transgenic animals. In the cDNA-based Swedish mutant APP transgenic
(Tg2576), increased cortical BACE1 expression levels directly corresponded with ~3- fold higher cortical Aβ deposits, which is strikingly different from the hippocampal Aβ deposition pattern observed with endogenous levels of Bace1 (Hsiao et al., 1996; Lehman et al., 2003). Similarly, Aβ deposits are strictly limited to the cortex in aged YAC APP transgenics expressing BACE1. Interestingly, this deposition profile is only present in animals homozygous for the YAC APP transgene and never observed in hemizygous
YAC APP animals (data not shown). This indicates that increased dosage of APP might be necessary for Aβ deposition since both homozygous YAC APP transgenics and
Tg2576 APP transgenics express approximately 2-fold and 4-fold higher levels of APP,
125 respectively, than the hemizygous YAC APP transgenics. Thus, while increased APP dosage may not alter Aβ production, a threshold of Aβ accumulation may be required for development of Aβ deposition, which may be reflective of APP expression levels in aged animals. Taken together, although minimal levels of APP are necessary for deposition onset, the current studies indicate BACE1 regional expression may dictate the pattern of
Aβ deposition.
A Model of Human AD Pathogenesis?
Based on our studies, the expression of human BACE1 in APP transgenic animals significantly alters the Aβ deposition profile. Previous studies suggest that Aβ deposits in humans follow a distinct brain-regional profile (Braak and Braak, 1997b). The regional patterns of Aβ deposition due to BACE1 expression correspond with the pattern of Aβ neuropathology in human AD. Thus, according to the amyloid hypothesis, if Aβ does indeed initiate AD pathogenesis, it is possible that BACE1 not only influences the pattern of Aβ deposition, but may also play an integral role in AD pathogenesis. These studies provide invaluable insight into the role of BACE1 in both APP processing and the complex pathogenesis of AD.
There are, however, limitations to this model of human AD that do not fully explain the role of BACE1 in APP processing and AD pathogenesis. The studies described here have explored how modulation of β-secretase activity alters the competitive balance of the APP processing pathway in an age-dependent manner.
Relative expression of BACE1 and APP steady-state levels has been examined in hopes of understanding the critical rate-limiting event involved in Aβ production. Yet, γ- secretase activity also plays a crucial role in the pathway, particularly as the key enzyme
126 involved in the production of the fibrillogenic Aβ1-42/3 peptide. These studies do not explore how, if at all, BACE1 overexpression alters γ-secretase expression and activity.
As an integral member of the γ-secretase complex, does PSEN1 expression change in response to increased β-secretase activity? Given that Aβ production is altered in an age-dependent manner, we have not addressed how the profile of Aβ peptides changes upon the onset of Aβ deposition. Is the ratio of Aβ1-42/3: Aβ1-40 altered in a brain- regional pattern consistent with regional BACE1 protein expression and Aβ deposition?
Alternatively, examining only steady-state protein levels cannot fully explain the relationship between APP processing and Aβ generation. These studies are inherently limited in describing how expression levels of APP or BACE1 relate to Aβ production without knowing the rate of Aβ turnover. Our studies suggest that regional fluctuations in BACE1 expression influence the age-related phenomenon of Aβ deposition, but these studies do not address how degradation of Aβ may change in a regional- or age- dependent manner. For example, our studies have not examined how the brain-regional rate of expression or turnover of APP, Aβ, or even BACE1 changes across the mouse lifespan. Further studies addressing these limitations might help explain how BACE1 regulates APP processing and may offer key clues in understanding AD neuropathology.
Future Directions
The studies described in this thesis characterize alterations in the APP processing pathway upon genomic BACE1 expression, which is responsible for distinct patterns of
Aβ neuropathology. These studies lead to compelling questions regarding the mechanism of BACE1’s role in AD pathogenesis. Future studies proposed here could
127 address these questions and further define how β-secretase transgenic mice represent a model of AD pathogenesis. Specifically, these studies include (1) further analysis of brain-regional alterations in APP processing to examine the properties of amyloid deposition in the BACE1 transgenics, (2) describing the impact of BACE1 overexpression on learning and memory, and (3) investigation of Aβ degradation rates to help explain the age-related regional alterations of APP processing observed in the BACE1 transgenic mice. In addition, further studies involving BACE2 as an alternative α-secretase may provide unique insight into APP processing as well as uncovering a potentially protective agent against AD-like neuropathology.
Characterization of brain-regional APP processing in BACE1xAPP animals
Understanding the influence of regional BACE1 expression on APP processing across the animal lifespan was a major area of investigation throughout this thesis work.
Most notably, we explored the effect of BACE1 regional expression in aged, depositing cDNA-based APP transgenics (Chapter 2). To characterize APP processing in aged, depositing animals expressing the genomic-based YAC APP transgene, it will be necessary to examine regional BACE1 expression patterns and Aβ levels. These studies will complement the age-dependent alterations of brain-regional Aβ production observed in adult BACE1xAPP transgenic animals (Chapter 3). If differences in brain-regional
Aβ production are established in adult animals, then one would expect that Aβ levels in aged animals follow a consistent regional pattern, given our observations regarding the cortical-specific deposition pattern. Using methods described in Chapter 2, one could examine the ratio of soluble: insoluble Aβ peptides as well the production of Aβ1-42 peptides in these aged animals to understand how the biochemical profile of Aβ peptides
128 changes according to brain region. Since Aβ deposits are present only in the cortex of these aged animals, one would expect that insoluble Aβ and the fibrillogenic Aβ1-42/3 peptides would predominate in the cortex relative to other brain regions, such as the hippocampus or cerebellum. These studies would explain how differences in regional
APP processing correspond with the selective cortical Aβ deposition in these aged animals. Yet, these correlative studies do not specifically address how deposition occurs.
To understand the impact of BACE1 on the process of amyloid deposition, one could compare regional BACE1 expression and Aβ production in YAC APPxBACE1 double transgenic animals prior to and subsequent to the onset of deposition.
Specifically, it will be necessary to examine APP phosphorylation patterns in these animals. Recently, Lee et al. (Lee et al., 2005) have demonstrated that BACE1 overexpression influences Aβ deposition by cleaving mature, phosphorylated APP species, while deposition does not occur when BACE1 cleaves immature APP species.
Since cortical deposits predominate in the BACE1+APP double transgenics, one would expect to see an increased proportion of cleaved phosphorylated APP species in the cortex upon, and not before, the onset of deposition. Similarly, one would predict that cleavage of phosphorylated APP is reduced in brain regions where human BACE1 expression is lowest, particularly the hippocampus and cerebellum. These studies could offer unique insight into APP processing and the age-related regional pattern of Aβ deposition observed in the BACE1 transgenic animals. Additionally, if phosphorylation of APP determines whether Aβ deposition will occur, these experiments could provide a mechanism explaining how BACE1 expression influences Aβ deposition.
129 To fully characterize alterations in APP processing, future studies could explore the impact of regional BACE1 expression on Psen1 expression levels. Specifically, does
γ-secretase activity change in response to increased β-secretase expression? Fluctuations in Psen1 may affect the cleavage activity of the γ-secretase complex in a brain regional, age-dependent manner as observed in the BACE1 transgenic animals. If BACE1 overexpression modulates processing of APP by brain region, one would predict that the brain regions with the highest Psen1 expression reflect the regional production of Aβ, especially as amyloid deposition occurs. These experiments, however, make assumptions regarding the relationship between Aβ production and γ-secretase activity. Nevertheless, by examining expression levels of the APP secretases, these studies could help answer the question about which enzyme is rate limiting for Aβ generation in vivo.
Targeted temporal and regional BACE1 expression
To characterize the age-related regional specificity of BACE1’s influence on APP processing, one could employ a targeted transgene inactivation approach to address specific questions about how BACE1 expression regulates age-related Aβ production and deposition by brain region. For example, one could inactivate the BACE1 transgene in specific brain regions and at particular times during the mouse lifespan by the Cre/loxP system combined with an inducible ligand binding protein (Lewandoski, 2001). Using the mouse promoter/enhancer D6 Cre recombinase line (van den Bout et al., 2002), the
BACE1 transgene, fused to the estrogen receptor binding domains and flanked by loxP sites, could be inactivated in the cortex by administering tamofixen in D6-Cre and
BACE1 double transgenic mice. After targeted inactivation in adult animals, one would predict that elimination of BACE1 results in reduced cortical-specific Aβ levels and Aβ
130 deposits. This outcome would further suggest that BACE1 expression is required to alter
APP processing in an age-dependent manner. However, like the previous BACE1 knockout studies (Cai et al., 2001; Luo et al., 2001; Roberds et al., 2001), this experiment does not explain if BACE1 expression is sufficient to induce alterations in Aβ production and deposition. Using this Cre line provides a distinct advantage because the promoter/enhancer D6 drives expression in the cortex and hippocampus. In addition to inactivating the BACE1 transgene, one could drive overexpression in the hippocampus, for example, to determine if ectopic BACE1 expression in vivo is sufficient to induce alterations in APP processing. One distinct disadvantage, however, is that this approach eliminates any consideration of how BACE1 is endogenously regulated. Nonetheless, future studies described here could provide direct evidence as to how BACE1 regulates age-related APP processing in specific brain regions, particularly the regions most affected in AD, and controls regional alterations in Aβ deposition.
