Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Expression of human ABCG1 in mice modulates cholesterol metabolite levels and does not influence guanidine-extractable… Burgess, Braydon 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2007-0040a.pdf [ 8.35MB ]
Metadata
JSON: 831-1.0100915.json
JSON-LD: 831-1.0100915-ld.json
RDF/XML (Pretty): 831-1.0100915-rdf.xml
RDF/JSON: 831-1.0100915-rdf.json
Turtle: 831-1.0100915-turtle.txt
N-Triples: 831-1.0100915-rdf-ntriples.txt
Original Record: 831-1.0100915-source.json
Full Text
831-1.0100915-fulltext.txt
Citation
831-1.0100915.ris

Full Text

E X P R E S S I O N OF HUMAN ABCG1 IN MICE MODULATES C H O L E S T E R O L METABOLITE LEVELS AND DOES NOT INFLUENCE GUANIDINE-EXTRACTABLE ABETA LEVELS IN BRAIN by Braydon Burgess B.Sc. Honors, Molecular Biology and Biochemistry, Simon Fraser University, 2003 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Pathology and Laboratory Medicine) THE UNIVERSITY OF BRITISH COLUMBIA April 2007 © Braydon Leslie Burgess, 2007 ABSTRACT Growing evidence suggests that cholesterol plays a key role in Alzheimer's Disease (AD). Importantly, membrane cholesterol levels influence generation of the amyloid-beta (Ali) peptide from amyloid precursor protein (APP). The amyloid cascade hypothesis of A D posits that increased accumulation of (AB) peptides and their deposition as amyloid in the brain, promote neurodegeneration in AD. The ATP-binding-cassette-G1 (ABCG1) transporter is highly expressed in the brain and periphery and is believed to regulate cholesterol levels in peripheral cells and tissues. This suggests that the ABCG1-mediated effects on cholesterol homeostasis may influence AU production and AD. To determine the effect of A B C G 1 overexpression in brain and to evaluate whether ABCG1 could influence AS production or amyloid deposition, A B C G 1 overexpression was studied both in vitro and using a novel A B C G 1 transgenic mouse. We generated an A B C G 1 overexpressing mouse containing the entire A B C G 1 genomic locus and regulatory sequence, to promote "physiological" expression patterns of overexpression. We found that the A B C G 1 transgene is physiologically regulated in vivo and overexpressed 3-6-fold at the protein level. Using a LacZ reporter driven by endogenous A B C G 1 regulatory sequences, we showed that A B C G 1 is highly expressed in hippocampal neurons. Importantly, ABCG1 overexpressing mice express the A B C G 1 transgene in primary neurons and astrocytes. Unexpectedly, the level of cholesterol biosynthetic precursors and 24S-hydroxycholesterol levels were reduced in adult ABCG1 overexpressing brain, suggesting a novel function of A B C G 1 . Total levels of cholesterol and other lipids were unaffected. In contrast to numerous published studies that studied constitutive overexpression of ABCG1 in vitro, overexpression of physiologically regulated A B C G 1 in primary astrocytes did not facilitate cholesterol efflux to HDL. These experiments suggest that the role of ABCG1 in ii vivo is not entirely understood and that ABCG1 may influence cholesterol biosynthesis or cholesterol precursor transport. A B C G 1 overexpression increased A P P processing by B-secretase and the production of Af i in transfected HEK cells. However, in vivo A B C G 1 overexpression or deficiency did not significantly affect murine AR levels nor did overexpression affect amyloid formation or fi-secretase mediated A P P cleavage. These experiments suggest that A B C G 1 overexpression does not impact amyloid formation as hypothesized. TABLE OF CONTENTS A B S T R A C T ii LIST OF T A B L E S Vii LIST OF F IGURES Viii LIST OF ABBREVIATIONS: ix A C K N O W L E D G E M E N T S Xi CO-AUTHORSHIP STATEMENT XM 1. B A C K G R O U N D 1 1.1. Alzheimer's Disease 1 1.1.1. Alzheimer's Disease - epidemiology and impact 1 1.1.2. A D and the brain 1 1.1.3. A D neuropathology 2 1.1.4. Tau 2 1.1.5. Amyloid plaques 3 1.1.6. Amyloid and Inflammation 3 1.1.7. The amyloid cascade hypothesis and familial AD 4 1.1.8. Amyloid Precursor Protein 5 1.1.9. A P P Processing 5 1.2. Cholesterol 7 1.2.1. Cholesterol in mammalian cells 7 1.2.2. Cholesterol and the brain: The blood brain barrier 8 1.2.3. Cholesterol and the brain: metabolism ..: 9 1.2.4. Cholesterol and Alzheimer's Disease 11 1.2.4.1. Apolipoprotein E 11 1.2.4.2. Serum Cholesterol 12 1.2.4.3. Statins 14 1.3. Down Syndrome and AD 16 1.4. ABC Transporters 17 1.4.1. The ATP-Binding-Cassette superfamily 17 1.4.2. The A B C G 1 transporter.. 17 1.4.3. Transcriptional regulation of Abcg l 19 1.4.4. Transporter assembly 19 1.4.5. A B C G 1 function 20 iv 1.4.6. Human disease and Abcgl 21 1.5. AD mouse models 22 1.5.1. Modeling AD: Amyloid vs Tau 22 1.5.2. The PDAPP Mouse model of AD 22 1.6. Research Hypothesis and Rationale 23 1.6.1. Rationale 23 1.6.2. Experimental Hypothesis and Specific Aims 23 2. CHAPTER 2: THE ROLE OF ABCG1 IN BRAIN AND AD 25 2.1. EXPERIMENTAL REAGENTS AND PROCEDURES 25 2.1.1. Transgenic Animals 25 2.1.1.1. Generation of ABCG1 BAC Transgenic Mice 25 2.1.1.2. PDAPP Mice 25 2.1.1.3. Abcgl-deficient mice: 25 2.1.1.4. Generation of Experimental Cohorts 26 2.1.1.5. Diet 26 2.1.1.6. Sacrifice and Tissue Preparation ' 26 2.1.2. Cell Culture 27 2.1.2.1. HEK 293 Cells 27 2.1.2.2. Primary Astrocyte Culture 27 2.1.2.3. Primary Neuron Culture 28 2.1.3. Protein Analysis 29 2.1.3.1. Sample Preparation 29 2.1.3.2. Membrane Fraction Preparation 29 2.1.3.3. Protein Determination assay 29 2.1.3.4. SDS-Polyacrylamide Gel Electrophoresis (SDS PAGE) 30 2.1.3.5. Western blotting 31 2.1.4. Messenger RNA Analysis 33 2.1.4.1. mRNA Extraction 33 2.1.4.2. QRT-PCR 33 2.1.5. Lipid Analysis 34 2.1.5.1. Total Lipid Analysis 34 2.1.5.2. Cholesterol metabolites 35 2.1.5.3. Cholesterol efflux 36 2.1.6. LacZ staining 37 2.1.7. AB Measurements 37 2.1.7.1. AB Secretion Assays 37 2.1.7.2. Murine AB measurements 38 2.1.7.3. Human AB measurements 38 2.1.8. ApoEELISA 39 2.1.9. Statistical Analyses: 39 2.2. RESULTS 40 2.2.1. ABCG1 overexpression in vitro 40 2.2.1.1. ABCG1 overexpression increases cholesterol efflux to HDL 40 2.2.1.2. ABCG1 overexpression promotes AB secretion in vitro 40 2.2.1.3. ABCG1 overexpression increases APP processing in vitro 42 2.2.2. ABCG1 expression in brain 44 2.2.2.1. Abcgl is highly expressed in hippocampal neurons 44 v 2.2.3. Generation of a novel animal model of ABCG1 overexpression 45 2.2.3.1. Transgenic strategy 47 2.2.3.2. Expression of the Abcgl transgene is physiologically regulated in vivo 48 2.2.3.3. Murine and human expression of Abcgl 49 2.2.3.4. The ABCG1 BAC is expressed in neurons 49 2.2.3.5. ABCG1 transgenic mice overexpress ABCG1 protein 3-8 fold in peripheral tissues 52 2.2.3.6. ABCG1 is overexpressed 3-6-fold in cortex, hippocampus and cerebellum. 53 2.2.3.7. ABCG1 remains overexpressed in PDAPP mice with amyloid 53 2.2.4. Characterization of the ABCG1 BAC transgenic mice 55 2.2.4.1. Cholesterol efflux to HDL is not increased from ABCG1 Tg astrocytes 55 2.2.4.2. Abcgl expression is increased following acetylated LDL treatment 56 2.2.4.3. Steady-state lipid levels are not changed in ABCG1 BAC Tg brain 58 2.2.4.4. Cholesterol precursors and metabolites are reduced in ABCG1 Tg brains... 59 2.2.4.5. Apolipoprotein E levels are unchanged in brain of ABCG1 transgenic mice. 61 2.2.5. ABCG1 overexpression in vivo, AB production and amyloid 63 2.2.5.1. ABCG1 overexpression does not influence amyloid deposition in vivo 64 2.2.5.2. Total APP and C99 levels are not affected by ABCG1 overexpression 66 2.2.5.3. Murine AB levels are not affected by ABCG1 protein levels 66 3.1. Introduction 70 3.2. Discussion 70 3.2.1. ABCG1 overexpression and AB production in vitro 70 3.2.2. The ABCG1 BAC transgenic mouse 72 3.2.3. Human and murine Abcgl expression 73 3.2.4. Cell-specific expression of Abcgl in brain 73 3.2.5. The in vivo function of ABCG1 74 3.2.5.1. Cholesterol efflux: 74 3.2.5.2. Brain lipid levels 76 3.2.5.3. Cholesterol metabolites in brain of ABCG1 BAC transgenic mice 77 3.2.6. ABCG1 and ApoE 78 3.2.7. ABCG1 and AD 78 3.2.7.1. ABCG1 and amyloid 78 3.2.7.2. ABCG1 and Murine AB 79 3.2.8. Conclusion: The function of ABCG1 in brain 79 3.2.9. Conclusion: The role of ABCG1 in AD 80 3.3. Future Directions 81 3.3.1. Determine how ABCG1 regulates cholesterol metabolite levels 81 3.3.2. Determine the cell-specific patterns of expression in brain 81 3.3.3. Determine the relevant Abcgl isoforms in brain and their function 82 3.3.4. Test the functionality of the human ABCG1 BAC in ABCG1 knockout mice 82 REFERENCES: 83 APPENDIX 1: Copy of UBC Animal Care Certificates 103 vi LIST OF TABLES Table 2-1: Tissue specific levels of Abcg l mRNA in humans and mice 51 Table 2-2: Total lipid levels in brain are not significantly affected by A B C G 1 overexpression 60 Table 2-3: Cholesterol precursors and 24S-hydroxycholesterol levels are reduced in brain of adult A B C G 1 overexpressing mice but not very young mice 62 LIST OF FIGURES Figure 1-1: Processing of the amyloid precuror protein by secretases 6 Figure 1-1: Summary of the cholesterol biosynthetic pathway 10 Figure 1-2: Predicted topology of t h e A B C G I monomer 18 Figure 2-1: A B C G 1 overexpression increases cholesterol efflux to HDL and increases the secretion of AIS40 and AB42 41 Figure 2-2: A B C G 1 overexpression increases cellular A P P levels and increases the processing of A P P into CTFs 43 Figure 2-3: Abcg l expression in brain 46 Figure 2-4: A B C G 1 BAC expression in brain of transgenic founders 48 Figure 2-5 The tissue-specific pattern of the ABCG1 BAC transgene expression is similar to murine Abcg l 50 Figure 2-6: The A B C G 1 BAC is expressed in primary neurons and responds to an LXR agonist 52 Figure 2-7: A B C G 1 protein is overexpressed in ABCG1 BAC Tg tissues 54 Figure 2-8: A B C G 1 overexpression in primary astrocytes does not promote cholesterol efflux to HDL 57 Figure 2-9: ApoE levels are not affected by ABCG1 overexpression 63 Figure 2-10: Soluble and insoluble human AB levels in the hippocampus of P D A P P mice either wildtype or overexpressing A B C G 1 65 Figure 2-11: C99 and total A P P levels are not affected by A B C G 1 overexpression 67 Figure 2-12: A B C G 1 overexpression or deficiency do not affect murine A l l levels in brain 69 viii LIST OF ABBREVIATIONS: Af i Amyloid-beta peptide A B C G 1 ATP-Binding Cassette G1 ABCA1 ATP-Binding Cassette A1 AcLDL Acetylated low-density lipoprotein A D Alzheimer's Disease AICD A P P intracellular domain Apo Apolipoprotein A P P S-Amyloid precursor protein A T P Adenosine triphosphate BAC Bacterial artificial chromosome BACE-1 Beta-site A P P cleaving enzyme-1 BBB Blood brain barrier B S A Bovine serum albumin IJMCD beta-methylcyclodextrin C83 83-amino acid CTF C99 99-amino acid CTF C N S Central nervous system C S F Cerebrospinal fluid C T F C-terminal fragment of A P P D M E M Dulbecco's modified eagle medium DS Down Syndrome ELISA Enzyme-linked immunosorbent assay ER Endoplasmic Reticulum FAD Familial Alzheimer's Disease F B S Fetal bovine serum G C : M S Gas chromatography : Mass spectrometry G T P Guanosine triphosphate HDL High-density lipoprotein HEK Human embryonic kidney H M G R 3-hydroxy-3-methyl-glutaryl-CoA reductase LAC-Z Beta-galactosidase LDL Low-density lipoprotein LXR Liver-X-receptor L X R E Liver-X-receptor response element MAC Membrane attack complex mRNA Messenger ribonucleic acid NFT Neurofibrillary tangle NSAID Non-Steroidal anti-inflammatory drug P B S Phosphate buffered saline PDGF- f i Platelet derived growth factor-beta PPAR -y Peroxisome proliferator-activated receptor -gamma Q R T - P C R Quantitative real-time polymerase chain reaction R X R Retinoid-X-receptor s A P P a Soluble A P P from alpha cleavage s A P P B Soluble A P P from beta cleavage S D S Sodium dodecylsulfate Tg Transgenic ACKNOWLEDGEMENTS First and foremost I am indebted to my graduate supervisor Dr. Cheryl Wellington for her support, creativity and vision that made this research and my personal growth throughout this degree a reality. Above all I have appreciated her positivity, lead by example approach, commitment to her students and taste for fine Scotch. What started out as a 4 month co-operative education semester has been over three years of intense personal growth and rewarding personal achievements and throughout the entire process Cheryl has provided support and direction above and beyond the call of duty. For this I am truly grateful. I would also like to acknowledge my supervisory committee, who have provided invaluable guidance and feedback on this research and letters of reference. The members of the Wellington lab, who provided immeasurable assistance, feedback and whose company turned a room full of toxic chemicals, whirring mechanical devices and endless microfuge tubes into a home away from home. Of note, these last years would have been much more difficult and less enjoyable without Sean Mclsaac, Jeniffer Chan, Kate Naus, Gavin Tansley, Jon Pearce, George Ou, Pam Parkinson, Anna Wilkinson and James Donkin. I would like to make special note of fellow graduate student Veronica Hirsch-Reinshagen who has been an inspiration throughout this entire journey. Vero's coolheaded approach and genuine spirit is refreshing and her competency in all aspects of research has motivated me to aim higher. This work would not be the same without her. xi CO-AUTHORSHIP STATEMENT The experiments presented in this thesis were planned and performed by the author with the following exceptions. Q R T - P C R Measurements were performed by Jeniffer Chan. Total brain lipids were extracted by the author, however, quantification of lipid species were performed in collaboration with Dr. Sheila Innis, Department of Nutrition, University of British Columbia. Cholesterol metabolite levels were measured in collaboration with Dr. Dieter Lutjohann, Department of Clinical Pharmacology, University of Bonn. All studies involving HEK 293 cells were performed by the author, however, representative experiments shown were performed in conjunction with Gavin Tansley. Murine and human AR. extraction and measurements were performed in collaboration with Dr. Ronald DeMattos, Lilly Research Laboratories, Indianapolis. xii BACKGROUND 1.1. Alzheimer's Disease 1.1.1. Alzheimer's Disease - epidemiology and impact Alzheimer's Disease (AD) is a neurodegenerative disease characterized by progressive loss of cognitive function and, eventually premature death 1 2 . A D is the most common form of senile dementia affecting more than 4.5 million North Americans in 2003, triple the number diagnosed in 1980 and by 2050, more than 15 million people may be affected with AD 3 . The average age of an individual diagnosed with probable A D is 73 years of a g e 4 , and the average life span following a positive diagnosis is 4-6 years, or half of the average life expectancy 2 . Current therapeutic options for mild to severe A D exist and fall into two classes: acetylcholinesterase inhibitors such as donepezil or galantamine or the NMDA receptor agonist memantine 5 . Both classes of drugs target neurotransmitter dysfunction that occurs in AD, however, neither of these classes of drugs are considered "disease modifying" therapies 6 . The benefits of these drugs are slight, resulting in short-term increases in measures of cognition, functioning and daily living that, in some studies, are ineffective at reducing patient c a r e 6 . The current drug regimens do not preclude the development of more effective treatments that by one estimate, could reduce the prevalence of A D by 50% if successful in delaying the onset of A D by 5 yea rs 1 . 