Learning and memory analysis in BACE1xAPP animals
The defining clinical feature of Alzheimer’s disease is memory loss and cognitive decline. Numerous behavioral and memory studies have been employed in models of AD to examine the effect of AD-like pathology on specific attributes of learning and memory
(Tremml et al., 1998; Chen et al., 2000; Koistinaho et al., 2001; Lattal et al., 2003). The most popular of these, the Morris water maze (Morris, 1984), measures spatial memory, as mice have to learn to swim to a platform placed in a tank of water. Submersion of the platform is a hippocampal-dependent task that requires the animal to learn and remember spatial cues associated with the location of the hidden platform.
131 To determine if BACE1-induced alterations in Aβ production influence cognitive performance, BACE1xAPP double transgenic mice and single APP transgenic control mice could be analyzed for performance in specific learning and memory tasks such as the Morris water maze. This analysis could address specific questions regarding the regional pattern of Aβ production in the BACE1-expressing animals. Since the double transgenic animals are spared of Aβ deposits in the hippocampus, one would predict that double transgenic animals might perform better in the hippocampal-dependent hidden platform task compared to control animals. In contrast to previous studies describing the rescue of memory deficits upon Bace1 removal in APP transgenics (Ohno et al., 2004),
BACE1 expression may induce learning and memory deficits, providing further insight into the role of altered regional Aβ production on cognitive impairments associated with
AD.
However, since the double transgenics exhibit the most Aβ deposition in the cortex, this assay may not result in any discernible differences between animals expressing BACE1 and control animals. Other behavioral assays, such as the Y-maze, can detect learning deficits of working memory by examining spontaneous alternation throughout the maze arms. Although no deficits were identified for the Morris water maze, learning and memory was significantly altered in YAC APP transgenics animals in the Y-maze task (Hock and Lamb, unpublished results). Since these animals also develop a similar cortical-specific Aβ deposition profile at old ages, the BACE1 transgenics could be analyzed for working memory in the Y-maze. By characterizing learning and memory, these studies could help establish BACE1 transgenics as a model of
AD behavior.
132 Characterization of Aβ turnover in BACE1 transgenics
Primary Neuronal Cultures
As described above, while steady state measurements of Aβ levels offer a glimpse into APP processing, metabolic labeling studies provide a more comprehensive analysis of the generation and turnover of Aβ over time (Busciglio et al., 1993). These studies could more fully characterize brain regional differences of APP processing by examining the relationship between Aβ production and degradation. To examine regional Aβ synthesis and turnover in BACE1 transgenics, cortical and hippocampal neurons would be cultured from embryonic animals expressing BACE1 and APP. These primary neuronal cultures could be incubated with [35S]methionine and cells would be harvested at specific time intervals. The amount of radiolabeled APP, C-terminal fragments, and Aβ could be measured to determine differences between cortical and hippocampal APP processing.
Since steady-state levels of Aβ are higher in the hippocampus at early post-natal ages
(Chapter 3), one would predict that the amount of radiolabeled Aβ is more abundant in hippocampal cultures.
Alternatively, Aβ turnover rates might be altered according to brain region.
Pulse-chase studies might be particularly helpful in characterizing degradation of Aβ in these cultures (Busciglio et al., 1993). Through the incorporation of radiolabeled methionine, generation and turnover of Aβ levels in the presence of unlabeled methionine could be investigated at regular timepoints. This experiment would help determine if there are there regional differences in Aβ turnover across extended pulse intervals and might help explain the regional Aβ production profile observed in young animals. These experiments could uncover distinct patterns of regional Aβ turnover, providing further
133 insight into how BACE1 influences regional APP processing. These studies, however, make specific assumptions regarding Aβ turnover. By harvesting primary neurons in culture, in vitro conditions do not correspond with neuronal function in vivo and may not represent how the brain-regional environment alters Aβ degradation. Additionally, these experiments do not address the impact of BACE1 on age-related Aβ production.
In vivo radiolabeling in adult mice
To characterize the age-related regional differences of Aβ levels in BACE1 transgenics, one could examine Aβ production and turnover by infusion of radiolabeled methionine in adult mice (Savage et al., 1998; Lee et al., 2005). As observed in Chapter
3, increased cortical BACE1 expression leads to significantly elevated Aβ production in the cortex. By infusion of [35S]methionine into the femoral vein of adult mice, Aβ degradation could be examined in each brain region over time. Given that APP processing derivatives turnover within one day (Savage et al., 1998), groups of animals could be separated into regular hourly intervals and the relative amount of APP, C- terminal fragments, and Aβ species could be assayed. These studies could address questions about how BACE1 expression selectively alters regional Aβ production.
Specifically, is Aβ degradation reduced upon increased BACE1 expression? Does the regional pattern of Aβ production in adult animals correspond to the rate of Aβ turnover in each brain region? For example, since cortical Aβ production is increased in adult animals, one would predict that less Aβ is degraded in the cortex relative to other brain regions. Conversely, if there are no differences in Aβ degradation according to brain region, then one could hypothesize that the regional expression of BACE1 establishes and maintains the distinct profile of Aβ production. By examining the “fate” of Aβ from
134 synthesis to turnover, these studies could lay the foundation for understanding how
BACE1 overexpression modulates the APP processing pathway throughout the mammalian lifespan.
Characterization of BACE2 as alternative “α-secretase”
Several studies have identified the unique function of BACE2 in reducing Aβ production (Farzan et al., 2000; Yan et al., 2001a; Basi et al., 2003; Fluhrer et al., 2003).
The approach used in these studies was to analyze BACE2 cleavage of APP in cell culture, whereby BACE2 and APP co-localization is engineered in vitro. In vivo studies have shown that BACE2 and APP do not co-localize in neuronal cells (Bennett et al.,
2000) and thus, BACE2 cannot alter APP processing in the brain. To characterize the
“α-secretase”-like function of BACE2 cleavage in vivo, one could drive neuronal expression of BACE2. For BACE2 neuronal-specific expression, a BACE2 construct can be fused to the neuron-specific rat enolase promoter (Cinato et al., 2001). Under the control of this promoter, BACE2 expression should be present in neuronal populations of the developing and adult brain. By driving expression in neurons, this experiment provides a distinct advantage in trying to determine if BACE2 could effectively compete with Bace1 for cleavage of APP.
To examine the impact of neuronal-specific BACE2 expression on APP processing, these BACE2 transgenic animals could then be crossed to APP transgenic AD models and Aβ levels can be assayed in young and aged mice. If BACE2 preferentially cleaves within the Aβ peptide, then neuronal-specific BACE2 expression would alter amyloidogenic processing of APP and reduce Aβ production, potentially leading to significantly reduced Aβ deposition compared to control APP transgenics. This approach
135 specifically addresses how neuronal-specific BACE2 overexpression could alter APP processing. Other approaches provide unique advantages that could address other questions about BACE2 function. For example, one could create a knock-in mouse with the BACE2 cDNA inserted into the BACE1 locus. This knock-in transgenic could specifically address how BACE2 functions in the context of the predicted β-secretase expression pattern (represented by BACE1). Although these approaches make assumptions about the endogenous regulation of BACE2, these analyses may provide further insight into unique aspects of the regulation of the APP processing pathway as mediated by β-secretase and, in this case, an α-secretase-like protein. These “ectopic”
BACE2 transgenics also present a unique opportunity to test the amyloid hypothesis by exploring the relationship between Aβ generation and AD pathogenesis.
136 Appendix
Figure A1: Aβ production by brain region in hemizgyous BACE1 x homozygous APP transgenics. Aβ1-40 ELISA analysis of brain region extracts from P7 (A) and P60 (B) BACE1/Homo- R1.40 YAC APP double transgenic mice (n= 12) and age-matched Homo-R1.40 YAC APP single transgenic mice (n= 12). Total Aβ1-40 levels do not differ in olfactory bulb, cerebellum, hippocampus, or cortex at either age. Increased dosage of APP in the context of human BACE1 overexpression does not alter Aβ production.
137
Figure A2: Profile of Aβ species in human BACE1 transgenics by immunoprecipitation/mass spectrometry.
A, IP/MS analysis of brain extracts from 12-month old BACE1xTg2576 animals. Whole brain extracts were homogenized in phosphate-buffered saline and were immunoprecipitated using mAb 468 antibody, which is directed towards aa 17-24 of Aβ sequence. Peptides were analyzed by matrix-assisted, laser desorption/ionization, time of flight mass spectrometry as previously described (Wang et al., 1996). Left, Graph representing the peak intensities of pooled samples from BACE1/Tg2576 double transgenics (n= 6) and Tg2576 transgenics alone (n= 6) for Aβ1-40 and Aβ1-42 peptides. Right, Synthetic Aβ12-28 was added to the extracts before immunoprecipitation and used to quantify the relative amount of Aβ1-40 or Aβ1-42 species. No significant differences were identified for either Aβ1-40 or Aβ1-42. B, Aβx-40 ELISA analysis of whole brain extracts from 2-month old BACE1x R1.40 YAC APP transgenic animals. Brain extracts were analyzed with an Aβ ELISA assay kit (Immuno-Biological Laboratories, Japan) that uses an antibody against Aβ11-28 as secondary conjugated antibody to detect all Aβ40 peptides generated, particularly any N- terminally-truncated Aβ species. There were no significant differences between the BACE1/R1.40 transgenic animals (n= 9) and R1.40 transgenics alone (n= 9).
138 Bibliography
Acquati F, Accarino M, Nucci C, Fumagalli P, Jovine L, Ottolenghi S, Taramelli R (2000) The gene encoding DRAP (BACE2), a glycosylated transmembrane protein of the aspartic protease family, maps to the down critical region. FEBS Lett 468:59-64.
Allsop D, Haga S, Bruton C, Ishii T, Roberts GW (1990) Neurofibrillary tangles in some cases of dementia pugilistica share antigens with amyloid beta-protein of Alzheimer's disease. Am J Pathol 136:255-260.
Ancolio K, Dumanchin C, Barelli H, Warter JM, Brice A, Campion D, Frebourg T, Checler F (1999) Unusual phenotypic alteration of beta amyloid precursor protein (betaAPP) maturation by a new Val-715 --> Met betaAPP-770 mutation responsible for probable early-onset Alzheimer's disease. Proc Natl Acad Sci U S A 96:4119-4124.
Ashford JW (2004) APOE genotype effects on Alzheimer's disease onset and epidemiology. J Mol Neurosci 23:157-165.
Ashford JW, Mortimer JA (2002) Non-familial Alzheimer's disease is mainly due to genetic factors. J Alzheimers Dis 4:169-177.
Association As (2004) Basic Facts for Alzheimer's Disease. In.
Basha MR, Wei W, Bakheet SA, Benitez N, Siddiqi HK, Ge YW, Lahiri DK, Zawia NH (2005) The fetal basis of amyloidogenesis: exposure to lead and latent overexpression of amyloid precursor protein and beta-amyloid in the aging brain. J Neurosci 25:823-829.
Basi G, Frigon N, Barbour R, Doan T, Gordon G, McConlogue L, Sinha S, Zeller M (2003) Antagonistic effects of beta-site amyloid precursor protein-cleaving enzymes 1 and 2 on beta-amyloid peptide production in cells. J Biol Chem 278:31512-31520.
Benjannet S, Elagoz A, Wickham L, Mamarbachi M, Munzer JS, Basak A, Lazure C, Cromlish JA, Sisodia S, Checler F, Chretien M, Seidah NG (2001) Post- translational processing of beta-secretase (beta-amyloid-converting enzyme) and
139 its ectodomain shedding. The pro- and transmembrane/cytosolic domains affect its cellular activity and amyloid-beta production. J Biol Chem 276:10879-10887.
Bennett BD, Babu-Khan S, Loeloff R, Louis JC, Curran E, Citron M, Vassar R (2000) Expression analysis of BACE2 in brain and peripheral tissues. J Biol Chem 275:20647-20651.
Bertram L, Blacker D, Mullin K, Keeney D, Jones J, Basu S, Yhu S, McInnis MG, Go RC, Vekrellis K, Selkoe DJ, Saunders AJ, Tanzi RE (2000) Evidence for genetic linkage of Alzheimer's disease to chromosome 10q. Science 290:2302-2303.
Bodendorf U, Fischer F, Bodian D, Multhaup G, Paganetti P (2001) A splice variant of beta-secretase deficient in the amyloidogenic processing of the amyloid precursor protein. J Biol Chem 276:12019-12023.
Bodendorf U, Danner S, Fischer F, Stefani M, Sturchler-Pierrat C, Wiederhold KH, Staufenbiel M, Paganetti P (2002) Expression of human beta-secretase in the mouse brain increases the steady-state level of beta-amyloid. J Neurochem 80:799-806.
Bossy-Wetzel E, Schwarzenbacher R, Lipton SA (2004) Molecular pathways to neurodegeneration. Nat Med 10 Suppl:S2-9.
Braak H, Braak E (1997a) Diagnostic criteria for neuropathologic assessment of Alzheimer's disease. Neurobiol Aging 18:S85-88.
Braak H, Braak E (1997b) Frequency of stages of Alzheimer-related lesions in different age categories. Neurobiol Aging 18:351-357.
Breteler MM, Claus JJ, van Duijn CM, Launer LJ, Hofman A (1992) Epidemiology of Alzheimer's disease. Epidemiol Rev 14:59-82.
Brinton RD (2004) Impact of estrogen therapy on Alzheimer's disease: a fork in the road? CNS Drugs 18:405-422.
Brookmeyer R, Gray S, Kawas C (1998) Projections of Alzheimer's disease in the United States and the public health impact of delaying disease onset. Am J Public Health 88:1337-1342.
140 Busciglio J, Gabuzda DH, Matsudaira P, Yankner BA (1993) Generation of beta-amyloid in the secretory pathway in neuronal and nonneuronal cells. Proc Natl Acad Sci U S A 90:2092-2096.
Buxbaum JD, Liu KN, Luo Y, Slack JL, Stocking KL, Peschon JJ, Johnson RS, Castner BJ, Cerretti DP, Black RA (1998) Evidence that tumor necrosis factor alpha converting enzyme is involved in regulated alpha-secretase cleavage of the Alzheimer amyloid protein precursor. J Biol Chem 273:27765-27767.
Cai H, Wang Y, McCarthy D, Wen H, Borchelt DR, Price DL, Wong PC (2001) BACE1 is the major beta-secretase for generation of Abeta peptides by neurons. Nat Neurosci 4:233-234.
Cao X, Sudhof TC (2001) A transcriptionally [correction of transcriptively] active complex of APP with Fe65 and histone acetyltransferase Tip60. Science 293:115- 120.
Capell A, Meyn L, Fluhrer R, Teplow DB, Walter J, Haass C (2002) Apical sorting of beta-secretase limits amyloid beta-peptide production. J Biol Chem 277:5637- 5643.
Capell A, Steiner H, Willem M, Kaiser H, Meyer C, Walter J, Lammich S, Multhaup G, Haass C (2000) Maturation and pro-peptide cleavage of beta-secretase. J Biol Chem 275:30849-30854.
Chen G, Chen KS, Knox J, Inglis J, Bernard A, Martin SJ, Justice A, McConlogue L, Games D, Freedman SB, Morris RG (2000) A learning deficit related to age and beta-amyloid plaques in a mouse model of Alzheimer's disease. Nature 408:975- 979.
Chiocco MJ, Kulnane LS, Younkin L, Younkin S, Evin G, Lamb BT (2004) Altered Amyloid-{beta} Metabolism and Deposition in Genomic-based {beta}-Secretase Transgenic Mice. J Biol Chem 279:52535-52542.
Christen-Zaech S, Kraftsik R, Pillevuit O, Kiraly M, Martins R, Khalili K, Miklossy J (2003) Early olfactory involvement in Alzheimer's disease. Can J Neurol Sci 30:20-25.
141 Christensen MA, Zhou W, Qing H, Lehman A, Philipsen S, Song W (2004) Transcriptional regulation of BACE1, the beta-amyloid precursor protein beta- secretase, by Sp1. Mol Cell Biol 24:865-874.
Cinato E, Mirotsou M, Sablitzky F (2001) Cre-mediated transgene activation in the developing and adult mouse brain. Genesis 31:118-125.
Citron M, Oltersdorf T, Haass C, McConlogue L, Hung AY, Seubert P, Vigo-Pelfrey C, Lieberburg I, Selkoe DJ (1992) Mutations of the β-amyloid precursor protein in familial Alzheimer's disease increase β-protein production. Nature 360:672-674.
Clarimon J, Bertranpetit J, Calafell F, Boada M, Tarraga L, Comas D (2003) Association study between Alzheimer's disease and genes involved in Abeta biosynthesis, aggregation and degradation: suggestive results with BACE1. J Neurol 250:956- 961.
Cordy JM, Hussain I, Dingwall C, Hooper NM, Turner AJ (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 100:11735-11740.
Cruts M, Dermaut B, Rademakers R, Roks G, Van den Broeck M, Munteanu G, van Duijn CM, Van Broeckhoven C (2001) Amyloid beta secretase gene (BACE) is neither mutated in nor associated with early-onset Alzheimer's disease. Neurosci Lett 313:105-107.
De Strooper B, Saftig P, Craessaerts K, Vanderstichele H, Guhde G, Annaert W, Von Figura K, Van Leuven F (1998) Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein. Nature 391:387-390.
De Strooper B, Annaert W, Cupers P, Saftig P, Craessaerts K, Mumm JS, Schroeter EH, Schrijvers V, Wolfe MS, Ray WJ, Goate A, Kopan R (1999) A presenilin-1- dependent gamma-secretase-like protease mediates release of Notch intracellular domain. Nature 398:518-522.