1.1.2. AD and the brain A D is associated with progressive impairments in memory, learning, judgment and changes in behavior. As the disease progresses, individuals lose the ability to care for themselves, and often become aggressive and exhibit signs of psychoses 8 . In the final stages of AD, individuals are severely demented, often mute, and bedridden. 1 1.1.3. A D neuropathology Behavioral tests are used to differentiate AD-related cognitive decline from other forms of dementia, and allow a diagnosis of possible or probable A D 9 . A formal diagnosis of AD, however, is done post-mortem and requires histological evidence of the two hallmarks of A D neuropathology, which are extracellular amyloid plaques in the neural parenchyma and cerebrovasculature, and intracellular neurofibrillary tangles (NFT) of hyperphosphorylated tau protein 9 . Cerebral atrophy from neuronal death is also observed at autopsy and by MRI in living patients with A D 1 0 . The recent development of an amyloid-binding compound, Pittsburgh Compound-B (PIB), which is visible in positron emission tomography, has facilitated the detection of amyloid deposits in living patients 1 1 . Comparison of PIB with cerebrospinal (CSF) fluid biomarkers suggest that quantifying amyloid in living patients may be effective at detecting AD prior to recognizable behavioral changes 1 2 . 1.1.4. Tau NFTs are intraneuronal deposits of filamentous protein aggregates, composed primarily of abnormally phosphorylated tau protein 1 3 , 1 4 . Tau-protein is usually present in the axon, however, in some forms of dementia, including AD, tau protein becomes hyperphosphorylated forming cytoplasmic aggregates in the neuron 1 5 . NFTs and hyperphosphorylated tau are observed in other forms of neurodegeneration such as Pick's disease. Furthermore, mutations in tau are associated with an autosomal dominant form of frontotemporal dementia, suggesting that tau and NFTs are closely associated with and may be sufficient to cause dementia in the absence of amyloid 1 5 , 1 6 . In AD, the density of NFTs in hippocampus and neocortex correlate well with dementia scores obtained shortly before death 1 7 and an inverse correlation between NFTs and surviving neurites is reported 1 8 . Taken together these studies suggest that NFTs play an important role in AD. 2 1.1.5. Amyloid plaques Amyloid plaques are extracellular protein aggregates deposited around neurites and are associated with activated microglia. The core of the plaques are enriched in fibrillar peptides primarily of 40 or 42 amino acids in length 1 9 and mature plaques contain a variety of aggregated proteins including Apolipoprotein E, Apolipoprotein J , anti-chymotrypsin and lipids 2 0~ 2 2. High concentrations of AB40 and AB42 (which possesses an isoleucine and alanine residue at its C-terminus) have been shown to spontaneously form amyloid fibrils in vitro. In vivo, histological analysis of newly formed plaques reveals that fibrillar AB42 is deposited first and may catalyze the aggregation of other proteins 2 3 . Although amyloid fibrils have been shown to be directly toxic to neurons in vi t ro 2 4 , more recent research suggests that the toxic AS species may be oligomeric AB, formed prior to deposition as amyloid 2 5 . In addition to the direct toxicity of AB and amyloid, the propensity of amyloid to promote microglial activation may also be important in AD. 1.1.6. Amyloid and Inflammation Histological investigation of the amyloid plaque in AD brains reveals that plaques are often infiltrated with activated microglia 2 6 , 2 7 . In the brain, microglia mediate innate immunity in the C N S much like macrophages in the periphery. The consequence of microglial activation may be two-fold. First, activated microglia emit a respiratory burst designed to kill invasive microbes that releases neurotoxic free radicals into the parenchyma and can contribute to toxic oxidative stress upon the neurons 2 8 . Secondly, microglia play an important role in the activation of complement, that promotes the formation of the membrane attack complex (MAC) on invasive cells leading to their lysis and death. Although the MAC is usually prevented from assembling on host cells, both amyloid and microglia secrete factors that have been shown to promote MAC formation and may lead to the accidental targeting of host cells 2 8 . Histological analysis on dystrophic neurons in AD reveals that many of the apoptotic neurons are associated with MACs , suggesting that chronic microglia activation may be responsible for 3 much of the neuronal loss in the AD brain . The importance of inflammation in A D is underscored by retrospective studies that observed a 50% reduction in AD prevalence in arthritic individuals who chronically used NSAIDs 3 0 . NSAIDs, which reduce the production of prostaglandin inflammatory mediators, are believed to be protective either by reducing the damaging effects of cerebral inflammation associated with AD, or by reducing the activity of y-secretase 3 1 . 1.1.7. The amyloid cascade hypothesis and familial A D The amyloid cascade hypothesis of A D is the most widely accepted theory of AD, asserting that an increased abundance of AB in the brain and its eventual deposition as amyloid is the driving force behind other AD neuropathologies including NFT formation and neuron loss 1 ' 3 2 . The AB peptide was first sequenced in 1984 in a paper that reported that the primary component of amyloid in the AD brain were peptides deposited in a B-pleated sheet conformation 1 9 . However, the strongest evidence supporting AB as a critical toxic mediator in AD came from genetic linkage studies of families afflicted with early-onset AD. In contrast to sporadic A D that accounts for ~95% of all reported A D cases, Familial AD (FAD) is associated with autosomal dominant mutations that cause AD between 35 and 60 years of a g e 3 2 . When FAD mutations were first mapped to individual genes in the 1990s it was hypothesized that they would map to a gene containing the AB peptide sequence 3 2 . FAD mutations were found in amyloid precursor protein (APP), a large protein containing the AB peptide domain. However, these early studies found at least 3 FAD associated genes. Two well known FAD mutations in A P P are the Swedish (KM670/671NL) and Indiana (V717F) mutations that both map to the boundaries of the AB peptide region 3 3 , 3 4 . Other FAD mutations mapped to proteins with protease activity, presenilin-1 and presenilin-2 that were shown to cleave A P P at the C-terminal boundary of the AB region 3 5 , 3 6 . The convergent conclusion of these studies was that FAD mutations acted to increase production of AB from A P P , or shift the production from the shorter 40 amino acid (AB40) peptide, to the 4 longer 42 reside form (AS42) • . Together these studies demonstrate that an increase in AS peptide production, specifically AS42, is sufficient to cause AD as early as 35 years of age, and formed the basis for the amyloid cascade hypothesis 3 2 , 3 8 . Although the cause of A S accumulation in FAD cases is well documented, the cause of A S accumulation in sporadic AD remains the subject of debate. 1.1.8. Amyloid Precursor Protein The A S peptide is produced by proteolytic processing of the A P P protein. A P P is a type-l transmembrane glycoprotein existing in three main isoforms of 695, 751 and 770 amino a c i d s 3 9 and is expressed in many cell types in the body with very high expression in neurons. The A P P protein shares sequence homology to well characterized proteins involved in cell signaling 4 0 and in neurons, A P P undergoes anterograde transport. A P P is found in synapses, especially at sites of synaptogenesis 4 1 . It has been shown that A P P forms signaling complexes with Go coupled receptors involved in GTP-signaling pathways 4 2 suggesting that A P P may function as a membrane bound signaling receptor. Additionally, A P P undergoes cleavage to liberate a soluble ectodomain (sAPP) that may act as a growth regulator in addition to a soluble intracellular domain (AICD) that co-localizes with the nucleus and is believed to be a transcriptional activator 4 3 , 4 4 . During the liberation of the two soluble domains from the extracellular and intracellular face of the membrane, the transmembrane spanning domain of A P P remains associated with the membrane. 1.1.9. A P P Processing The processing of full-length A P P into extracellular and intracellular domains depends upon the proteolytic activity of a, S and y secretase (Figure 1-1) 4 5 . The cleavage of A P P first occurs at the plasma membrane by either a or S secre tase 4 6 . a-secretase cleaves A P P within the AS region, to liberate the s A P P a domain and an 83 amino acid (C83) C-terminal fragment 5 Proteolytic Processing of Amyloid Precursor Protein (APP) A P P CYSTEINE R I C H ACIDIC G L Y C O S Y L A T E D B-a-< < < s A P P a sAPPfc a -cleavage ## :3L Cytosol ic C -terminus C83 2a -cleavagej p3 •I 1b G-cleavage C99 2b | y -cleavage AB40 AIS42 AICD I I AICD n • Figure 1-1: P rocess ing of the amyloid precursor protein by secre tases. Full length A P P undergoes cleavage at the plasma membrane by either (1a) a-secretase or (1b) fi-secretase to produce a soluble extracellular domain and membrane bound CTF denoted C83 or C99. (2a) C83 is further cleaved by y-secretase to produce non-pathogenic p3 fragments and the AICD. (2b) C99 is cleaved by y-secretase to produce either AG.40 or AG.42 and the ACID. ## Location of the Swedish KM670/671NL mutation ** Indiana V717F mutation 6 (CTF) that traverses the plasma membrane. Because a-secretase catalyzes proteolysis within the AB region of A P P , a -cleavage precludes the formation of A l i . a -secretase activity is thought to localize to phospholipid-rich domains of the plasma membrane and may be achieved through the actions of multiple proteins shown to have a-secretase activity including T A C E and ADAM10 4 7 . Alternatively, the first step in A P P processing can occur through cleavage by B-secretase-activity-enzyme 1 (BACE-1). The BACE-1 protein is an aspartyl-protease found in cholesterol-rich microdomains of the plasma membrane that cleaves A P P to liberate the s A P P B ectodomain and a 99 amino acid (C99) CTF. Mice deficient for BACE-1 do not generate AB, confirming that BACE-1 is essential to B-secretase cleavage 4 8 , 4 9 . Following cleavage at the plasma membrane, C83 and C99 CTFs are internalized where they are degraded or further cleaved by y-secretase to liberate the AICD. y-secretase is a complex composed of presenilin-1, presenilin-2, nicastrin and APH-1 and cleaves C83 substrates within the phospholipid bilayer to release the soluble AICD from the transmembrane-spanning p3 peptide, or alternatively y-secretase cleaves C99 to liberate the AICD and leave the AB peptide of either 40 or 42 amino acids associated with the membrane 4 5 . 1.2. Cholesterol 1.2.1. Cholesterol in mammalian cells Cholesterol is a 27 carbon molecule composed of 4 adjoining carbon rings, hydrocarbon tail and hydroxyl group. Cholesterol plays a key role in maintaining the proper fluidity of eukaryotic cell membranes and as a precursor for many hormones. The plasma membrane is enriched in cholesterol, possessing more than 80% of unesterified cholesterol in the cell and comprising 20-25% of the total lipid content of the plasma membrane 5 0 . In health, cells balance their membrane cholesterol content by regulating its uptake, synthesis, storage and export, as excess membrane cholesterol can be cytotoxic. In the brain, neurons and glia import cholesterol and other lipids from lipoprotein-like particles containing apoE through receptors of the LDL-receptor family 5 1 . Additionally, cells can synthesize their own cholesterol 7 in the endoplasmic-reticulum (ER). The committed and rate limiting step in cholesterol biosynthesis is the enzymatic combination of 2 acetyl-CoA molecules to form mevalonate by 3-hydroxy-3-methyl-glutaryl-CoA reductase (HMGR). Under conditions of low cellular cholesterol or mevalonate the transcription of HMGR is increased, whereas under conditions of high cholesterol the degradation of H M G R is increased. These properties allow the abundance of the H M G R protein to regulate cholesterol synthesis 5 2 , 5 3 . Under conditions of excess cholesterol, plasma membrane cholesterol is either effluxed to extracellular acceptors or is returned to the ER where it can be esterified with fatty acids by AcetyCoA Transferase for storage in the hydrophobic core of cytoplasmic lipid droplets 5 4 . Cells require cholesterol efflux to eliminate excess cholesterol as mammalian cells cannot degrade the sterol ring. To remove cholesterol from the body, cholesterol or sterol-derivatives such as oxysterols are returned to the liver for enzymatic conversion into bile-salts and bile-acids, which are ultimately eliminated in bile. 1.2.2. Cholesterol and the brain: The blood brain barrier The brain accounts for approximately 2% of total body mass yet it contains 23% of the sterols in the entire body 5 0 . It is estimated that 70-80% of the cholesterol in the brain surrounds neurons in myelin sheaths that maintain conductivity of the neuron, with the remainder present in cell membranes or lipoprotein-like particles in the C S F 5 5 . In the periphery, cells obtain cholesterol from endogenous synthesis, or from circulating lipoproteins enriched in dietary cholesterol. However, the brain does not appear to import cholesterol from the peripheral circulation. This was best shown by an experiment that monitored uptake of isotopically labeled cholesterol given intravenously to baboons over 85 days. After 70-85 days all organs contained large quantities of labeled cholesterol, with the exception of the brain that acquired little or no cholesterol 5 6 . Furthermore, experiments in mice fed a high-cholesterol diet found that while serum lipids were increased 7-8-fold above chow-fed mice, no evidence of cellular accumulation of cholesterol, or reduced synthesis was observed in the brain, in contrast 8 to peripheral organs 5 7 . The insulation of peripheral and brain cholesterol occurs at the Blood brain barrier (BBB), which consists of two layers of cells that surround capillaries of the brain. The first layer is a specialized subset of endothelial cells that form tight junctions with neighboring cells that prevent bulk-flow around the cells and show minimal rates of endocytosis and vesicular transport across their membranes 5 8 . The second line of defense is a sheath of astrocytic cell processes that surround the endothelial cells and, although they are essential for the impermeability of the membrane, their function is less well understood. This BBB, although impermeable to the influx of cholesterol, contains transporters to permit the flow of certain compounds such as glucose and is permeable to some hydrophobic molecules, including hydroxysterols that diffuse down concentration gradients across the BBB 5 0 . Because the brain is impermeable to cholesterol influx, all cholesterol must be endogenously synthesized. 