Deane R, Du Yan S, Submamaryan RK, LaRue B, Jovanovic S, Hogg E, Welch D, Manness L, Lin C, Yu J, Zhu H, Ghiso J, Frangione B, Stern A, Schmidt AM, Armstrong DL, Arnold B, Liliensiek B, Nawroth P, Hofman F, Kindy M, Stern D, Zlokovic B (2003) RAGE mediates amyloid-beta peptide transport across the blood-brain barrier and accumulation in brain. Nat Med 9:907-913.
142 Delatour B, Blanchard V, Pradier L, Duyckaerts C (2004) Alzheimer pathology disorganizes cortico-cortical circuitry: direct evidence from a transgenic animal model. Neurobiol Dis 16:41-47.
DeMattos RB, O'Dell M A, Parsadanian M, Taylor JW, Harmony JA, Bales KR, Paul SM, Aronow BJ, Holtzman DM (2002) Clusterin promotes amyloid plaque formation and is critical for neuritic toxicity in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A 99:10843-10848.
Dickson TC, Vickers JC (2001) The morphological phenotype of beta-amyloid plaques and associated neuritic changes in Alzheimer's disease. Neuroscience 105:99-107.
Doty RL, Reyes PF, Gregor T (1987) Presence of both odor identification and detection deficits in Alzheimer's disease. Brain Res Bull 18:597-600.
Duff K, Eckman C, Zehr C, Yu X, Prada CM, Perez-tur J, Hutton M, Buee L, Harigaya Y, Yager D, Morgan D, Gordon MN, Holcomb L, Refolo L, Zenk B, Hardy J, Younkin S (1996) Increased amyloid-beta42(43) in brains of mice expressing mutant presenilin 1. Nature 383:710-713.
Eckman CB, Mehta ND, Crook R, Perez-tur J, Prihar G, Pfeiffer E, Graff-Radford N, Hinder P, Yager D, Zenk B, Refolo LM, Prada CM, Younkin SG, Hutton M, Hardy J (1997) A new pathogenic mutation in the APP gene (I716V) increases the relative proportion of A beta 42(43). Hum Mol Genet 6:2087-2089.
Ehehalt R, Keller P, Haass C, Thiele C, Simons K (2003) Amyloidogenic processing of the Alzheimer beta-amyloid precursor protein depends on lipid rafts. J Cell Biol 160:113-123.
Farris W, Mansourian S, Leissring MA, Eckman EA, Bertram L, Eckman CB, Tanzi RE, Selkoe DJ (2004) Partial loss-of-function mutations in insulin-degrading enzyme that induce diabetes also impair degradation of amyloid beta-protein. Am J Pathol 164:1425-1434.
Farzan M, Schnitzler CE, Vasilieva N, Leung D, Choe H (2000) BACE2, a β-secretase homolog, cleaves at the β site and within the amyloid-β region of the amyloid-β precursor protein. Proc Natl Acad Sci U S A 97:9712-9717.
143 Fischer F, Molinari M, Bodendorf U, Paganetti P (2002) The disulphide bonds in the catalytic domain of BACE are critical but not essential for amyloid precursor protein processing activity. J Neurochem 80:1079-1088.
Fluhrer R, Capell A, Westmeyer G, Willem M, Hartung B, Condron MM, Teplow DB, Haass C, Walter J (2002) A non-amyloidogenic function of BACE-2 in the secretory pathway. J Neurochem 81:1011-1020.
Fluhrer R, Multhaup G, Schlicksupp A, Okochi M, Takeda M, Lammich S, Willem M, Westmeyer G, Bode W, Walter J, Haass C (2003) Identification of a beta- secretase activity, which truncates amyloid beta-peptide after its presenilin- dependent generation. J Biol Chem 278:5531-5538.
Foundation AHA (2004) The Facts on Alzheimer's Disease. In.
Fukumoto H, Cheung BS, Hyman BT, Irizarry MC (2002) Beta-secretase protein and activity are increased in the neocortex in Alzheimer disease. Arch Neurol 59:1381-1389.
Fukumoto H, Rosene DL, Moss MB, Raju S, Hyman BT, Irizarry MC (2004) Beta- secretase activity increases with aging in human, monkey, and mouse brain. Am J Pathol 164:719-725.
Galton CJ, Patterson K, Xuereb JH, Hodges JR (2000) Atypical and typical presentations of Alzheimer's disease: a clinical, neuropsychological, neuroimaging and pathological study of 13 cases. Brain 123 Pt 3:484-498.
Gatz M, Fratiglioni L, Johansson B, Berg S, Mortimer JA, Reynolds CA, Fiske A, Pedersen NL (2005) Complete ascertainment of dementia in the Swedish Twin Registry: the HARMONY study. Neurobiol Aging 26:439-447.
Glenner GG, Wong CW (1984) Alzheimer's disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochemical and Biophysical Research Communications 120:885-890.
Goate A, Chartier-Harlin M-C, Mullan M, Brown J, Crawford F, Fidani L, Giuffra L, Haynes, A., Irving N, James L, Mant R, Newton P, Rooke K, Roques P, Talbot C, Perciak-Vance M, Roses A, Williamson R, Rossor M, Owen M, Hardy J (1991) Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer's disease. Nature 349:704-706.
144 Gold G, Blouin JL, Herrmann FR, Michon A, Mulligan R, Duriaux Sail G, Bouras C, Giannakopoulos P, Antonarakis SE (2003) Specific BACE1 genotypes provide additional risk for late-onset alzheimer disease in APOE varepsilon 4 carriers. Am J Med Genet 119B:44-47.
Goldgaber D, Lerman MI, McBride OW, Saffiotti U, Gajdusek DC (1987) Characterization and chromosomal localization of a cDNA encoding brain amyloid of Alzheimer's disease. Science 235:877-880.
Grabowski TJ, Cho HS, Vonsattel JP, Rebeck GW, Greenberg SM (2001) Novel amyloid precursor protein mutation in an Iowa family with dementia and severe cerebral amyloid angiopathy. Ann Neurol 49:697-705.
Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M, Wisniewski HM, Binder LI (1986) Abnormal phosphorylation of the microtubule-associated protein tau (tau) in Alzheimer cytoskeletal pathology. Proc Natl Acad Sci U S A 83:4913-4917.
Guenette SY, Tanzi RE (1999) Progress toward valid transgenic mouse models for Alzheimer's disease. Neurobiol Aging 20:201-211.
Haass C, Hung AY, Selkoe DJ, Teplow DB (1994) Mutations associated with a locus for familial Alzheimer's disease result in alternative processing of amyloid β-protein precursor. Journal of Biological Chemistry 269:17741-17748.
Han X, D MH, McKeel DW, Jr., Kelley J, Morris JC (2002) Substantial sulfatide deficiency and ceramide elevation in very early Alzheimer's disease: potential role in disease pathogenesis. J Neurochem 82:809-818.
Haniu M, Denis P, Young Y, Mendiaz EA, Fuller J, Hui JO, Bennett BD, Kahn S, Ross S, Burgess T, Katta V, Rogers G, Vassar R, Citron M (2000) Characterization of Alzheimer's β-secretase protein BACE. A pepsin family member with unusual properties. J Biol Chem 275:21099-21106.
Hardy J, Selkoe DJ (2002) The amyloid hypothesis of Alzheimer's disease: progress and problems on the road to therapeutics. Science 297:353-356.
Haroutunian V, Purohit DP, Perl DP, Marin D, Khan K, Lantz M, Davis KL, Mohs RC (1999) Neurofibrillary tangles in nondemented elderly subjects and mild Alzheimer disease. Arch Neurol 56:713-718.
145 He W, Lu Y, Qahwash I, Hu XY, Chang A, Yan R (2004) Reticulon family members modulate BACE1 activity and amyloid-beta peptide generation. Nat Med 10:959- 965.
Hendriks L, van Duijn CM, Cras P, Cruts M, Van Hul W, van Harskamp F, Warren A, McInnis MG, Antonarakis SE, Martin JJ, et al. (1992) Presenile dementia and cerebral haemorrhage linked to a mutation at codon 692 of the beta-amyloid precursor protein gene. Nat Genet 1:218-221.
Herms J, Anliker B, Heber S, Ring S, Fuhrmann M, Kretzschmar H, Sisodia S, Muller U (2004) Cortical dysplasia resembling human type 2 lissencephaly in mice lacking all three APP family members. Embo J 23:4106-4115.
Hogan B, Beddington R, Constantini F, Lacy E (1994) Manipulating the Mouse Embryo, 2nd Edition. Cold Spring Harbor: Cold Spring Harbor Laboratory Press.
Holsinger RM, McLean CA, Beyreuther K, Masters CL, Evin G (2002) Increased expression of the amyloid precursor beta-secretase in Alzheimer's disease. Ann Neurol 51:783-786.
Hsiao K, Chapman P, Nilsen S, Eckman C, Harigaya Y, S. Y, Yang F, Cole G (1996) Correlative memory deficits, Aβ elevation and amyoid plaques in transgenic mice. Science 274:99-102.
Hsiao KK, Borchelt DR, Olson K, Johannsdottir R, Kitt C, Yunis W, Xu S, Eckman C, Younkin S, Price D, et al. (1995) Age-related CNS disorder and early death in transgenic FVB/N mice overexpressing Alzheimer amyloid precursor proteins. Neuron 15:1203-1218.
Huse JT, Pijak DS, Leslie GJ, Lee VM, Doms RW (2000) Maturation and endosomal targeting of beta-site amyloid precursor protein-cleaving enzyme. The Alzheimer's disease beta-secretase. J Biol Chem 275:33729-33737.
Huse JT, Liu K, Pijak DS, Carlin D, Lee VM, Doms RW (2002) Beta-secretase processing in the trans-Golgi network preferentially generates truncated amyloid species that accumulate in Alzheimer's disease brain. J Biol Chem 277:16278- 16284.
Huse JT, Byant D, Yang Y, Pijak DS, D'Souza I, Lah JJ, Lee VM, Doms RW, Cook DG (2003) Endoproteolysis of beta-secretase (beta-site amyloid precursor protein-
146 cleaving enzyme) within its catalytic domain. A potential mechanism for regulation. J Biol Chem 278:17141-17149.
Hussain I, Powell D, Howlett DR, Tew DG, Meek TD, Chapman C, Gloger IS, Murphy KE, Southan CD, Ryan DM, Smith TS, Simmons DL, Walsh FS, Dingwall C, Christie G (1999) Identification of a novel aspartic protease (Asp 2) as beta- secretase. Mol Cell Neurosci 14:419-427.
Hussain I, Powell DJ, Howlett DR, Chapman GA, Gilmour L, Murdock PR, Tew DG, Meek TD, Chapman C, Schneider K, Ratcliffe SJ, Tattersall D, Testa TT, Southan C, Ryan DM, Simmons DL, Walsh FS, Dingwall C, Christie G (2000) ASP1 (BACE2) cleaves the amyloid precursor protein at the beta-secretase site. Mol Cell Neurosci 16:609-619.
Irizarry MC, Locascio JJ, Hyman BT (2001) beta-site APP cleaving enzyme mRNA expression in APP transgenic mice: anatomical overlap with transgene expression and static levels with aging. Am J Pathol 158:173-177.
Iwata N, Tsubuki S, Takaki Y, Shirotani K, Lu B, Gerard NP, Gerard C, Hama E, Lee HJ, Saido TC (2001) Metabolic regulation of brain Abeta by neprilysin. Science 292:1550-1552.
Jankowsky JL, Fadale DJ, Anderson J, Xu GM, Gonzales V, Jenkins NA, Copeland NG, Lee MK, Younkin LH, Wagner SL, Younkin SG, Borchelt DR (2004) Mutant presenilins specifically elevate the levels of the 42 residue beta-amyloid peptide in vivo: evidence for augmentation of a 42-specific gamma secretase. Hum Mol Genet 13:159-170.
Jellinger KA, Bancher C (1997) Proposals for re-evaluation of current autopsy criteria for the diagnosis of Alzheimer's disease. Neurobiol Aging 18:S55-65.
Johnson-Wood K, Lee M, Motter R, Hu K, Gordon G, Barbour R, Khan K, Gordon M, Tan H, Games D, Lieberburg I, Schenk D, Seubert P, McConlogue L (1997) Amyloid precursor protein processing and A beta42 deposition in a transgenic mouse model of Alzheimer disease. Proc Natl Acad Sci U S A 94:1550-1555.
Kamal A, Stokin GB, Yang Z, Xia CH, Goldstein LS (2000) Axonal transport of amyloid precursor protein is mediated by direct binding to the kinesin light chain subunit of kinesin-I. Neuron 28:449-459.
147 Kamal A, Almenar-Queralt A, LeBlanc JF, Roberts EA, Goldstein LS (2001) Kinesin- mediated axonal transport of a membrane compartment containing beta-secretase and presenilin-1 requires APP. Nature 414:643-648.
Kamino K, Orr HT, Payami H, Wijsman EM, Alonso ME, Pulst SM, Anderson L, O'Dahl S, Nemens E, White JA, et al. (1992) Linkage and mutational analysis of familial Alzheimer disease kindreds for the APP gene region. Am J Hum Genet 51:998- 1014.
Kandel ER, Schwartz JH, Jessel TM, eds (2000) Principles of Neural Science, Fourth Edition, Fourth Edition Edition. New York: McGraw-Hill.
Kang DE, Pietrzik CU, Baum L, Chevallier N, Merriam DE, Kounnas MZ, Wagner SL, Troncoso JC, Kawas CH, Katzman R, Koo EH (2000) Modulation of amyloid beta-protein clearance and Alzheimer's disease susceptibility by the LDL receptor-related protein pathway. J Clin Invest 106:1159-1166.
Kawarabayashi T, Younkin LH, Saido TC, Shoji M, Ashe KH, Younkin SG (2001) Age- dependent changes in brain, CSF, and plasma amyloid (beta) protein in the Tg2576 transgenic mouse model of Alzheimer's disease. J Neurosci 21:372-381.
Kidd M (1963) Paired helical filaments in electron microscopy of Alzheimer's disease. Nature 197:192-193.
Kidd M (1964) Alzheimer's Disease--An Electron Microscopical Study. Brain 87:307- 320.
Kinoshita A, Whelan CM, Berezovska O, Hyman BT (2002) The gamma secretase- generated carboxyl-terminal domain of the amyloid precursor protein induces apoptosis via Tip60 in H4 cells. J Biol Chem 277:28530-28536.
Kinoshita A, Fukumoto H, Shah T, Whelan CM, Irizarry MC, Hyman BT (2003) Demonstration by FRET of BACE interaction with the amyloid precursor protein at the cell surface and in early endosomes. J Cell Sci 116:3339-3346.
Kitazume S, Tachida Y, Oka R, Shirotani K, Saido TC, Hashimoto Y (2001) Alzheimer's beta-secretase, beta-site amyloid precursor protein-cleaving enzyme, is responsible for cleavage secretion of a Golgi-resident sialyltransferase. Proc Natl Acad Sci U S A 98:13554-13559.
148 Koistinaho M, Ort M, Cimadevilla JM, Vondrous R, Cordell B, Koistinaho J, Bures J, Higgins LS (2001) Specific spatial learning deficits become severe with age in beta -amyloid precursor protein transgenic mice that harbor diffuse beta -amyloid deposits but do not form plaques. Proc Natl Acad Sci U S A 98:14675-14680.
Koo EH, Sisodia SS, Archer DR, Martin LJ, Weidemann A, Beyreuther K, Fischer P, Masters CL, Price DL (1990) Precursor of amyloid protein in Alzheimer disease undergoes fast anterograde axonal transport. Proc Natl Acad Sci U S A 87:1561- 1565.
Kovacs T (2004) Mechanisms of olfactory dysfunction in aging and neurodegenerative disorders. Ageing Res Rev 3:215-232.
Kovacs T, Cairns NJ, Lantos PL (2001) Olfactory centres in Alzheimer's disease: olfactory bulb is involved in early Braak's stages. Neuroreport 12:285-288.
Krezowski J, Knudson D, Ebeling C, Pitstick R, Giri RK, Schenk D, Westaway D, Younkin L, Younkin SG, Ashe KH, Carlson GA (2004) Identification of loci determining susceptibility to the lethal effects of amyloid precursor protein transgene overexpression. Hum Mol Genet 13:1989-1997.
Kulnane LS, Lamb BT (2001) Neuropathological characterization of mutant amyloid precursor protein yeast artificial chromosome transgenic mice. Neurobiol Dis 8:982-992.
Lahiri DK, Nall C, Chen D, Zaphiriou M, Morgan C, Nurnberger JI, Sr. (2002) Developmental expression of the beta-amyloid precursor protein and heat-shock protein 70 in the cerebral hemisphere region of the rat brain. Ann N Y Acad Sci 965:324-333.
Lai F, Williams RS (1989) A prospective study of Alzheimer disease in Down syndrome. Arch Neurol 46:849-853.
Lamb BT, Bardel KA, Kulnane LS, Anderson JJ, Holtz G, Wagner SL, Sisodia SS, Hoeger EJ (1999) Amyloid production and deposition in mutant amyloid precursor protein and presenilin-1 yeast artificial chromosome transgenic mice. Nature Neuroscience 2:695-697.
149 Lamb BT, Sisodia SS, Lawler AM, Slunt HH, Kitt CA, Kearns WG, Pearson PL, Price DL, Gearhart JD (1993) Introduction and expression of the 400 kilobase amyloid precursor protein gene in transgenic mice [corrected]. Nat Genet 5:22-30.
Lamb BT, Call LM, Slunt HH, Bardel KA, Lawler AM, Eckman CB, Younkin SG, Holtz G, Wagner SL, Price DL, Sisodia SS, Gearhart JD (1997) Altered metabolism of familial Alzheimer's disease-linked precursor protein variants in yeast artificial chromosome transgenic mice. Human Molecular Genetics 6:1535-1541.
Lammich S, Kojro E, Postina R, Gilbert S, Pfeiffer R, Jasionowski M, Haass C, Fahrenholz F (1999) Constitutive and regulated alpha-secretase cleavage of Alzheimer's amyloid precursor protein by a disintegrin metalloprotease. Proc Natl Acad Sci U S A 96:3922-3927.