1.2.3. Cholesterol and the brain: metabolism In the adult brain many, if not all cells, have the capability to synthesize cholesterol, and to do so continuously (Figure 1-2) 5 0. However, terminally differentiated adult neurons often do not synthesize sufficient cholesterol for processes such as axonal growth and synaptogenesis, and rely largely upon cholesterol obtained from glia. Because cholesterol cannot readily traverse the BBB into the periphery, cholesterol is eliminated from the brain mostly via conversion to the oxysterol 24S-hydroxycholesterol. 24S-hydroxycholesterol is not restricted by the BBB and passively diffuses down a concentration gradient into the circulation where it can be transported to the liver for elimination as bile acids 5 9 . 24S-hydroxycholesterol is produced enzymatically by Cyp46A1, which, in humans, is expressed exclusively in sub-sets of neurons 6 0 . The mechanism by which glial cells clear cholesterol is less well understood and may be coupled to cholesterol efflux to glial-derived apolipoproteins, as glia do not express Cyp46A1 5 0 ' 6 0 . Genetic polymorphisms in Cyp46A1 are associated with increased risk of AD in some s tud ies 6 1 . However a meta-analysis of Cyp46A1 polymorphisms finds that only 11 of the 22 studies found a significant association between Cyp46A1 and A D and concluded that Cyp46A1 was not a significant contributor to AD in the general popu lace 6 2 . 9 Summary of the Cholesterol Biosynthetic Pathway HMGR 2 X A c e t v l C o A — • Mevalonate T Lanosterol Figure 1-2: Summary of the cholesterol biosynthetic pathway. Two molecules of AcetylCoA are converted to mevalonate by HMGR, in the first and rate limiting step of cholesterol biosynthesis. The pathway branches at lanosterol, through pathways involving desmosterol or lathosterol that are both precursors to cholesterol. In the brain, cholesterol is converted by Cyp46A1 to 24S-hydroxycholesterol that diffuses across the BBB for elimination into bile. Dotted lines represent multistep reactions, solid lines represent single step reactions and catalytic enzyme. Extracellular cholesterol transport in the C N S , involves lipoprotein-particles of similar size and density to HDL, which are enriched in phospholipids and cholesterol and complexed with apoE and to a lesser degree apoJ 6 3 . CNS-lipoproteins are thought to be primarily formed by glia that synthesize the majority of the apolipoprotein in the brain 6 3 . In culture, the cholesterol delivered by apoE containing lipoproteins have been shown to dramatically increase the elongation of axons required in times of growth or neuronal repair and are required for efficient formation and maintenance of synapses 6 4 , 6 5 . These experiments suggest a central role of glia in brain lipid homeostasis, which is central to the growth, maintenance, and repair of neuronal membranes. 1.2.4. Cholesterol and Alzheimer's Disease Several converging lines of evidence suggest that the transport and levels of cholesterol in the brain may play an important role in the pathology of AD. Specifically, apoE genotype, serum cholesterol levels, and the therapeutic use of cholesterol biosynthesis inhibitors known collectively as statins, are implicated in the pathogenesis of AD. 1.2.4.1. Apolipoprotein E The strongest genetic risk-factor for sporadic A D is allelic variations in the apolipoprotein E (apoE) gene, the main cholesterol carrying apolipoprotein in the brain. The human apoE locus is present in three main alleles, of which apoE3 is present in 70-78% of the population and is the most common allele, apoE4 is present in 20-30% of the population and apoE2 is the rarest allele, being present in only 7-8% of individuals 6 6" 6 8 . It is well established that the apoE4 allele acts to decrease the average age of onset of A D relative to the apoE3 allele, and that a single allele of apoE2 exerts a protective effect. For example, apoE3/E2 carriers develop A D latest at an average of 84 years compared to an average of 68 years age in apoE4/E4 carriers. Furthermore, the prevalence of AD in non-apoE4 carriers at 75 years of age is 20% compared to 60% in the population carrying one-apoE4 allele, or 80% in apoE4/E4 11 carriers. These studies support the hypothesis that apoE4 promotes A D in a dose dependant manner. The mechanisms by which substitutions in one or two amino acids among apoE2 (Cys112, Cys158), apoE3 (Cys112, Arg158) and apoE4(Arg112, Arg158) has been the focus of much research and has not reached consensus. It has been shown that apoE4 delivers less cholesterol to cultured neurons than apoE3, with apoE2 delivering the mos t 6 9 . Neurite growth, particularly in the adult brain, is heavily dependant on apoE-derived cholesterol, and apoE4 may be less efficient at repairing damaged neurons in AD 6 5 , 7 ° . An alternative line of evidence suggests that AB:apoE interactions may account for the impact of apoE genotype. ApoE3 but not apoE4 has been shown to form SDS-resistant complexes with A l i 7 1 , 7 2 that may impair the clearance of AIJ42 from the brain of apoE4 carriers via apoE receptors 7 3 . Furthermore, apoE deficient mice do not form fibrillar amyloid, suggesting that apoE may catalyze the conversion of soluble A l l into amyloid fibrils in vivo in an isoform specific manner 7 4 . Although the mechanism underlying the apoE and AD interaction remains the subject of considerable debate, apoE remains the only well established genetic risk-factor for sporadic AD in the general population 6 2 . Because sporadic A D accounts for more than 95% of A D cases, understanding how apoE contributes to A D pathogenesis may be of higher impact to worldwide disease burden than any treatments based on the pathways illuminated by FAD mutations that affect less than 5% of A D patients. 1.2.4.2. Serum Cholesterol The second line of evidence supporting a role for cholesterol in AD are population studies that suggest elevated serum cholesterol is a risk factor for AD. The strongest evidence for this connection comes from studies with long follow-up periods. One study of 1449 individuals that obtained 4 measures of serum cholesterol from individuals over an average of 21 years found that cholesterol levels greater than 6.5mM in midlife were associated with 2.6 greater chance of being diagnosed with A D 7 5 . This study confirmed that apoE isoform was a 12 genetic risk factor, but that the effect of serum cholesterol was independent of apoE. Similarly, the Finnish Male cohort study of the seven countries and the Honolulu aging study found that total cholesterol serum levels >6.5mM in midlife were associated with 3.1 greater risk of AD during a 15-25 year follow up period 7 8 " 7 8 . However, the Framingham study and Gotborg study, both with 10+ years of follow up concluded that there was either no association between total serum cholesterol and AD, or that higher cholesterol levels were associated with decreased dementia risk 7 9 , 8 0 . One criticism with the Framingham study however, with respect to the connection to AD, was the use of broadly stratified age groups that may not discriminate midlife measurements from older individuals. Numerous other studies have been performed, many with only 2-3 year follow up periods and, taken together, the number of studies reporting negative associations between markers of serum lipids and A D outnumber the positive association studies, reviewed 8 1 . Despite the contradictory results, the evidence suggests that cholesterol levels in midlife may be a better indicator of A D risk than cholesterol levels immediately prior to or upon diagnosis. Studies using animal models have attempted to recapitulate the human studies with the use of high-cholesterol diets. Rabbits fed a diet high in cholesterol develop intraneuronal AB deposits in as few as 4 weeks of feeding, compared to chow-fed rabbits that have undetectable levels of A f i 8 2 . Furthermore, when fed a diet high in cholesterol for 7 weeks the P S A P P mouse model of AD, which is transgenic for human A P P and human mutant presnilin-1, was found to have larger and more numerous amyloid deposits than chow fed mice. The authors also found a positive correlation between total plasma cholesterol and insoluble AG. load suggesting that a high dietary intake of cholesterol could influence A P P processing and exacerbate amyloid deposition 8 3 . Although not all animal studies agree that hypercholesterolemia increases amyloid deposition, all studies have found evidence of altered A P P processing in mice fed a high fat/cholesterol diet 8 4 , 8 5 . Taken together, these studies show that a high intake of dietary cholesterol can influence A P P processing in the brain, supporting the hypothesis that high serum cholesterol may promote AD pathology in humans. The underlying mechanism behind 13 the correlation of high-serum cholesterol and AD is unclear however, as cholesterol in the central nervous system is believed to be insulated from peripheral sources 5 0 1.2.4.3. Statins The third line of evidence connecting cholesterol metabolism and AD is derived from retrospective studies finding that statin use was associated with a reduced prevalence of AD. Statins inhibit the activity of HMGR, which catalyzes the rate limiting step in cholesterol biosynthesis, and are prescribed to reduce LDL-cholesterol in patients at risk for cardiovascular disease. In the initial retrospective studies, the use of lovastatin or pravastatin was shown to reduce the prevalence of probable AD by 60-73% when compared to either no-treatment or patients taking other medications for hypertension or cardiovascular disease 8 6 . Despite the marked success of these early retrospective trials, the results of prospective trials have not reached a clear consensus. Two trials designed to validate the efficacy of simvastatin and pravastatin in hypercholesterolemic individuals 80 years of age or older, did not detect a difference in the incidence of A D between statin treated and non-treated individuals over a 5 year follow-up period 8 7 , 8 8 . However, both of these prospective studies have received criticism for their use of phone interviews to assess AD during the follow up period, and may also have lacked the sensitivity to detect small changes in AD incidence ra tes 6 2 . In contrast, a 2006 prospective trial conducted on individuals with early to mid stages of A D did find that atorvastatin improved cognition scores compared to placebo after 6 months and by a greater degree after 1 y e a r 8 9 . A slight beneficial effect of simvastatin on cognition was also confirmed in another prospective 26 week study of individuals with mild-moderate AD in the absence of preexisting hypercholesterolemia 9 0 . The discrepancies in statin studies may be related to the outcome measures used by individual studies. The prospective studies examining statin use in non-demented elderly failed to observe a significant impact on AD incidence, however, statin therapy was shown to improve cognitive markers in patients with mild AD. These studies suggest that while the initial benefit 14 statin therapy observed in retrospective trials may have been overestimated, statins may prove to be effective in treatment if not prevention of mild AD. The mechanism by which statins may impact AD pathology is also the subject of debate. One line of evidence suggests that statins may reduce the production of AB in the brain. It was shown that hippocampal rat neurons exposed to statins secreted 60% fewer AB40 peptides in cul ture 9 1 . In the same study, directly depleting membrane cholesterol using |3-methylcyclodextrin (PMCD) produced a similar reduction in AB. Fassbender and colleagues went on to confirm their results in vivo, by demonstrating that AB40 and 42 levels in the brain of statin-fed guinea pigs was reduced by 40-60% 9 1 . Similarly, HEK 293 cells treated with lovastatin, or (3MCD, produced fewer AB peptides and had increased a-secretase activity 9 2 . Both these studies, however, used very high statin doses, ~220-400x greater in guinea pigs than the average weight adjusted human dose. These experiments are interpreted to suggest that neuronal cholesterol levels modify A P P processing, possibly by trafficking A P P into B A C E -1 containing cholesterol-rich lipid raft domains under conditions of abundant cholesterol, or by trafficking A P P into phospholipid rich a-secretase containing membrane domains under conditions of low cholesterol 4 7 . Accordingly, statins influence neuronal cholesterol levels and thus influence A P P processing and AB production. Despite the caveats of the animal studies, human trials suggest that statin therapy of therapeutic doses may have a similarly beneficial effect. Patients treated with simvastatin were shown to have reduced s A P P a and s A P P B in C S F , suggesting that statin therapy reduces the processing of A P P 9 3 . Furthermore, individuals treated for 3 months with atorvastatin of various therapeutically approved doses were found to have a dose dependant decrease in plasma AB 9 4 . Although plasma AB is not a reliable biomarker for A D and the significance of these results to AD is unclear, the studies provide evidence that statin treatment may reduce the production of AB in humans 9 4 . The role of statins in AD however, may be more complex than cholesterol-mediated effects on AB production. The H M G R enzyme produces the precursors for both cholesterol 15 and a class of molecules known as isoprenoids, and statin therapy inhibits the synthesis of both cholesterol and isoprenoids 9 5 . Isoprenoids are involved in signaling, especially in mediating inflammatory reactions. Studies on rat brain-slice cultures show that statins strongly reduce microglial activation, and are especially effective at reducing AU-induced microglia activation 9 5 , 9 6 . The anti-inflammatory effects of statins are also believed to be important for their protective role in atherosclerosis, suggesting that any beneficial effects on A D may be pleiotropic, resulting from both reduced synthesis of cholesterol and inhibition of isoprenoid-mediated inflammatory responses 9 7 . 