Lattal KM, Mullen MT, Abel T (2003) Extinction, renewal, and spontaneous recovery of a spatial preference in the water maze. Behav Neurosci 117:1017-1028.
Lee EB, Skovronsky DM, Abtahian F, Doms RW, Lee VM (2003) Secretion and intracellular generation of truncated Abeta in beta-site amyloid-beta precursor protein-cleaving enzyme expressing human neurons. J Biol Chem 278:4458-4466.
Lee EB, Zhang B, Liu K, Greenbaum EA, Doms RW, Trojanowski JQ, Lee VM (2005) BACE overexpression alters the subcellular processing of APP and inhibits A{beta} deposition in vivo. J Cell Biol 168:291-302.
Lehman EJ, Kulnane LS, Lamb BT (2003) Alterations in beta-amyloid production and deposition in brain regions of two transgenic models. Neurobiol Aging 24:645- 653.
Leissring MA, Farris W, Chang AY, Walsh DM, Wu X, Sun X, Frosch MP, Selkoe DJ (2003) Enhanced proteolysis of beta-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death. Neuron 40:1087- 1093.
Lemere CA, Blusztajn JK, Yamaguchi H, Wisniewski T, Saido TC, Selkoe DJ (1996) Sequence of deposition of heterogeneous amyloid beta-peptides and APO E in Down syndrome: implications for initial events in amyloid plaque formation. Neurobiol Dis 3:16-32.
150 Lendon CL, Ashall F, Goate AM (1997) Exploring the etiology of Alzheimer disease using molecular genetics. JAMA 277:825-831.
Levitan D, Greenwald I (1995) Facilitation of lin-12-mediated signalling by sel-12, a Caenorhabditis elegans S182 Alzheimer's disease gene. Nature 377:351-354.
Levy-Lahad E, Wasco W, Poorkaj P, Romano DM, Oshima J, Pettingell WH, Yu C-E, Jondro PD, Schmidt SD, Wang K, Crowley AC, Fu Y-H, Guenette SY, Galas D, Nemens E, Wijsman EM, Bird TD, Schellenberg GD (1995) Candidate gene for the chromosome 1 familial Alzheimer's disease locus. Science 269:973-977.
Lewandoski M (2001) Conditional control of gene expression in the mouse. Nat Rev Genet 2:743-755.
Li R, Shen Y (2005) Estrogen and brain: synthesis, function and diseases. Front Biosci 10:257-267.
Li R, Lindholm K, Yang LB, Yue X, Citron M, Yan R, Beach T, Sue L, Sabbagh M, Cai H, Wong P, Price D, Shen Y (2004) Amyloid beta peptide load is correlated with increased beta-secretase activity in sporadic Alzheimer's disease patients. Proc Natl Acad Sci U S A 101:3632-3637.
Lichtenthaler SF, Dominguez DI, Westmeyer GG, Reiss K, Haass C, Saftig P, De Strooper B, Seed B (2003) The cell adhesion protein P-selectin glycoprotein ligand-1 is a substrate for the aspartyl protease BACE1. J Biol Chem 278:48713- 48719.
Liu HC, Leu SJ, Chang JG, Sung SM, Hsu WC, Lee LS, Hu CJ (2003) The association of beta-site APP cleaving enzyme (BACE) C786G polymorphism with Alzheimer's disease. Brain Res 961:88-91.
Liu K, Doms RW, Lee VM (2002) Glu11 site cleavage and N-terminally truncated A beta production upon BACE overexpression. Biochemistry 41:3128-3136.
Loffler J, Huber G (1992) Beta-amyloid precursor protein isoforms in various rat brain regions and during brain development. J Neurochem 59:1316-1324.
Lopez-Perez E, Seidah NG, Checler F (1999) Proprotein convertase activity contributes to the processing of the Alzheimer's beta-amyloid precursor protein in human
151 cells: evidence for a role of the prohormone convertase PC7 in the constitutive alpha-secretase pathway. J Neurochem 73:2056-2062.
Lue LF, Kuo YM, Roher AE, Brachova L, Shen Y, Sue L, Beach T, Kurth JH, Rydel RE, Rogers J (1999) Soluble amyloid beta peptide concentration as a predictor of synaptic change in Alzheimer's disease. Am J Pathol 155:853-862.
Luo Y, Bolon B, Kahn S, Bennett BD, Babu-Khan S, Denis P, Fan W, Kha H, Zhang J, Gong Y, Martin L, Louis JC, Yan Q, Richards WG, Citron M, Vassar R (2001) Mice deficient in BACE1, the Alzheimer's beta-secretase, have normal phenotype and abolished beta-amyloid generation. Nat Neurosci 4:231-232.
Marcinkiewicz M, Seidah NG (2000) Coordinated expression of beta-amyloid precursor protein and the putative beta-secretase BACE and alpha-secretase ADAM10 in mouse and human brain. J Neurochem 75:2133-2143.
Marr RA, Rockenstein E, Mukherjee A, Kindy MS, Hersh LB, Gage FH, Verma IM, Masliah E (2003) Neprilysin gene transfer reduces human amyloid pathology in transgenic mice. J Neurosci 23:1992-1996.
Masters CL, Simms G, Weinman NA, Multhaup G, McDonald BL, Beyreuther K (1985) Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc Natl Acad Sci U S A 82:4245-4249.
Miller SA, Dykes DD, Polesky HF (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res 16:1215.
Mohajeri MH, Saini KD, Nitsch RM (2004) Transgenic BACE expression in mouse neurons accelerates amyloid plaque pathology. J Neural Transm 111:413-425.
Morelli L, Llovera R, Ibendahl S, Castano EM (2002) The degradation of amyloid beta as a therapeutic strategy in Alzheimer's disease and cerebrovascular amyloidoses. Neurochem Res 27:1387-1399.
Morris R (1984) Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Methods 11:47-60.
152 Mullan M, Crawford F, Axelman K, Houlden H, Lilius L, Winblad B, Lannfelt L (1992) A pathogenetic mutation for probable Alzheimer's disease in the APP gene at the N-terminus of β-amyloid. Nature Genetics 1:345-347.
Naslund J, Haroutunian V, Mohs R, Davis KL, Davies P, Greengard P, Buxbaum JD (2000) Correlation between elevated levels of amyloid beta-peptide in the brain and cognitive decline. Jama 283:1571-1577.
Naslund J, Schierhorn A, Hellman U, Lannfelt L, Roses AD, Tjernberg LO, Silberring J, Gandy SE, Winblad B, Greengard P, et al. (1994) Relative abundance of Alzheimer A beta amyloid peptide variants in Alzheimer disease and normal aging. Proc Natl Acad Sci U S A 91:8378-8382.
Nestor PJ, Scheltens P, Hodges JR (2004) Advances in the early detection of Alzheimer's disease. Nat Med 10 Suppl:S34-41.
Nicolaou M, Song YQ, Sato CA, Orlacchio A, Kawarai T, Medeiros H, Liang Y, Sorbi S, Richard E, Rogaev EI, Moliaka Y, Bruni AC, Jorge R, Percy M, Duara R, Farrer LA, St Georg-Hyslop P, Rogaeva EA (2001) Mutations in the open reading frame of the beta-site APP cleaving enzyme (BACE) locus are not a common cause of Alzheimer's disease. Neurogenetics 3:203-206.
Nowotny P, Kwon JM, Chakraverty S, Nowotny V, Morris JC, Goate AM (2001) Association studies using novel polymorphisms in BACE1 and BACE2. Neuroreport 12:1799-1802.
Ohno M, Sametsky EA, Younkin LH, Oakley H, Younkin SG, Citron M, Vassar R, Disterhoft JF (2004) BACE1 Deficiency Rescues Memory Deficits and Cholinergic Dysfunction in a Mouse Model of Alzheimer's Disease. Neuron 41:27-33.
Pasalar P, Najmabadi H, Noorian AR, Moghimi B, Jannati A, Soltanzadeh A, Krefft T, Crook R, Hardy J (2002) An Iranian family with Alzheimer's disease caused by a novel APP mutation (Thr714Ala). Neurology 58:1574-1575.
Paxinos GF, Keith B.J. (2001) The Mouse Brain in Stereotaxic Coordinates, Second Edition Edition. San Diego: Academic Press.
Payami H, Zareparsi S, Montee KR, Sexton GJ, Kaye JA, Bird TD, Yu CE, Wijsman EM, Heston LL, Litt M, Schellenberg GD (1996) Gender difference in
153 apolipoprotein E-associated risk for familial Alzheimer disease: a possible clue to the higher incidence of Alzheimer disease in women. Am J Hum Genet 58:803- 811.
Peacock ML, Murman DL, Sima AA, Warren JT, Jr., Roses AD, Fink JK (1994) Novel amyloid precursor protein gene mutation (codon 665Asp) in a patient with late- onset Alzheimer's disease. Ann Neurol 35:432-438.
Pike CJ, Overman MJ, Cotman CW (1995) Amino-terminal deletions enhance aggregation of beta-amyloid peptides in vitro. J Biol Chem 270:23895-23898.