1.3. Down Syndrome and AD The A P P gene is located on chromosome 21 and triplication of chromosome 21 in Down Syndrome (DS) inevitably results in the rapid accumulation of AD-like neuropathology by the age of 40 y e a r s 9 8 . Interestingly, the triplication of A P P alone through small chromosomal rearrangements is also associated with early-onset AD. A recent phylogenetic analysis of five separate families with duplicated A P P found that inheritance of three copies of A P P was sufficient to cause AD-like dementia between 42 to 59 years of age in 21 out of 21 individuals studied, in the absence of DS. The presence of AD neuropathology was confirmed at autopsy 9 9 . In comparison, severe AD-like neuropathology is observed in almost all DS cases at autopsy by 40 years of age 1 0 ° , reviewed in 1 0 1 . The specific age of onset for A D dementia in DS is difficult to assess because of the underlying and highly variable degree of mental retardation in DS. Because of this, the onset of AD-related dementia in DS cannot be directly compared to A P P triplication alone. However, a comparison of the reported onset of pathology between the two studies suggests that A D neuropathology develops more quickly in the DS population. Almost 100% of DS individuals have pathology by 40 years of age versus A P P triplication who develop dementia at an average of 52 years and as late as 59. This comparison suggests that other genes on chromosome 21 may play a role in A D when overexpressed in DS subjects. 16 1.4. A B C Transporters 1.4.1. The ATP-Binding-Cassette superfamily The ATP-Binding-Cassette (ABC) superfamily of transporters contains 48 genes sub-divided into 8 sub-classes classified from A B C A to A B C G . Typical A B C transporters consist of two sets of 6 transmembrane a-helix domains, dual A T P binding-domains and a characteristic signature motif specific to A B C transporters, reviewed in 1 0 2 . Eukaryotic A B C transporters are either full-size, containing all 12 a-helices on one polypeptide chain, or half-size requiring homodimerization, or heterodimerization with a similar A B C transporter to form a functional unit 1 0 3 . The prototypical function of an A B C transporter is to promote the uni-directional movement of compounds or peptides across lipid membranes using energy provided by the hydrolysis of A T P . A B C transporters typically transport their substrate from cytoplasm to extracellular space, or in the case of lipid transporters, from the inner leaflet of the plasma membrane to the outer leaflet or to an extracellular lipid acceptor 1 0 2 . Specific examples of A B C transporters include ABCB1 involved in multidrug resitance 1 t M , the nucleoside transporter A B C C 4 , and ABCA1 that transports cholesterol and phospholipids to Apolipopoprotein Al to form high density lipoprotein ( H D L ) 1 0 5 . 1.4.2. The A B C G 1 transporter A B C G 1 is a half-size transporter with a proposed role in sterol and lipid transport. Human Abcgl is encoded on chromosome 21 q.22.3, contains 23 exons and is 97% identical to murine ABCG1 at the protein l e v e l 1 0 6 . In humans Abcgl is expressed in most tissues, and is especially abundant in liver, spleen, thymus and brain 1 0 7 . Primary sequence topology prediction reveals a typical A B C transporter design, with alternative splicing occurring in the N-terminal intracellular loop, and the constant exons (exons 11-23) encoding the canonical 6 17 ABCG1: Predicted Topology Figure 1-3: Predicted topology of the A B C G 1 monomer Each ABCG1 monomer is predicted to encode 6 transmembrane alpha-helices and the Walker A and Walker B ATP binding domains that comprise the constant domain. The N-terminus of the protein is subject to alternative splicing and more than 11 isoforms have been detected in macrophages. transmembrane helices and the Walker A and Walker B A T P binding domains (Figure 1-3) 1 0 8 . As many as 11 alternatively spiced isoforms have been reported in macrophages, although only 4 of these have been validated 1 0 6 , 1 0 9 . All identified transcripts possess one or more of the first 10 exons, and all of exons 11-23. The transcript corresponding to exons 5,7 and 11-23 codes for a protein of 678 amino acids in length and shows the greatest induction to oxysterol treatment both in vitro and vivo 1 0 6 . Overexpression of 4 different Abcg 1 isoforms has been shown to have similar effects on sterol transport, however, the significance of the different isoforms, or the tissue specific expression patterns of different isoforms have not been well characterized. 18 1.4.3. Transcriptional regulation of Abcgl Several A B C transporters involved in the regulation of sterol homeostasis are transcriptionally regulated by agonists of the Liver-X-Receptor (LXR) nuclear receptors 1 1 0 The promoter region of the Abcgl gene in both mice and humans possesses multiple LXR-Response-elements (LXREs) upstream of exon one and upstream of exon 8 1 0 6 , 1 0 9 . L X R E -mediated oxysterol-sensitive transcription is achieved through the binding of LXR and R X R transcription factor dimers, which bind oxysterol and retinoid ligands, respectively, in the cytosol, and translocate to the nucleus where they bind L X R E s and promote transcription Because oxysterol ligands are produced as cholesterol byproducts in cholesterol laden cells, L X R E responsive genes such as Abcgl will be expressed under conditions of excess cholesterol suggestive of a role in cholesterol metabolism Functional L X R E sensitivity of the Abcgl gene was confirmed by mRNA overexpression following treatment with 20S-hydroxycholesterol, 22R-hydroxycholesterol and the synthetic LXR-ligand TO901317, and ablation of expression in cells deficient for LXR transcription factors The Abcgl gene is also regulated by the peroxisome proliferator-activated-receptor-gamma (PPAR-y) transcription factor in vivo 1 1 2 . Abcgl expression was induced in macrophages and aorta following treatment with the PPAR -y ligand rosiglitizone, and this activation was shown to be independent of LXR transcription factors 1 1 2 . PPAR -y activation suppresses the inflammatory response and is reported in activated macrophages and microglia 1 1 3 . 1.4.4. Transporter assembly Because the A B C G transporters are half-transporters and A B C G 5 and A B C G 8 , require heterodimerization to function, the assembly of functional A B C G 1 transporters is an important consideration for overexpression studies. Human Abcgl is considered a functional homologue of the Drosophila melanogaster white gene, which transports amino acids for pigment synthesis in the retina. Studies in D.melanogaster suggest that white may require heterodimerization with scarlet, another A B C protein, to form a functional transporter 1 1 \ Accordingly, mammalian 19 A B C G 1 was suspected to form heterodimers with a highly homologous transporter known as A B C G 4 , which was supported by one report suggesting that co-expression of ABCG1 and A B C G 4 was required to promote A T P hydrolysis using an insect cell system 1 1 5 . However, selective overexpression of A B C G 1 alone in various mammalian cell types produced functional transporter activity, which was not augmented by co-expression with A B C G 4 1 1 6 suggesting that A B C G 1 can form functional homodimers. 1.4.5. A B C G 1 function The established function of ABCG1 is to facilitate the efflux of cellular cholesterol to lipidated apolipoproteins such as HDL. Although the specific mechanism by which ABCG1 facilitates the transport of lipids remains elusive, the role of ABCG1 in lipid efflux has been studied in more than 30 peer reviewed articles between 2005 and 2007. The majority of these reports focus on sterol transport in the periphery, because of the importance of sterol-efflux to atherosclerosis, in which the export of cholesterol from lipid laden macrophages is thought to be the rate limiting step in preventing lesions in the vessel wall. In brief, the overexpression of Abcgl cDNAs has been shown in various cell types to promote cholesterol efflux to HDL, but not to lipid-poor apolipoproteins, and that this activity is likely mediated by enriching the outer leaflet of the plasma membrane with f ree -cho les te ro l 1 0 6 , 1 1 6 , 1 1 7 . A B C G 1 deficiency in cells impairs efflux of cholesterol and choline-based phospholipids to HDL and, furthermore, ABCG1 deficient mice have been reported to accumulate more cholesterol and neutral lipids in a variety of tissues when fed a high fat diet compared to wildtype animals 1 0 7 , 1 1 8 . In summary, ABCG1 in vivo appears to play an important role in maintaining tissue lipid homeostasis, likely though facilitating lipid export. In contrast to ABCA1 that specifically transports lipids to poorly lipidated or lipid-free apolipoproteins, ABCG1 facilitates efflux to various lipid containing particles including discoidal apoAl, LDL, HDL and synthetic phospholipids vesicles, suggesting that the function of ABCG1 is distinct from A B C A 1 1 1 9 ' 1 2 0 . Because ABCG1 was shown to regulate sterol transport from cells, multiple groups undertook studies in mice exploring the role of ABCG1 in atherosclerotic lesion development. Despite predictions from the in vitro studies that ABCG1 overexpression would prevent atherosclerotic lesions and conversely that ABCG1 deficiency would increase lesions, the results thus far have been contradictory 1 2 1 1 2 5 . While some studies show the expected inverse correlation between ABCG1 protein levels and atherosclerotic lesions, other studies report fewer lesions in Abcgl knockout mice or an increased accumulation of tissue lipid in ABCG1 overexpressing mice on a high fat diet (Burgess B L , et al. manuscript submitted to J . Lipid Res). Taken together these studies suggest that while A B C G 1 is undoubtedly capable of facilitating the export of lipids from cultured cells, the in vivo role of A B C G 1 is not entirely understood. 1.4.6. Human disease and Abcgl There are currently no known diseases with well established genetic linkage to Abcgl. However, because of its position on chromosome 21, Abcgl is expected to be overexpressed by approximately 1.4-2-fold in with the inheritance of a third copy of chromosome 21 in DS 1 2 6 , 1 2 7 . Interestingly, individuals with DS do show some evidence of dysregulated lipid metabolism, although it should be noted that many of these studies have used very small cohorts with questionable controls. First, trisomic fetuses are reported to be hypercholesterolemic in utero 1 2 8 . Secondly, young children with DS are reported to be 23 times more likely to develop cholelithiasis 1 2 9 , which are biliary gall-stones associated with cholesterol-saturated bile and high serum cholesterol levels 1 3 0 . Third, several groups have measured serum lipids in adults with DS and report that trisomic adults have atherogenic serum lipids, yet paradoxically, appear to be protected from atherosclerosis when compared to similarly institutionalized mentally-retarded controls 1 3 1 , 1 3 2 . Although not conclusive, these results suggest that DS may be associated with aberrations in cholesterol homeostasis and that genes on chromosome 21 such as ABCG1 may contribute to these findings. 21 1.5. AD mouse models 1.5.1. Modeling AD: Amyloid vs Tau Over 100 separate animal murine models for studying A D have been reported. These animal models group loosely into two classes, first, tau based models that overexpress tau, or express mutant tau in an attempt to induce A D through tau and tangle pathologies, and second, APP-related models that attempt to induce AD-like pathologies through the formation of Af i or amyloid deposits. While tau and A P P techniques have been successful at producing NFTs and amyloid plaques respectively, neither technique has recapitulated both aspects of pathology without co-expressing mutant forms of tau and A P P in the same m o u s e 1 3 3 , 1 3 4 . A major obstacle to creating a complete model of AD may result from the use of mice as a model species. For instance, humans expressing A P P at 1.5-fold normal levels develop autosomal dominant A D that is 100% penetrant by midlife, where as mice need to overexpress A P P more than 8-fold in order to develop amyloid pathology 1 3 4 . While no mouse model of A D exists that fully recapitulates A D neuropathology, APP-related amyloid models of AD provide the best in vivo available system for testing treatments based on the amyloid hypothesis and recapitulate amyloid plaques, cerebral inflammation and cognitive deficits consistent with AD. 1.5.2. The P D A P P Mouse model of A D The P D A P P mouse model of AD is a well-established model of amyloid formation. P D A P P mice express an A P P minigene containing introns 7,8 that permits alternative splicing to produce all the three major human A P P isoforms, A P P 695, 751 and 770 1 3 5 . In the P D A P P mouse all isoforms contain the FAD associated A P P V717F "Indiana" mutation that increases the production of AIJ42 by y-secretase. The A P P minigene is expressed under the PDGF-B promoter that drives high expression in hippocampal neurons and cortex 1 3 6 By 12 months of age, P D A P P mice develop amyloid plaques associated with dystrophic neurites and G F A P -positive glial cells that are similar to human pathology, and developed age-related deficits in 22 learning and memory 1 3 7 . Unlike human AD, however, no significant neuronal loss has been observed in P D A P P mice 1 3 6 . P D A P P mice expressing humanized apoE3 or apoE4 alleles (ie replacement of the murine apoE gene with either human apoE3 or apoE4) confirm that apoE4 accelerates amyloid formation, which is consistent with human studies 1 3 8 . The P D A P P mouse is well suited for these experiments as it is well-established model for studies of cholesterol metabolism and A D and recapitulates multiple aspects of A D including learning and memory deficits, parenchymal amyloid, intracellular amyloid and cerebral inflammation. 1.6. Research Hypothesis and Rationale 1.6.1. Rationale The amyloid cascade hypothesis of A D asserts that A P P processing into AU peptides and the accumulation of soluble and fibrillar AB in the brain, either by increased production or decreased elimination, are the main drivers of A D neuropathology 3 2 . It has been shown that cholesterol levels of the plasma membrane play an important role in AB production by modulating the proteolytic processing of A P P by secretases 9 1 , 9 2 4 7 . Thus, because ABCG1 is an important regulator of cholesterol levels and distribution within cells and tissues 1 0 7 , 1 1 6 , 1 1 7 1 1 8 , A B C G 1 may influence the production of AB. Furthermore, individuals with DS overexpress Abcgl and A P P on chromosome 21 1 2 6 , 1 2 7 , 1 3 9 and may accumulate amyloid pathology more rapidly than individuals that overexpress A P P a l o n e 9 9 , 1 0 ° . Thus, the main goal of this thesis is to evaluate the role of ABCG1 overexpression on AB production and amyloid formation. Furthermore, because little is known about the role of ABCG1 in the brain in vivo, this thesis has a secondary focus on the role of ABCG1 in mediating cholesterol homeostasis in the brain. 1.6.2. Experimental Hypothesis and Specific Aims The hypothesis of this thesis is that ABCG1 influences AB production and amyloid formation. Specifically, we hypothesize that ABCG1 overexpression will promote amyloid formation by facilitating the production of AB peptides in a cell model and from neurons in vivo. 23 Specific Aims: 1. To determine if ABCG1 over-expression influences AU production in vitro 2. To characterize Abcgl expression patterns in a novel murine model of Abcgl overexpression 3. To determine the impact of ABCG1 overexpression on brain lipid metabolism A B C G 1 transgenic mice. 4. To establish if ABCG1 overexpression or deficiency influences amyloid neuropathology in a mouse model of amyloid formation. 2 . CHAPTER 2 : THE ROLE OF A B C G 1 IN BRAIN AND AD 2 . 1 . EXPERIMENTAL REAGENTS AND PROCEDURES 2.1.1. Transgenic Animals 2.1.1.1. Generation of ABCG1 BAC Transgenic Mice ABCG1 bacterial artificial chromosome (BAC) transgenic mice were generated using the 141 kb BAC CTD-201013 (Children's Hospital Oakland Research Institute). The BAC insert contained the entire human Abcgl genomic sequence plus 30.6 kb 5' and 13.1 kb of 3' flanking sequences. The BAC was purified through pulse-field gradient gels and the presence of all 23 exons was confirmed by P C R . The validated BAC was microinjected into F1 C57BI/6/CBA murine oocytes, and founder animals were identified by human specific P C R . The stable integration of all 23 exons was verified. Founders were backcrossed onto a pure C57BI/6 genetic background for at least 5 generations before being used for experiments. 2.1.1.2. P D A P P Mice P D A P P mice were generously donated by Dr. Ronald DeMattos, Lilly Research Laboratories, Indianapolis, Indiana USA. P D A P P mice express a human A P P mini-gene under the control of the PDGF-ft promoter that promotes expression of the transgene in hippocampal and cortical neurons 1 4 0 . The A P P mini-gene encodes the FAD associated Indiana mutation (V717F) in addition to A P P introns 6-8 to produce A P P isoforms 695, 751 and 770 through alternative splicing. P D A P P mice are maintained by brother-sister matings as transgenic homozygotes on a mixed genetic background derived from C57/BI6, DBA and Swiss-Webster origins. 2.1.1.3. Aocgrf-deficient mice: Abcg 1 -deficient mice were obtained from Deltagen Inc. The targeting vector used to generate these mice contained 7 kb of 5' and 1.4 kb of 3' murine genomic DNA flanking a 7 kb 25 Internal Ribosome Entry Site -LacZ-Neo-pA cassette that places the B-galactosidase gene under the control of endogenous Abcgl regulatory elements. Homologous recombination results in the deletion of 7 amino acids ( G P S G A G K ) within the Walker A motif in exon 3 of the murine Abcgl gene to ablate function of the transporter. Chimeric animals were generated using embryonic stem cells derived from the 129/OlaHsd genetic background, and were backcrossed to C57/BI6 mice for at least 7 generations before use. 2.1.1.4. Generation of Experimental Cohorts Heterozygous ABCG1 BAC transgenic males were crossed to P D A P P A P P homozygous females to produce (ABCG1 BAC+ / APP+) and (ABCG1 wt, APP- ) progeny. Female progeny were sacrificed at wean to generate male cohorts for aging experiments. 2.1.1.5. Diet Animals were maintained on a standard chow diet (PMI LabDiet 5010, containing 24% protein, 5.1% fat, and 0.03% cholesterol) 2.1.1.6. Sacrifice and Tissue Preparation Animals were anesthetized with 1-1.5mL Avertin™ administered intraperitonealy and prepared according to manufacturer's instructions. Anesthetized mice were perfused transcardially with P B S containing 2500U/mL heparin at 5-10 mL/min for 7min. Whole tissues were taken or brains were dissected into regions on ice before snap-freezing on dry ice for storage at -80°C or immersion fixation in 10% formalin in P B S and stored at 4°C. All procedures involving experimental animals were performed in accordance with protocols from the Canadian Council of Animal Care and the University of British Columbia Committee on Animal Care. See Appendix 1. 26 2.1.2. Cell Culture 2.1.2.1. HEK 293 Cells In vitro assays were performed with HEK 293 cells stably transfected with human APP695 Swe (KM670/671NL) that were generously donated by Dr. Weihong Song (University of British Columbia, Vancouver, Canada). HEK 293 cells were chosen for these experiments because they are yield high transfection efficiencies and exhibit a,p and y secretase activities required to cleave A P P into AB peptides. HEK 293 cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) containing 10% heat inactivated fetal bovine serum (FBS), 4.5g/L glucose, 2mM L-glutamine, Penicillin (100u/ml_) and Streptomycin (100mg/mL) and 200ug/mL G418 (All from Invitrogen) to maintain expression of the A P P transgene. Transfection: HEK:APP695Swe cells were seeded in 10cm plates at 50% confluency 24h prior to transfection. For transfection, cells were incubated with 25uL F u G E N E transfection reagent (Roche) and either 8ug empty vector pFRT (Invitrogen) or pFRT vector containing an Abcgl cDNA (IMAGE Consortium) containing exons 5,7,11-23, as described 1 0 6 . Cells were transfected for 24h, then media was changed to regular growth media for 6h before use. Transfection efficiency was estimated by adding 1ug of e G F P cDNA to the transfection mixture and visualizing G F P fluorescence by inverted microscopy. Transfection efficiency was estimated to be 30-50%. 2.1.2.2. Primary Astrocyte Culture Primary astrocytes were obtained from 1-2 day post-natal pups. Pups were sacrificed by cervical dissection and brains were removed and placed in cold Hanks Balanced Salt Solution (HBSS) adjusted to 6g/L glucose and 10mM H E P E S buffer. Using a dissecting microscope, cortices and hippocampi were isolated, meninges were removed, and brains were triturated by shearing through a 5mL sterile tissue culture pipette. Triturated brain samples 27 were spun at 1000RCF for 5min, supernatant was discarded and homogenized tissue samples were resuspended in 1 mL of Growth Media (DMEM:F12 containing 10% F B S , 2mM L-Glutamine, Penicillin (100u/mL) and Streptomycin (100mg/mL) all reagents from Invitrogen)). The cell suspensions each from individual pup were added to one T75 tissue culture flask containing 19mL Growth Media for maintenance. Media was changed every 3-5 days until cells were fully confluent, which occurred after approximately 12-14 days in vitro. 2.1.2.3. Primary Neuron Culture Primary cortical neurons were prepared from E16 or E17 murine embryos. To harvest neurons, embryonic brains were removed and placed in cold H B S S containing 6g/L Glucose, 50mM H E P E S . Using a dissecting microscope, cortices were isolated, stripped of meninges and minced using forceps. To free neurons from the extracellular matrix, minced cortices were incubated 3-10min in 0.05% Trypsin at 37°C. Trypsinization was stopped with addition of 2 volumes of Stop Media (Neurobasal media, 10% F B S , 0.5mM L-Glutamine, 1xB27 Supplement, Penicillin (100u/mL) and Streptomycin (100mg/ml, all reagents obtained from Invitrogen) then cells were triturated gently through a 5mL sterile tissue culture pipette and spun 3 min at 1000RFC to collect cells. Crude tissue pellets were resuspended Neuron Growth Media (NGM): (Neurobasal Media, 0.5mM L-Glutamine, 1xB27 Supplement, 1% Penicillin/Streptomycin (Invitrogen)) and were gently triturated through fire-polished glass pipettes to generate a single cell suspension. Crude cellular debris was allowed to settle for 1min before transferring the cell-containing supernatant transferred to a new tube. Neurons were seeded at a density of one mouse per 24 well plate that had previously been coated over-night with 0.1mg/mL Poly-D-Lysine to facilitate cell adhesion. Neurons were maintained in 800u.L/well of N G M that was 72 changed with fresh N G M every 4 days for 14 days in vitro prior to experiments to allow the extension of neurites. 28 2.1.3. Protein Analysis 2.1.3.1. Sample Preparation To prepare crude cell or tissue lysates, starting material was homogenized in at least 10 volumes of RIPA Lysis Buffer (20mM tris pH 7.5, 5mM EDTA, 50mM NaCl, 10mM Na Pyrophosphate, 50mM NaF, 1% NP40 alternative, 1x complete protease inhibitor (Roche). Homogenates were sonicated on ice for 15s and centrifuged at 21000 R C F 5min to remove cellular debris. 2.1.3.2. Membrane Fraction Preparation To improve detection of ABCG1 protein in mouse tissue, crude membrane fractions were purified from murine tissue to eliminate cytosolic proteins prior to Western blotting 1 2 7 . First, 100mg of tissue was homogenized in 1250uA. Low-ionic Strength Buffer (50mM mannitol, 2mM EDTA, 50mM Tris-HCL pH7.6, 1x Complete Protease Inhibitor (Roche)), centrifuged at 500g for 5min to pellet nuclei and debris, followed by collection of the supernatant. To obtain the purified membrane fraction, 800u,L of supernatant was layered on a 500u,L mannitol cushion (300mM mannitol, 2mM EDTA, 50mM Tris-HCL pH7.6) and centrifuged at 100,000 R C F in a fixed angle rotor TLA 120.2 (Beckman) for 45min. The supernatant was discarded and the membrane-fraction pellet was re-suspended with 200u.L of Low-ionic Strength Buffer supplemented with 1% SDS. The resuspended membrane fractions were sonicated for 15s to shear DNA that could interfere with subsequent S D S - P A G E procedures. 2.1.3.3. Protein Determination assay The protein content of tissue and cell lysates was determined by DC Protein assay (BioRad) on 0.5-2ul of sample containing 0.25-12|ig protein and was compared to a standard curve of known delipidated-BSA concentrations. 29 2.1.3.4. SDS-Polyacrylamide Gel Electrophoresis (SDS PAGE) To facilitate immunodetection of proteins, protein samples were separated by size using S D S - P A G E and transferred to immobilized membranes. Typical running conditions were as follows. To prepare samples for S D S - P A G E , the protein content of sample lysates was determined by protein assay and diluted with RIPA buffer to contain equal protein concentrations in equal volumes. The normalized samples were mixed 5:1 with loading buffer concentrate containing 200 mM Tris/HCI pH 6.8, 20% p-Mercaptoethanol; 5% SDS, 20% Glycerol, Bromphenol Blue. S D S and p-mercaptoethanol were included to ensure complete solubilization and saturation of the protein with S D S and reduction of disulphide bonds. Once mixed, samples were heated at 94C for 5min to ensure hydrolysis of DNA and ensure complete denaturation of all proteins. Protein samples were loaded into polyacrylamide gels either 0.75mm or 1.5mm thick. Gels contained 3 volumes separating gel to 1 volume stacking gel to ensure good resolution. Separating gels contained 10% acrylamide/bis 375mM Tris pH 8.8, 0.01% S D S , 0.005% ammonium peroxydisulphate (APS), pH 8.8. Stacking gels contained 3% bis-acrylamide, 0.125mM Tris pH 6.8, 0.02% S D S , 0.01% A P S . Polymerization was initiated with T E M E D (Biorad). Anode and cathode buffer reservoirs were filled with Running Buffer containing 6mM Tris pH 8.4, 4.7mM glycine, 1u.M S D S . Gels were run 1-1.5h at 100V at room temperature. Tris-Tricine Gels To resolve C83 and C99 CTFs , a Tris-Tricine buffer system was used to separate proteins on step gradient gels that maximized resolution of the CTFs and A P P holoprotein. Samples were prepared by mixing 4:1 with Tris-Tricine loading buffer concentrate containing 30 (100mM Tris-HCI pH 6.8, 30% glycerol, 8% S D S , 0.02% Commassie Blue G-250, 3% DTT). Step gradient gels consisted of 3 layers: stacking gel (4% acrylamide, 0.75mM Tris pH 8.5, 0.02% APS) , middle separating layer (10% acrylamide, 1M Tris pH 8.45, 13% glycerol, 0.02% APS) and bottom separating layer (17% acrylamide, 1M Tris pH8.45, 13% glycerol, 0.02% APS) polymerized with T E M E D (BioRad). Gels were poured with a ratio of 1:1:3 (4%:10%:17%)(v/v/v). The cathode chamber was filled with Cathode Running buffer (100mM Tris-base pH 8.5, 100mM tricine, 0.1% SDS) and the anode buffer reservoir was filled with anode buffer (200mM Tris-base pH 8.9). Tris-Tricine gels were run at 110V for 5-6h at room temperature and transferred to nitrocellulose membranes. Transfer: To transfer proteins from polyacrylamide gels to immobilized nitrocellulose or polyvinylidine fluoride membranes for western blotting, gels were sandwiched with membrane and blotting paper and immersed in Transfer Buffer containing 25mM Tris pH 8.4, 190mM glycine, 10% methanol. Gels were transferred at 24-36V for 3-24h at 4°C. 2.1.3.5. Western blotting The following Western blot protocol was used to detect proteins except where