Postina R, Schroeder A, Dewachter I, Bohl J, Schmitt U, Kojro E, Prinzen C, Endres K, Hiemke C, Blessing M, Flamez P, Dequenne A, Godaux E, van Leuven F, Fahrenholz F (2004) A disintegrin-metalloproteinase prevents amyloid plaque formation and hippocampal defects in an Alzheimer disease mouse model. J Clin Invest 113:1456-1464.
Prasher VP, Farrer MJ, Kessling AM, Fisher EM, West RJ, Barber PC, Butler AC (1998) Molecular mapping of Alzheimer-type dementia in Down's syndrome. Ann Neurol 43:380-383.
Probst A, Anderton BH, Brion JP, Ulrich J (1989) Senile plaque neurites fail to demonstrate anti-paired helical filament and anti-microtubule-associated protein- tau immunoreactive proteins in the absence of neurofibrillary tangles in the neocortex. Acta Neuropathol (Berl) 77:430-436.
Puglielli L, Ellis BC, Saunders AJ, Kovacs DM (2003) Ceramide stabilizes beta-site amyloid precursor protein-cleaving enzyme 1 and promotes amyloid beta-peptide biogenesis. J Biol Chem 278:19777-19783.
Pype S, Moechars D, Dillen L, Mercken M (2003) Characterization of amyloid beta peptides from brain extracts of transgenic mice overexpressing the London mutant of human amyloid precursor protein. J Neurochem 84:602-609.
Qing H, Zhou W, Christensen MA, Sun X, Tong Y, Song W (2004) Degradation of BACE by the ubiquitin-proteasome pathway. Faseb J 18:1571-1573.
Refolo LM, Malester B, LaFrancois J, Bryant-Thomas T, Wang R, Tint GS, Sambamurti K, Duff K, Pappolla MA (2000) Hypercholesterolemia accelerates the
154 Alzheimer's amyloid pathology in a transgenic mouse model. Neurobiol Dis 7:321-331.
Riddell DR, Christie G, Hussain I, Dingwall C (2001) Compartmentalization of beta- secretase (Asp2) into low-buoyant density, noncaveolar lipid rafts. Curr Biol 11:1288-1293.
Riley KP, Snowdon DA, Markesbery WR (2002) Alzheimer's neurofibrillary pathology and the spectrum of cognitive function: findings from the Nun Study. Ann Neurol 51:567-577.
Roberds SL, Anderson J, Basi G, Bienkowski MJ, Branstetter DG, Chen KS, Freedman SB, Frigon NL, Games D, Hu K, Johnson-Wood K, Kappenman KE, Kawabe TT, Kola I, Kuehn R, Lee M, Liu W, Motter R, Nichols NF, Power M, Robertson DW, Schenk D, Schoor M, Shopp GM, Shuck ME, Sinha S, Svensson KA, Tatsuno G, Tintrup H, Wijsman J, Wright S, McConlogue L (2001) BACE knockout mice are healthy despite lacking the primary beta- secretase activity in brain: implications for Alzheimer's disease therapeutics. Hum Mol Genet 10:1317-1324.
Rocchi A, Pellegrini S, Siciliano G, Murri L (2003) Causative and susceptibility genes for Alzheimer's disease: a review. Brain Res Bull 61:1-24.
Rubinsztein DC (1997) The genetics of Alzheimer's disease. Progress in Neurobiology 52:447-454.
Sambrook JaR, David W. (2001) Molecular Cloning: A Laboratory Manual, Third Edition Edition. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press.
Saunders AJ, Kim T-W, Tanzi RE (1999) BACE maps to chromosome 11 and a BACE Homolog, BACE2, reside in the obligate Down Syndrome region of chromosome 21. Science 286:1255a.
Savage MJ, Trusko SP, Howland DS, Pinsker LR, Mistretta S, Reaume AG, Greenberg BD, Siman R, Scott RW (1998) Turnover of amyloid beta-protein in mouse brain and acute reduction of its level by phorbol ester. J Neurosci 18:1743-1752.
Scheuner D, Eckman C, Jensen M, Song X, Citron M, Suzuki N, Bird TD, Hardy J, Hutton M, Kukull W, Larson E, Levy-Lahad E, Viitanen M, Peskind E, Poorkaj
155 P, Schellenberg G, Tanzi R, Wasco W, Lannfelt L, Selkoe D, Younkin S (1996) Secreted amyloid beta-protein similar to that in the senile plaques of Alzheimer's disease is increased in vivo by the presenilin 1 and 2 and APP mutations linked to familial Alzheimer's disease [see comments]. Nat Med 2:864-870.
Scholefield Z, Yates EA, Wayne G, Amour A, McDowell W, Turnbull JE (2003) Heparan sulfate regulates amyloid precursor protein processing by BACE1, the Alzheimer's beta-secretase. J Cell Biol 163:97-107.
Selkoe DJ (1998) The cell biology of beta-amyloid precursor protein and presenilin in Alzheimer's disease. Trends Cell Biol 8:447-453.
Selkoe DJ (1999) Translating cell biology into therapeutic advances in Alzheimer's disease. Nature 399:A23-31.
Selkoe DJ (2001) Alzheimer's disease: genes, proteins, and therapy. Physiol Rev 81:741- 766.
Serby M, Larson P, Kalkstein D (1991) The nature and course of olfactory deficits in Alzheimer's disease. Am J Psychiatry 148:357-360.
Serby M, Mohan C, Aryan M, Williams L, Mohs RC, Davis KL (1996) Olfactory identification deficits in relatives of Alzheimer's disease patients. Biol Psychiatry 39:375-377.
Sheng JG, Price DL, Koliatsos VE (2003) The beta-amyloid-related proteins presenilin 1 and BACE1 are axonally transported to nerve terminals in the brain. Exp Neurol 184:1053-1057.
Sherrington R, Rogaev EI, Liang Y, Rogaeva EA, Levesque G, Ikeda M, Chi H, Lin C, Li G, Holman K, Tsuda T, Mar L, Foncin J-F, Bruni AC, Montesi MP, Sorbi S, Rainero I, Pinessi L, Nee L, Chumakov I, Pollen D, Brookes A, Sanseau P, Polinsky RJ, Wasco W, Da Silva HAR, Haines JL, Pericak-Vance MA, Tanzi RE, Roses AD, Fraser PE, Rommens JM, St. George-Hyslop PH (1995) Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease. Nature 375:754-760.
Shi J, Zhang S, Tang M, Liu X, Li T, Wang Y, Han H, Guo Y, Hao Y, Zheng K, Kong X, Su Z, Tong Y, Ma C (2004) The 1239G/C polymorphism in exon 5 of BACE1
156 gene may be associated with sporadic Alzheimer's disease in Chinese Hans. Am J Med Genet B Neuropsychiatr Genet 124:54-57.
Shi XP, Tugusheva K, Bruce JE, Lucka A, Wu GX, Chen-Dodson E, Price E, Li Y, Xu M, Huang Q, Sardana MK, Hazuda DJ (2003) Beta-secretase cleavage at amino acid residue 34 in the amyloid beta peptide is dependent upon gamma-secretase activity. J Biol Chem 278:21286-21294.
Shibata M, Yamada S, Kumar SR, Calero M, Bading J, Frangione B, Holtzman DM, Miller CA, Strickland DK, Ghiso J, Zlokovic BV (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 106:1489-1499.
Sinha S, Anderson JP, Barbour R, Basi GS, Caccavello R, Davis D, Doan M, Dovey HF, Frigon N, Hong J, Jacobson-Croak K, Jewett N, Keim P, Knops J, Lieberburg I, Power M, Tan H, Tatsuno G, Tung J, Schenk D, Seubert P, Suomensaari SM, Wang S, Walker D, Zhao J, McConlogue L, John V (1999) Purification and cloning of amyloid precursor protein β-secretase from human brain. Nature 402:537-540.
Snowdon DA, Greiner LH, Markesbery WR (2000) Linguistic ability in early life and the neuropathology of Alzheimer's disease and cerebrovascular disease. Findings from the Nun Study. Ann N Y Acad Sci 903:34-38.
Snowdon DA, Kemper SJ, Mortimer JA, Greiner LH, Wekstein DR, Markesbery WR (1996) Linguistic ability in early life and cognitive function and Alzheimer's disease in late life. Findings from the Nun Study. Jama 275:528-532.
St George-Hyslop PH, Tanzi RE, Polinsky RJ, Haines JL, Nee L, Watkins PC, Myers RH, Feldman RG, Pollen D, Drachman D, et al. (1987) The genetic defect causing familial Alzheimer's disease maps on chromosome 21. Science 235:885-890.
Stern Y, Gurland B, Tatemichi TK, Tang MX, Wilder D, Mayeux R (1994) Influence of education and occupation on the incidence of Alzheimer's disease. Jama 271:1004-1010.
Struhl G, Greenwald I (1999) Presenilin is required for activity and nuclear access of Notch in Drosophila. Nature 398:522-525.
157 Sun A, Koelsch G, Tang J, Bing G (2002) Localization of beta-secretase memapsin 2 in the brain of Alzheimer's patients and normal aged controls. Exp Neurol 175:10- 22.
Suzuki N, Cheung TT, Cai XD, Odaka A, Otvos L, Jr., Eckman C, Golde TE, Younkin SG (1994) An increased percentage of long amyloid beta protein secreted by familial amyloid beta protein precursor (beta APP717) mutants. Science 264:1336-1340.