noted. Membranes were blocked in 5% skim-milk in Phosphate Buffered Saline (PBS) containing (10mM N a 2 H P 0 4 , 1.4mM K H 2 P 0 4 , 137mM NaCl, 3mM KCI) for 1h to reduce non-specific antibody binding in subsequent steps. All incubations were performed with agitation. Primary antibodies were prepared in 5% skim milk powder in P B S . Antibody dilution and incubation time varied for each antibody as described below. Following incubation with primary antibodies, the membranes were washed 3x 10 minutes with P B S supplemented with 0.05% Tween-20 detergent (PBS-T). Secondary detection was performed using HRP-conjugated antibodies specific to the host-species of the primary antibody, diluted 1:2000 in 5% skim milk powder in P B S and incubated for 30min to 2h. Following the binding of secondary antibodies, 31 membranes were washed 3x 10min in PBS-T, immersed in Enhanced-Chemiluminescent reagent (Amersham) for 1 min and exposed to photographic film. A B C G 1 : Total cell lysates were used to detect ABCG1 in HEK cells, whereas crude membrane fractions were used to detect ABCG1 in murine tissues. Samples were diluted with loading dye and heated for 5min at 55°C. Western blot conditions were as previously listed, except P B S and PBS-T were replaced by Tris-Buffered-Saline (0.3% Tris pH 7.4, 0.8% NaCl, 0.02%KCI)(TBS) and TBS-T (TBS, 0.05%Tween-20). The primary antibody was rabbit anti-A B C G 1 NB400-132 (Novus) diluted 1:750. Secondary detection was by goat anti-rabbit-HRP conjugated (ImmunoLabs) diluted 1:2000 in 1% skim-milk T B S and incubated for 2h. G A P D H We determined G A P D H levels to control for protein loading. The primary antibody was mouse monoclonal mAB374 (Chemicon) diluted 1:5000 for 30min. Secondary antibody conditions were 1:2000 goat anti-mouse-HRP incubated for 45min. A P P - C T F For experiments involving the detection of A P P CTFs, nitrocellulose membranes were cut into holoprotein and CTF containing segments at a position corresponding to approximately 60kDa. The holoprotein membrane segment was incubated with rabbit anti-APP-C-terminal antibody (Sigma-Aldrich) diluted 1:5000k for 1h and with secondary bovine anti-rabbit-HRP diluted 1:2000 for 45min. The CTF containing membrane was incubated in rabbit ant i -APP-C terminal antibody (Sigma) diluted 1:5000 overnight. Secondary antibody conditions were goat anti-rabbit-HRP diluted 1:2000 for 2h. 32 A P P To detect A P P , the primary antibody was mouse monoclonal 6E10 (SantaCruz Biotech) diluted 1:1000 for 1h. Secondary antibody conditions were goat anti-mouse-HRP diluted 1:2000 and incubated for 1h. ApoE To detect apoE protein, proteins were transferred to Immobilon-P P V D F membranes (Millipore). The primary antibody was polyclonal rabbit anti-apoE M-20 (SantaCruz Biotech) diluted 1:750 for 2h. Secondary antibody conditions were bovine anti-rabbit-HRP diluted 1:2000 and incubated for 1hr. Quantitation Western blot data was quantitated using densitometry. Photographic film was scanned into TIFF format at 600dpi resolution and pixel counts were determined using ImageJ (NIH) software. 2.1.4. Messenger RNA Analysis 2.1.4.1. mRNA Extraction Total mRNA content was extracted using Trizol (Invitrogen) according to the manufacturer's protocol from cells or snap-frozen tissue that was previously stored at -80°C. Extracted mRNA was and stored at -80°C for fewer than 3 months prior to use. To reduce DNA contamination, samples were treated with DNAsel prior to cDNA synthesis. 2.1.4.2. Q R T - P C R The level of mRNA was determined using quantitative real-time polymerase chain reaction (QRT-PCR). Prior to Q R T - P C R , mRNA transcripts were converted to cDNA using oligo-dT primers and Taqman Reverse transcription reagents (Applied Biosystems). Primers 33 were designed using PrimerExpress software (Applied Biosystems) and spanned exons 22-23 of human Abcgl. Sequences are: Human Abcgl forward (5' A C A C C A T C C C C A C G T A C C T A 3') and reverse (5' G A T G A C C C C T T C G A A C C C A 3') and murine Abcgl forward (5' C A A C G T G G A T G A G G T T G A G A 3') and reverse (5' C T G G G C C T C T G T G A A G T T G T 3'), and murine B-actin forward (5' A C G G C C A G G T C A T C A C T A T T G 3') and reverse (5' C A A G A A G G A A G G C T G G A A A A G 3'). Q R T - P C R was done with Sybr green reagents (Applied Biosystems) on an ABI 7000. Cycling conditions were 50°C for 2 min, 95°C for 10 min, then 40 cycles at 95°C for 15 sec and 60°C for 1min, followed by dissociation at 95°C for 15 sec, 60°C for 20 sec, and 95°C for 15 sec. Each sample was assayed in triplicate, normalized to p-actin to control for RNA loading and analysed with 7000 system S D S software v1.2 (Applied Biosystems) using the relative standard curve method. 2.1.5. Lipid Analysis 2.1.5.1. Total Lipid Analysis Lipid Extraction: To determine total lipid content of brain tissue, 25mg of tissue was homogenized in 0.9% saline and protein content was determined by Lowy assay (BioRad). Brain homogenates were mixed with 2:1 chloroform:methanol mixture to obtain a final dilution of 6:3:1.25 chloroform:methanol:saline containing betulin as an internal standard. Samples were spun at 400RCF for 5 minutes to separate the aqueous and organic phases. The organic layer was isolated and an equal volume of chloroform was added back to the tissue homogenate to re-extract lipids from the aqueous phase. Samples were vortexed and centrifuged as before and the organic layers were pooled. The pooled organic fractions containing total lipid extracts were evaporated under nitrogen and solubilized in chloroform-methanol-acetone-hexane (2.0:3.0:0.5:0.5) for storage at -20C until analysis. Lipid Measurements Lipid analysis was performed in collaboration with Dr. Sheila Innis (University of British Columbia, Vancouver) as described 1 4 1 . Separation of polar and non-polar lipids, including individual classes, was achieved using a H P L C (Waters 2690 Alliance HPLC, (Milford MA), equipped with an auto-sampler and column heater. The sample chamber was kept at 18°C and the column heater at 35°C. The column was a Waters YMC-Pack Diol 120NP, 25 cm 4.6 mm id, 5 m particle size and 12 nm pore size. Using a quaternary solvent system of hexane-petroleum ether, 97:3 (v,v); methanol-triethylamine-acetic acid, 765:15:13 (v/v/v); acetone-triethylamine-acetic acid, 765:15:13 (v/v/v); isopropanol-acetic acid, 800:40 (v/v) in a linear gradient with a flow rate of 2 mL/min. The column eluant was split 10:90 to an E L S D (Alltech, model 2000; Mandel Scientific, Guelph, Canada) and a fraction collector (Gilson FC204, Mandel Scientific).ELSD detection and quantitation of the separated lipid classes was performed with a nitrogen flow rate of 1.8 ml/min, a drift tube temperature of 60°C. Calibration curves to determine the linear range of the analysis were established using authentic standards for each lipid class and samples were quantified using the external standard method. 2.1.5.2. Cholesterol metabolites Lipids were extracted from tissues by a modified Folch procedure. Briefly, dried tissue samples were submerged in chloroform:methanol (2:1) overnight in a capped dram vial. The following day, solvent volume was adjusted to 20X the dry weight, and 0.25x volumes of 0.9% NaCl was added, mixed vigorously for 1 min, and equilibrated by shaking for 2h at RT. Samples were centrifuged at 2000 rpm for 5 min to separate phases. The upper aqueous phase was removed, and lipids were dried under an N 2 stream. Sterol levels were determined after derivatization to the corresponding trimethylsilyl-ethers by gas chromatography-flame ionization detection (GC-FID) and gas chromatography-mass spectrometry (GC-MS) as reported previously 1 4 2 . 35 2.1.5.3. Cholesterol efflux The mass of cholesterol transferred to apolipoprotein, or lipoprotein acceptors was used to assess cholesterol transporter activity as described 1 4 3 . HEK cells: Cells were seeded in 24 well plates at 250,000 cells / well for efflux. Cellular cholesterol was labeled with 1u.Ci/mL of H 3 cholesterol for 24h during transfection in growth media. Transfected and labeled cells were washed once with Efflux Media (DMEM, 0.2% BSA, 2mM L-Glutamine) and incubated for 4h efflux in either efflux media or efflux media containing human HDL2/3 (25ug/mL HDL protein). Primary Astrocytes: Astrocytes were seeded in 24 well plates and allowed to adhere overnight. To prepare efflux Loading Media, Acetylated-LDL (Intracel) was incubated with H 3cholesterol(New England Nuclear) for 30min at 37°C before dilution in DMEM, 1% F B S , 2mM L-glutamine, 100U/ml penicillin, 100mg/mL streptomycin to obtain a final concentration of 40|ag/mL AcLDL protein and iLiCi/mL H 3cholesterol. Cells were incubated in Loading Media for 24h, washed once in efflux media and equilibrated 4h in Efflux Media to internalize membrane bound lipoproteins and distribute radiolabeled cholesterol within the cell. Cells were washed once in efflux in media prior to 4h efflux in 250u l Efflux Media supplemented with either nothing, 15u.g/mL delipidated human apoAl (Calbiochem), 15u.g/mL delipidated human apoE3 (Calbiochem) or human HDL2/3 containing 25u,g/mL HDL protein. Quantification After incubation with efflux media, the cell media was collected and spun for 5 min at 21000 R C F to remove cells and was mixed with two volumes of OptiPhase SuperMix 36 scintillation fluid (PerkinElmer). Cells were lysed with 100^L 0.1% NaOH, 0.2% S D S and mixed with 3 volumes of scintillation fluid. The H 3 content of cells and media were determined by scintillation counting to calculate efflux. Efflux = (media counts) / (media counts + cell counts). 2.1.6. LacZ staining To prepare tissue for Lac-Z staining, wildtype and Abcgl hemizygous mice were perfused with P B S . Brains were immersion fixed in 2% paraformaldehyde P B S for 24h and cryoprotected in 30% sucrose in P B S at 4°C. Frozen coronal sections were prepared on a cryostat and mounted on Superfrost Plus Slides (Fisher). LacZ staining was performed in the presence of 1mg/mLX-gal. Sections were counterstained with Neutral Red and dehydrated prior to being coverslipped. 2.1.7. AU Measurements 2.1.7.1. Af i Secretion Assays The secretion of AU from HEK:APPSwe cells was measured from transiently transfected cells seeded in 10cm plates. To measure AU secretion, 10 ml of fresh Growth Media was added to cells for 6h, then removed and spun at 21000 R C F for 5 min to pellet contaminating cells. The cell-free media was supplemented with complete protease inhibitor (Roche) to reduce degradation of AU and was frozen at -80°C until further use. Plated cells were scraped into 1mL of P B S , spun at 21000 R C F to remove supernatant, then homogenized in 200uL RIPA buffer as previously described. The total protein content of RIPA lysates was determined by Lowry assay. To determine the AU40 and AU42 content of conditioned media, media samples were diluted 5-1 Ox and measured by ELISA (Biosource) specific to human AU40 or 42 as per manufacturers instructions. Raw AU values were normalized to total plate protein levels to account for plate to plate differences in seeding density and differences in cell-survival between A B C G 1 and empty vector transfected cells. 37 2.1.7.2. Murine Aft measurements Endogenous murine A S levels were quantified using 75 ug of protein from total brain homogenates by Dr. Ronald DeMattos of Lilly Research Laboratories, Indianapolis. The coating antibody used was R163 for AG-40 (a generous gift from Dr P.D. Mehta, New York Institute for Basic Research, Staten Island, New York). Plates were coated overnight at 4°C and then blocked with P B S containing 0.1% BSA for 2h at room temperature. Following 5 washes with TBS-T (Tris-buffered saline containing 0.1% Tween-20), samples were incubated for 2h with biotinylated 4G8 (Signet Laboratories Inc) for 1h with agitation. The plates were washed 5X and streptavidin-alkaline phosphatase complex was added for 1h. Plates were finally washed 5x with TBS-T and once with water and AttoPhos reagent (Calbiochem) was added for 30-60 min and fluorescence was determined using a Bio-Tek FL600 fluorescence microplate reader. Murine AG. levels were normalized to total protein as measured by DC Protein Assay (BioRad). 2.1.7.3. Human AG measurements Human AG levels in tissues were determined by Dr.Ron DeMattos, Lilly Research Laboratories, Indianapolis as described 1 4 4 . Briefly, hippocampal tissue was homogenized in P B S and spun 5 min at 21K R F C to pellet insoluble debris and supernatant was used for soluble AG measurements. Cellular debris was resuspended in denaturing buffer containing 5 M guanidine plus protease inhibitors. The extracts were diluted and analyzed in denaturing ELISAs containing a final concentration of 500 mM guanidine for AG T otai (m266 as capture antibody and biotinylated m3D6 as reporter antibody) or AG42 (m21F12 as capture antibody and biotinylated m3D6 as reporter antibody). AG40 levels were determined by subtraction. 38 2.1.8. ApoE ELISA apoE levels in C S F were determined by ELISA as described 1 4 5 . To prepare ELISA plates, MaxiSorp(Nunc) 96-well plates were coated with capture antibody W U E 4 generously provided by Dr. Holtzman (Washington University, St.Louis) diluted at 4mg/mL in Carbonate ELISA Buffer (50mM N a H C 0 3 p H 9.6) overnight at 4°C. All incubations were performed with agitation. Plates were washed 5X with P B S supplemented with 0.25% Tween-20 (PBS-T) and blocked with 1% skim milk powder in PBS-T for 1h. Samples and standards were prepared ELISA Buffer (PBS-T, 0.5% BSA). apoE standards were prepared from Swiss-Webster mouse plasma calibrated to contain 61.7u.g/mL apoE. The standard curve ranged from 0.8ng/mL to 50ng/mL apoE. Samples and standards were loaded onto the plate and incubated overnight at 4°C. To wash the plate, wells were rinsed 5x with PBS-T. The detection antibody was goat polyclonal anti-apoE (Calbiochem cat#178479) diluted 1:6000 in ELISA Buffer and was incubated for 1.5h at room temperature. The plate was then washed 5x with PBS-T prior to incubation with biotinylated anti-Goat (Vector BA-9500) diluted 1:50,000 in ELISA Buffer for 1.5h. The plate was washed again 8x with PBS-T prior to incubation with streptavidin-poly-HRP (Pierce) diluted 1:6000 in ELISA buffer for 1.5h. To quantify apoE, plates were washed 8x with PBS-T prior to the addition of ELISA-TMB (Sigma-Aldrich) until adequate development of the coloured product. The ELISA reaction was stopped with the addition of 1N HCL. Absorbance at 450nM was measured with a MultiScan microplate reader (Fischer). 