Takahashi M, Dore S, Ferris CD, Tomita T, Sawa A, Wolosker H, Borchelt DR, Iwatsubo T, Kim SH, Thinakaran G, Sisodia SS, Snyder SH (2000) Amyloid precursor proteins inhibit heme oxygenase activity and augment neurotoxicity in Alzheimer's disease. Neuron 28:461-473.
Tanzi RE (1999) A genetic dichotomy model for the inheritance of Alzheimer's disease and common age-related disorders. J Clin Invest 104:1175-1179.
Tanzi RE, Moir RD, Wagner SL (2004) Clearance of Alzheimer's Abeta peptide: the many roads to perdition. Neuron 43:605-608.
Tanzi RE, Gusella JF, Watkins PC, Bruns GA, St George-Hyslop P, Van Keuren ML, Patterson D, Pagan S, Kurnit DM, Neve RL (1987) Amyloid beta protein gene: cDNA, mRNA distribution, and genetic linkage near the Alzheimer locus. Science 235:880-884.
Terry RD (1963) The Fine Structure Of Neurofibrillary Tangles In Alzheimer's Disease. J Neuropathol Exp Neurol 22:629-642.
Terry RD, Gonatas NK, Weiss M (1964) The ultrastructure of the cerebral cortex in Alzheimer's disease. Trans Am Neurol Assoc 89:12.
Thinakaran G (1999) The role of presenilins in Alzheimer's disease. J Clin Invest 104:1321-1327.
Tomita T, Maruyama K, Saido TC, Kume H, Shinozaki K, Tokuhiro S, Capell A, Walter J, Grunberg J, Haass C, Iwatsubo T, Obata K (1997) The presenilin 2 mutation (N141I) linked to familial Alzheimer disease (Volga German families) increases the secretion of amyloid beta protein ending at the 42nd (or 43rd) residue [see comments]. Proc Natl Acad Sci U S A 94:2025-2030.
158 Tremml P, Lipp HP, Muller U, Ricceri L, Wolfer DP (1998) Neurobehavioral development, adult openfield exploration and swimming navigation learning in mice with a modified beta-amyloid precursor protein gene. Behav Brain Res 95:65-76.
Tyler SJ, Dawbarn D, Wilcock GK, Allen SJ (2002) alpha- and beta-secretase: profound changes in Alzheimer's disease. Biochem Biophys Res Commun 299:373-376. van den Bout CJ, Machon O, Rosok O, Backman M, Krauss S (2002) The mouse enhancer element D6 directs Cre recombinase activity in the neocortex and the hippocampus. Mech Dev 110:179-182. van Duijn CM, Tanja TA, Haaxma R, Schulte W, Saan RJ, Lameris AJ, Antonides- Hendriks G, Hofman A (1992) Head trauma and the risk of Alzheimer's disease. Am J Epidemiol 135:775-782.
Van Gassen G, Annaert W, Van Broeckhoven C (2000) Binding partners of Alzheimer's disease proteins: are they physiologically relevant? Neurobiol Dis 7:135-151.
Vassar R, Bennett BD, Babu-Khan S, Kahn S, Mendiaz EA, Denis P, Teplow DB, Ross S, Amarante P, Loeloff R, Luo Y, Fisher S, Fuller J, Edenson S, Lile J, Jarosinski MA, Biere AL, Curran E, Burgess T, Louis JC, Collins F, Treanor J, Rogers G, Citron M (1999) β-Secretase cleavage of Alzheimer's amyloid precursor protein by the transmembrane aspartic protease BACE. Science 286:735-741. von Koch CS, Zheng H, Chen H, Trumbauer M, Thinakaran G, van der Ploeg LH, Price DL, Sisodia SS (1997) Generation of APLP2 KO mice and early postnatal lethality in APLP2/APP double KO mice. Neurobiol Aging 18:661-669.
Wang R, Sweeney D, Gandy SE, Sisodia SS (1996) The profile of soluble amyloid beta protein in cultured cell media. Detection and quantification of amyloid beta protein and variants by immunoprecipitation-mass spectrometry. J Biol Chem 271:31894-31902.
Welsh KA, Butters N, Hughes JP, Mohs RC, Heyman A (1992) Detection and staging of dementia in Alzheimer's disease. Use of the neuropsychological measures developed for the Consortium to Establish a Registry for Alzheimer's Disease. Arch Neurol 49:448-452.
159 Willem M, Dewachter I, Smyth N, Van Dooren T, Borghgraef P, Haass C, Van Leuven F (2004) beta-Site amyloid precursor protein cleaving enzyme 1 increases amyloid deposition in brain parenchyma but reduces cerebrovascular amyloid angiopathy in aging BACE x APP[V717I] double-transgenic mice. Am J Pathol 165:1621- 1631.
Wisniewski KE, Wisniewski HM, Wen GY (1985) Occurrence of neuropathological changes and dementia of Alzheimer's disease in Down's syndrome. Ann Neurol 17:278-282.
Wolfe MS, Kopan R (2004) Intramembrane proteolysis: theme and variations. Science 305:1119-1123.
Xu H, Sweeney D, Wang R, Thinakaran G, Lo AC, Sisodia SS, Greengard P, Gandy S (1997) Generation of Alzheimer beta-amyloid protein in the trans-Golgi network in the apparent absence of vesicle formation. Proc Natl Acad Sci U S A 94:3748- 3752.
Yan R, Munzner JB, Shuck ME, Bienkowski MJ (2001a) BACE2 functions as an alternative alpha-secretase in cells. J Biol Chem 276:34019-34027.
Yan R, Han P, Miao H, Greengard P, Xu H (2001b) The transmembrane domain of the Alzheimer's beta-secretase (BACE1) determines its late Golgi localization and access to beta -amyloid precursor protein (APP) substrate. J Biol Chem 276:36788-36796.
Yan R, Bienkowski MJ, Shuck ME, Miao H, Tory MC, Pauley AM, Brashler JR, Stratman NC, Mathews WR, Buhl AE, Carter DB, Tomasselli AG, Parodi LA, Heinrikson RL, Gurney ME (1999) Membrane-anchored aspartyl protease with Alzheimer's disease β-secretase activity. Nature 402:533-537.
Yang HC, Chai X, Mosior M, Kohn W, Boggs LN, Erickson JA, McClure DB, Yeh WK, Zhang L, Gonzalez-DeWhitt P, Mayer JP, Martin JA, Hu J, Chen SH, Bueno AB, Little SP, McCarthy JR, May PC (2004) Biochemical and kinetic characterization of BACE1: investigation into the putative species-specificity for beta- and beta'- cleavage sites by human and murine BACE1. J Neurochem 91:1249-1259.
Yang LB, Lindholm K, Yan R, Citron M, Xia W, Yang XL, Beach T, Sue L, Wong P, Price D, Li R, Shen Y (2003) Elevated beta-secretase expression and enzymatic activity detected in sporadic Alzheimer disease. Nat Med 9:3-4.
160 Yang XW, Model P, Heintz N (1997) Homologous recombination based modification in Escherichia coli and germline transmission in transgenic mice of a bacterial artificial chromosome. Nat Biotechnol 15:859-865.
Yankner BA, Dawes LR, Fisher S, Villa-Komaroff L, Oster-Granite ML, Neve RL (1989) Neurotoxicity of a fragment of the amyloid precursor associated with Alzheimer's disease. Science 245:417-420.
Yoshikawa K, Aizawa T, Hayashi Y (1992) Degeneration in vitro of post-mitotic neurons overexpressing the Alzheimer amyloid protein precursor. Nature 359:64-67.
Yu H, Saura CA, Choi SY, Sun LD, Yang X, Handler M, Kawarabayashi T, Younkin L, Fedeles B, Wilson MA, Younkin S, Kandel ER, Kirkwood A, Shen J (2001) APP processing and synaptic plasticity in presenilin-1 conditional knockout mice. Neuron 31:713-726.
Zhang MY, Katzman R, Salmon D, Jin H, Cai GJ, Wang ZY, Qu GY, Grant I, Yu E, Levy P, et al. (1990) The prevalence of dementia and Alzheimer's disease in Shanghai, China: impact of age, gender, and education. Ann Neurol 27:428-437.
Zheng H, Jiang M, Trumbauer ME, Hopkins R, Sirinathsinghji DJ, Stevens KA, Conner MW, Slunt HH, Sisodia SS, Chen HY, Van der Ploeg LH (1996) Mice deficient for the amyloid precursor protein gene. Ann N Y Acad Sci 777:421-426.
Zimmermann M, Gardoni F, Marcello E, Colciaghi F, Borroni B, Padovani A, Cattabeni F, Di Luca M (2004) Acetylcholinesterase inhibitors increase ADAM10 activity by promoting its trafficking in neuroblastoma cell lines. J Neurochem 90:1489- 1499.
Zohar O, Cavallaro S, D'Agata V, Alkon DL (2003) Quantification and distribution of beta-secretase alternative splice variants in the rat and human brain. Brain Res Mol Brain Res 115:63-68.
161