2.1.9. Statistical Analyses: All p-values reported are derived from two-tailed Student's t-test analysed by Prism GraphPad software 3.0.4. Error bars reflect standard deviation. 39 2 . 2 . R E S U L T S 2.2.1. ABCG1 overexpression in vitro To determine the impact of ABCG1 overexpression on A P P processing, we first conducted experiments in human embryonic kidney (HEK) 293 cells stably expressing human APP695Swe (HEK:APP cells) that were transiently transfected with a human Abcgl cDNA or empty vector as a negative control. Although HEK cells are non-neuronal in origin, they do express all of the components of a, Ii and y secretase required to process A P P into AB and have been used extensively to study A P P metabolism 1 4 6 , 1 4 7 . 2.2.1.1. ABCG1 overexpression increases cholesterol efflux to HDL To validate the functionality of the Abcgl cDNA construct used in our transfection assay, we tested whether transient ABCG1 overexpression increases cholesterol efflux to HDL, which is the best-established assay for ABCG1 enzymatic activity 1 0 7 , 1 1 6 , 1 1 7 . The level of cholesterol efflux from H E K . A P P cells transfected with wildtype A B C G 1 was compared to cells transfected with either empty vector, or to ABCG1 containing a S220G mutation (ABCG1-S220G) in a conserved residue in the ATP.-binding domain that is expected to eliminate functional activity. H E K : A P P cells expressing wild-type A B C G 1 exhibited a 21% increase in cholesterol efflux to HDL (p=0.003 N=6 (Figure 2-1A)) compared to either empty vector or ABCG1-S220G, which were indistinguishable from each other. These results confirm that A B C G 1 functions as a cholesterol transporter in H E K : A P P cells and that the S220G variant is functionally inactive with respect to cholesterol efflux activity. 2.2.1.2. A B C G 1 overexpression promotes AG. secretion in vitro Having validated the H E K : A P P model system, we then assessed the impact of ABCG1 overexpression on AG production. H E K : A P P cells were transfected with empty vector, functionally inactive S220G ABCG1 or wild-type A B C G 1 , and conditioned media was collected 40 Cholesterol Efflux B ^1000 CO I 750 CD 500 CD 250 ca APP-vector M1APP-ABCG1 IZZl APP-S220G no acceptor | _ | D L A&40 ** ** AG42 I APP Vector 1APP ABCG1 APP ABCG1 S220G D Confirmation of ABCG1 expression vector A B C G 1 ~85kDa A B C G 1 G A P D H Figure 2-1: A B C G 1 overexpress ion increases cholesterol efflux to HDL and increases the secret ion of AB40 and AB42. (A) H E K - A P P cells were stably transfected with empty vector, ABCG1 or ABCG1 harbouring an S220G mutation in the catalytic domain. Expression of ABCG1 enhanced the efflux of cholesterol to HDL over 5 hours, compared to either empty vector or ABCG1 S220G. Data are representative of 2 independent experiments done in triplicate. (B) The overexpression of ABCG1 in H E K - A P P cells increases the secretion of AI340 (N=2) and (C) AB42 into cell culture medium over a 6 hour period (N=2). This effect requires the transporter activity of A B C G 1 . (D) Western blot analysis confirms overexpression of ABCG1 in H E K - A P P cells at the protein level. ns=p>0.05, *p<0.05, **p<0.01 41 to measure AB secretion from cells. Compared to empty vector, wild-type A B C G 1 increased the levels of secreted AB40 by 60% (p<0.05, N=2) (Figure 2-1B) and AB42 by 28% (p<0.05, N=2) (Figure 2-1C). AB secretion from A B C G 1 - S 2 2 0 G transfected cells was similar to empty vector. These results demonstrate that ABCG1 increases AB secretion from cells, and that the functional activity of ABCG1 as a cholesterol transporter is required for this effect. 2.2.1.3. A B C G 1 overexpression increases A P P processing in vitro The level of AB secreted from H E K : A P P cells was increased by the presence of functional ABCG1 however, additional experiments were required to determine the underlying mechanism. As changes in membrane cholesterol levels are reported to influence AB production by modulating the proteolytic processing of A P P by secretases, we measured A P P processing intermediates that reflect the activity of a and B secretase. During the proteolytic processing of A P P into AB, C-terminal processing intermediates of either 83 or 99 amino-acids in length are created at the plasma membrane through proteolysis by a or B secretase respectively (Figure 1-1) 1 4 8 , 1 4 9 . The membrane bound CTFs are internalized where they can be further cleaved by y-secretase. C83 is cleaved by y-secretase into a harmless product known as p3. Alternatively, C99 is processed into AB. Because C99 is the direct precursor to AB, the levels of intracellular C99 fragments are the gold standard to estimate the relative rate of AB production 1 4 6 , 1 4 7 . To determine if elevated A P P processing accounted for the increased AB secreted from ABCG1 transfected cells, C83 and C99 levels were measured by Western blot. Western blots were prepared from RIPA total cell lysates from H E K : A P P cells transfected with ABCG1 or empty vector, and were immunoblotted with antibodies specific to the last 19 amino-acids of the C-terminus of A P P to detect C83 and C99 (Figure 2A). Compared to vector-transfected controls, C83 levels in H E K . A P P cells transfected with A B C G 1 were increased 2.9 +/-.4-fold (p=0.014, N=3) (Figure 2-2B) and C99 levels were increased 4.5 +/-0.7-fold (p<0.006, N=3) (Figure 2-2D). Vector A B C G d_fi_>^ Bi_> {fife ' B 4 a I • X • CO U 0.0 C83 1 —.— p=0.0137 • — , — — Vector ABCG1 g 0.25H 1.00i o I 0.75 | o^ g 0.25 0.00 A P P C 9 9 C 8 3 | G A P D H C83 /APP p=0.0136 Vector ABCG1 C99 /APP p=0.0002 Vector ABCG1 Vector ABCG1 Figure 2-2: A B C G 1 overexpress ion increases cel lular A P P levels and increases the process ing of A P P into C T F s . (A) Western blots of H E K - A P P cell-homogenates previously transfected with either empty vector or ABCG1 were probed with antibodies specific to the C-terminus of A P P to detect full-length unprocessed A P P or CTFs . (B,D) C83 and C99 levels are increased in ABCG1 transfected cells. (C,E) When normalized to total A P P levels, C83 /APP and C99 /APP are significantly increased compared to empty vector. Total unprocessed A P P holoprotein in ABCG1 transfected cells was non-significantly increased by 1.6-fold (p=0.094, N=3) compared to cells transfected with empty vector. To determine whether the increase in CTF levels reflected an increase in A P P processing or could reflect the trend toward elevated A P P levels, C83 and C99 levels were expressed as a ratio to total A P P . W e found that compared to empty vector, ABCG1 overexpression increased the C83 to A P P ratio by 1.7 +/-0.16-fold (p<0.014, N=3) (Figure 2-2C) and the ratio of C99 to A P P by 2.8 +/-0.13-fold (p<0.0002, N=3) (Figure 2-2D). The finding that C99 levels are increased, even when corrected for total A P P levels, confirms that ABCG1 overexpression increases AU production under these experimental conditions. 2.2.2. A B C G 1 expression in brain In individuals with AD, neuropathology is first detected in the hippocampus and entorhinal cortex 1 5 ° . Although the precise mechanism behind amyloid and NFT formation in sporadic AD is not completely understood, the amyloid cascade hypothesis suggests that the production and secretion of AU peptides by neurons in these regions is a crucial step in AD pathogenesis 3 2 , 1 5 1 , 1 5 2 . AU peptides are produced by several cell types in the body including neurons and platelets, and neurons are reported to produce the majority of AU within the central nervous system 1 5 1 , 1 5 2 . Because our in vitro results show that A B C G 1 overexpression increases AU production, determining the physiological relevance of this finding requires knowledge of whether ABCG1 is expressed in neurons. 2.2.2.1. Abcgl is highly expressed in hippocampal neurons The presence of Abcgl mRNA in brain had been reported prior to this s tudy 1 0 8 , 1 2 6 , however the specific regions and cell types that express Abcgl were not well described. To further address the question of Abcgl expression within the brain, we used Abcgl knockout mice that were generated by inserting a U-galactosidase (LacZ) reporter construct into the coding region of Abcgl. Because the LacZ reporter gene is expressed under the control of 44 endogenous Abcgl regulatory sequences, Abcgl expression patterns in the C N S can be accurately determined by staining for U-galactosidase activity. Coronal sections of adult mouse brain hemizygous for Abcgl were therefore stained with X-gal to visualize Abcgl-expressing cells (Figure 2-3). Comparison of X-gal staining in Abcg1+I- sections (Figure 2-3A,2-3A) to Abcgl +/+ brains (Figure 2-3B, 2-3D) reveals marked Abcgl expression in the dentate gyrus and CA1 , CA2, and CA3 neuronal layers of the hippocampus, and detectable Abcgl expression in cortical layers (Figure 2-3F-I). These observations confirm that Abcgl is indeed expressed in neurons that are highly relevant to AD pathogenesis. Intermediate staining intensity indicative of moderate Abcgl expression was also observed in striatal and thalamic regions. Although these results suggest that Abcgl is predominately expressed in neurons, they do not exclude the possibility that Abcgl may also be expressed in other cell types in the brain. For example, because Abcgl is known to be expressed in peripheral macrophages, which are highly related to microglia in the brain, we expected to observe Abcgl expression in microglia. Abcgl expression has also been reported in astrocytes 1 5 3 . What is clear from the histological results is that expression of Abcgl is high in neurons under basal conditions, particularly for hippocampal neurons known to be highly relevant for AD. These observations support the hypothesis that A B C G 1 could accelerate AU production and AD neuropathology as observed in DS, by increasing the level of AU production in neurons. 2.2.3. Generation of a novel animal model of ABCG1 overexpression Our in vitro findings show that ABCG1 overexpression increases the rate of AU production from cells. The major caveat of these findings however, is the use of a non-neuronal cell model in which A B C G 1 and A P P were both highly overexpressed. To confirm our in vitro findings in a physiologically relevant context an A B C G 1 overexpressing mouse was therefore generated. 45 Figure 2-3: Abcgl express ion in brain. (A) Coronal sections from mice hemizygous for a LacZ reporter gene under the control of Abcgl regulatory sequence were stained with X-gal to detect the presence of the reporter and were counterstained with neutral red. Abcgl expression is detected in the hippocampus, cortex, striatum and thalamus. (B,F) Wildtype sections show no endogenous lacZ activity. (C) Abcgl is highly expressed in neurons of the CA1 , CA2 CA3, and Dentate Gyrus of the hippocampus. (E,G,H,I) Abcgl expression is enriched in cortical layers. 2.2.3.1. Transgenic strategy Because we were attempting to recapitulate overexpression of Abcgl through an increased gene dose as seen in DS, the transgenic strategy employed needed to overexpress the correct Abcgl mRNA isoforms in the correct cell types; information that was not available at the time this model was generated. We therefore selected genomic Abcgl maintained in a BAC as a transgene, as this approach has been shown to produce overexpression of many transgenes in a physiologically regulated way that maintains proper cell type and developmental expression patterns 1 5 4 . The BAC is essential to this approach as bacterial plasmid vectors cannot maintain sufficiently large inserts to accommodate genomic DNA spanning most eukaryotic genes 1 5 4 , 1 5 5 . The ABCG1 BAC insert used here contains a 141 kb region of human chromosome 21 that contains all introns and exons of Abcgl, 31 kb of upstream regulatory sequence containing all known promoter elements and no other known genes. Purified B A C DNA was injected into mouse oocytes for random integration into the genome. Tail DNA from the resulting progeny were tested by P C R for inheritance of the BAC using human specific primers (data not shown), and 3 transgenic founders from independent injections were verified. After 3 generations of backcrossing to congenic C57/BI6 mice, the relative levels of transgene expression was measured by Q R T - P C R in the brain. Brain was chosen for these experiments as A B C G BAC Tg mice would be used primarily to evaluate the role of A B C G 1 on lipid homeostasis and A D neuropathology in the brain. The expression of human Abcgl in mice derived from founder 2 was ~73-fold greater than mice derived from founder 3, that showed the lowest level of transgene expression (Figure 2-4). Line 2T was derived from the progeny of founder 2, however these mice lacked human exons 1-4 (data not shown) and expressed human Abcgl at low levels. We hypothesize that the line 2T created by the integration of a truncated BAC in founder 2, that segregated from the full-length BAC during backcrossing. Because Abcgl expression in line 2 was ~5-fold greater than the next highest expressor, line 2 was chosen for further experiments. The DNA sequence of human Abcgl 47 BAC Expression in Brain 125n c o 0 ' I ' ' I I I Line 1 Line 2 Line 2T Line 3 Figure 2-4: A B C G 1 B A C express ion in brain of t ransgenic founders. Expression of the ABCG1 BAC in brain of four independent transgenic strains was compared to select the highest expressing line for further experiments. Human Abcgl mRNA was determined by Q R T - P C R and normalized to the lowest expressing line for comDarison. exons 1-23 in line 2 was validated by P C R to confirm the absence of coding mutations (data not shown). 2.2.3.2. Expression of the Abcgl transgene is physiologically regulated in vivo Once the A B C G 1 BAC transgenic mice had been generated, transgene expression was characterized to establish whether the ABCG1 BAC Tg mouse was an accurate model of increased A B C G 1 gene dose. First we measured mRNA levels in 12 different tissues previously reported to express Abcgl to determine if Abcgl was properly expressed across tissue specific axes 1 0 7 . The levels of either human or murine A B C G G 1 mRNA were determined by Q R T - P C R and were normalized first to the levels of fi-actin to control for RNA loading and next (in a ratiometric manner) to Abcgl mRNA in liver for each species (Figure 2-5). Because the method used to quantify mRNA cannot directly compare the amount of human 48 to mouse mRNA, the results were normalized to an internal reference tissue and expressed as fold-increase. Liver was chosen as our internal reference as Abcgl was reported to be expressed in liver and would permit the comparison of our results to previously published data that were also normalized to liver. Relative levels of human and murine mRNA did not significantly differ in 10 of the 12 tissues examined as determined by Student's t-test (Figure 2-5). The two tissues that did show differences between endogenous and transgenic expression were adrenal gland and spleen, where human Abcgl expression was 76% lower than endogenous murine Abcgl in adrenal gland (p<0.05) and 75% lower in spleen (p<0.001)(N=3 human Abcgl and N=6 murine Abcgl for all levels measured). Human Abcgl mRNA in eye was non-significantly increased compared to murine expression. Other than these two tissues, the similarities between relative levels of human and murine Abcgl mRNA suggest that the BAC is expressed in a physiologically accurate way. Most importantly for this study, expression of the human Abcgl transgene in brain reflected that of endogenous Abcgl. 2.2.3.3. Murine and human expression of Abcgl We also compared murine Abcgl levels from this experiment with a published report detailing expression of Abcgl mRNA in human tissues 1 0 7 . The authors used Northern blots to quantify Abcgl mRNA levels, defined liver expression as 100% and other tissues were expressed as a percentage of liver. Because the mRNA levels in Figure 2-5 are normalized to liver, these murine data can be compared to the human data. Results of the studies are summarized in Table 2-1. These results suggest that patterns of Abcgl expression are substantially different between mice and humans. 2.2.3.4. The ABCG1 BAC is expressed in neurons Because neurons generate the majority of A S peptides in the brain, it was important to determine if the Abcgl transgene was expressed murine neurons 1 5 1 , 1 5 2 . To measure mRNA levels in neurons from the BAC transgenic mice, cortical neurons were cultured from ABCG1 49 Abcgl T issue Panel Murine Human Figure 2-5: The tissue-specific pattern of the ABCG1 BAC transgene expression is similar to murine Abcgl. Abcgl mRNA levels were determined by Q R T - P C R , normalized to beta-actin to control for loading and expressed as fold-induction relative to liver for comparison. (A) Murine Abcgl mRNA was obtained from wildtype male mice and measured using murine specific primers to exon 2 (N=4). (B) Human Abcgl mRNA was obtained from ABCG1 BAC Tg males and was amplified using human specific primers spanning exons 22-23 (N=3). BAC Tg and wildtype E16 embyros. Neurons were allowed differentiate for 14 days in vitro before treatment with TO901317, a synthetic LXR ligand known to stimulate expression of genes involved in cholesterol efflux including Abcgl 1 5 3 , 1 5 6 . Neuronal mRNA was harvested after 24 hours of treatment and murine and human Abcgl mRNA was quantified by Q R T - P C R . Compared to vehicle treatment, the levels of murine Abcgl mRNA were increased 1.41± 0.04-50 fold (p =0.005, N=4) by treatment with 0.5uJvl TO901317, or 1.34-fold ± 0.09 (p=0.039 N=4) following treatment with 2u.M TO901317 (Figure 2-6). Similarly, the levels of human Abcgl mRNA were significantly increased relative to vehicle by 1.57± 0.01-fold (p=0.002, N=4) after treatment with 0.5u.M TO901317 and by 1.96 ± 0.39-fold (p=0.05, N=4) after treatment with 2u,M TO901317 (Figure 2-6). This result shows that both human and murine Abcgl are expressed in Abcgl transgenic neurons. Furthermore TO901317 induced similar amounts of Tissue Relative Abcgl mRNA Human * Murine Adrena l G land 1.26 4.59 Lung 1.35 3.91 Heart 1.08 1.23 Skeleta l Musc le 0.54 0.82 Brain 0.66 6.45 Sp leen 1.06 4.45 Smal l Intestine 0.61 0.16 Kidney 0.75 1.05 Liver 1.00 1.00 Tes tes 0.23 0.75 * K lucken et a l . (2000) P N A S Table 2-1: Tissue specific levels of Abcgl mRNA in humans and mice. Human Abcgl mRNA levels measured by northern blot were obtained from a previously published report. Murine Abcgl mRNA levels in wildtype mice were determined by Q R T - P C R . Both mRNA species are expressed as relative units. Liver is defined as 1 for reference. 51 Primary Neurons murine human i i 25n • r ™ ' r i • i l i i V 0.5 2.0 V 0.5 2.0 L X R Agonis t T0901317(uM) Figure 2-6: The A B C G 1 B A C is expressed in primary neurons and responds to an L X R agonist. Primary neurons obtained from embryonic A B C G 1 B A C Tg mice were cultured 14 days in vitro and treated with either vehicle, or increasing concentrations of LXR agonist TO901317 for 24h and mRNA levels were determined by Q R T - P C R . (A) Compared to vehicle treated cells, LXR treatment increased murine Abcgl expression by ~1.4-fold (N=2). (B) Compared to vehicle, human Abcgl expression was increased by ~1.5-fold by 0.5LIM (N=2) and -1.9-fold (N=2) by 2uM TO901317 respectively. These data are representative of 2 experiments. * p<0.05 human and murine mRNA expression in neurons, suggesting that both the human transgene and murine Abcgl respond similarly to oxysterol ligands. 2.2.3.5. ABCG1 transgenic mice overexpress ABCG1 protein 3-8 fold in peripheral tissues Because the correlation coefficient between mRNA and protein levels is less than 0.5 1 5 7 , and absolute levels of human Abcgl mRNA cannot be directly compared to endogenous murine Abcgl mRNA levels, it is crucial and more relevant to determine the levels of ABCG1 overexpression in the ABCG1 BAC Tg mice at the protein level. Detection of ABCG1 protein in total tissue homogenates proved to be insensitive using all available ABCG1 antibodies (data not shown). We therefore enriched for ABCG1 by preparing crude membrane fractions from 52 total tissue homogenates over a mannitol gradient, and these membrane fractions were used to prepare Western blots of ABCG1 in tissues. Using this method, we observed that ABCG1 protein was robustly expressed in spleen, liver and lung of A B C G 1 BAC Tg mice compared to nontransgenic littermate controls (Figure 2-7A). In these tissues, the level of ABCG1 protein was 3-5 fold greater in A B C G 1 BAC Tg mice compared to nontransgenic littermates. Because human and murine ABCG1 are 97% similar at the protein level, the antibody used does not noticeably differentiate between human and murine A B C G 1 (data not shown). Thus, the levels of ABCG1 protein shown here reflect the sum of both endogenous murine and transgenic human A B C G 1 protein. 2.2.3.6. A B C G 1 is overexpressed 3-6-fold in cortex, hippocampus and cerebellum Because our A B C G 1 BAC Tg model would be used to assess amyloid pathology and amyloid develops in a model that develops these deposits in a regional specific manner, we performed ABCG1 Western blots on selected brain regions. Compared to wildtype mice, A B C G 1 protein levels in BAC Tg mice were increased 6.0± 0.47-fold (p=0.001, N=3) in cortex, 5.6± 1.5-fold (p=0.038, N=3) in hippocampus and 3.0± 0.48-fold (p=0.025, N=3) in cerebellum (Figure 2-7B). These results confirm that ABCG1 is overexpressed in brain regions relevant to A D in our ABCG1 BAC Tg mice, validating this model as suitable for determining whether physiologically accurate overexpression of ABCG1 is sufficient to accelerate amyloid pathology in vivo. 2.2.3.7. A B C G 1 remains overexpressed in P D A P P mice with amyloid The analyses of ABCG1 protein levels were performed using baseline ABCG1 BAC Tg mice prior to breeding to an animal model of AD. This is important because we had previously observed that the expression of a related A B C transporter, A B C A 1 , was altered upon crossing to the A P P / P S 1 mouse model of AD, suggesting that the presence of amyloid deposits may 53 ABCG1 in Peripheral Tissues Sp leen Lung Liver WT G1 WT G1 WT G1 ABCG1 <wHMNKg9 ABCG1 J ^ J p * * * * * ABCG1 NaKATPase W H » M M N a K A T P a s e « M M M M NaKATPase B A B C G 1 in Brain Reg ions C x Hp C b ABCG1 in Brain c B A C A B C G N a K A T P 10.O1 S 5.<M o 5 « • WT ABCG1 WT ABCG1 WT ABCG1 C x Hp C b H ippocampus C57/Bk6 APP+ (3mo) APP+ (13mo) B A C + + + A B C G 1 N a K A T P Cortex C57/Bk6 APP+ (3mo) APP+ (13mo) B A C + + + A B C G 1 I^^H^^IHHaHHHHHHHH N a K A T P Figure 2-7: A B C G 1 protein is overexpressed in A B C G 1 B A C Tg t issues. Crude membrane fractions were prepared from tissues of wildtype or ABCG1 BAC Tg mice for Western blot of A B C G 1 and NaKATPase transporter as a loading control. (A) Compared to wildtype mice, A B C G 1 is overexpressed 3-5-fold in spleen, lung and liver of ABCG1 BAC Tg mice (N=2). (B) ABCG1 is overexpressed 6-fold in cortex, 6-fold in hippocampus and 3-fold in cerebellum of ABCG1 BAC Tg mice compared to wildtype (N=3). Western blots were summarized by densitometry and wildtype was defined as 1. (C) Human and murine ABCG1 expression is unaffected by either the presence of amyloid and the A P P transgene in 13 month old P D A P P APP+ mice or the A P P transgene alone in 3 month old APP+ mice when compared to either wildtype or ABCG1 BAC+ mice (N=2). * = p<0.05 Cortex (Cx), Hippocampus (Hp), Cerebellum (Cb) 54 influence A B C transporter expression 1 5 8 . To determine if a similar confounding effect may be present after crossing the ABCG1 BAC Tg mice to the P D A P P A D mouse model of AD used in later experiments (2.2.5), brain levels of ABCG1 were also measured in A B C G 1 / P D A P P both prior to and after the onset of amyloid deposition. The Western blots indicate that robust overexpression of the ABCG1 transgene is retained in A B C G 1 / P D A P P mice in both hippocampus and cortex (Figure 2-7C). These results confirm that the A B C G 1 BAC Tg mouse is suitable for testing the impact of A B C G 1 overexpression on amyloid development in the P D A P P mouse model of AD. 2.2.4. Characterization of the ABCG1 BAC transgenic mice To validate the functionality of the ABCG1 BAC in vivo and further study the role of ABCG1 in the brain, previously reported phenotypes of ABCG1 overexpression were assessed in the ABCG1 BAC Tg mice. 2.2.4.1. Cholesterol efflux to HDL is not increased from ABCG1 Tg astrocytes The best established functionalassay for A B C G 1 function in cells is to facilitate cholesterol efflux specifically to lipid-rich apolipoproteins such as HDL and not to lipid-poor apolipoproteins. Importantly, although this function has been demonstrated in numerous studies using transfected cultured cells and A B C G 1 knockout cells 1 0 7 ' 1 1 6 , 1 1 7 much less is known about the ability of physiological overexpression of ABCG1 to mediate efflux. To determine if the ABCG1 BAC was functional in transgenic mice, we measured cholesterol efflux to HDL from primary astrocytes. Astrocytes were chosen for these experiments because they express Abcgl and play an important role in maintaining lipidation of apolipoproteins in brain 1 5 3 , 1 5 9 , 1 6 0 . Primary astrocytes were harvested from 1-2d post-natal mice and cells were grown in culture for 14 days prior to seeding for efflux experiments. In a typical cholesterol efflux experiment, cells of interest are loaded with excess cholesterol, which stimulates expression of genes including Abcgl and AbcA1 that are involved in promoting the efflux of this excess lipid. Cholesterol loading is thought to result in the production of 55 endogenous oxysterol ligands that activate the L X R / R X R pathway and stimulate expression of genes that restore cholesterol homeostasis 1 6 1 . Therefore, astrocytes were cholesterol loaded with acetylated LDL (AcLDL) for 24 hours to stimulate Abcgl expression prior to efflux. Cells were then washed, and cholesterol efflux was measured over 5 hours in either plain media, or in media containing apoAl as a positive control specific for ABCA1-mediated efflux, delipidated human apoE3, which is a natural cholesterol acceptor for ABCA1-mediated efflux from astrocytes and human H D L ^ t o detect A B C G 1-mediated efflux. W e expected that ABCG1 overexpression would result in increased cholesterol efflux to HDL 2 / 3 , but not to lipid-poor apoAl or apoE3. Compared to baseline cholesterol efflux in media-only treated cells, we observed that cholesterol efflux to delipidated apoAl control was similarly increased by 1.75-fold (p=0.0006 N=6) in wildtype astrocytes and 1.8-fold (p=0.003, N=6) in A B C G 1 BAC Tg astrocytes (Figure 2-8A). Cholesterol efflux to apoE was similar to that of apoAl. These results demonstrate that the ABCA1 -specific pathway of cholesterol efflux is operative in primary astrocytes under these experimental conditions. Contrary to our expectations, the efflux of cholesterol from ABCG1 overexpressing astrocytes to HDL was non-significantly increased 1.2+/-0.2 -fold (p=0.066, N=6) compared to wildtype astrocytes (Figure 2-8A). Because these data failed to show the expected increase efflux of cholesterol to HDL from A B C G 1 overexpressing cells, we verified that our cholesterol loading protocol was up-regulating the expression of Abcgl as intended. 2.2.4.2. Abcgl expression is increased following acetylated LDL treatment The AcLDL loading protocol used in these experiments was originally developed for macrophages to mimic exposure to oxidized-LDL that occurs in the artery wall, however astrocytes do not naturally accept lipids from oxidized LDL in vivo. LXR agonists such as TO901317 are also commonly used to induce ABCA1 and A B C G 1 expression in a wide variety of cell types, including those typically used in cholesterol efflux assays 1 1 8 , 1 5 3 . Therefore, primary astrocytes were incubated for 24 hours with either AcLDL in serum free media, regular 56 A Cholesterol Efflux Primary Astrocytes media apoAl apoE3 ABCG1 Expression Primary Astrocytes murine human W T A B C G 1 + Figure 2-8: A B C G 1 overexpress ion in primary astrocytes does not promote cholesterol efflux to HDL. (A) Efflux from either wildtype or A B C G 1 BAC Tg primary astrocytes was measured for 4 hours in the presence of either plain media containing 0.2% BSA or supplemented with 15uxj/mL delipidated ApoAl, 15u.g/mL delipidated ApoE3, or 25(j.g/mL HDL. Data are pooled from 2 experiments done in triplicate (N=6) and are representative of 4 independent experiments. (B) ABCG1 Q R T - P C R of primary astrocytes following treatment with basal media, basal media + 1LIM TO901317 or 0 .2%BSA +40u.g/mL AcLDL. Human Abcgl mRNA levels were determined in ABCG1 BAC+ mice and murine Abcgl mRNA levels were measured in wildtype and ABCG1 BAC+ mice. maintenance media that contains 10% FBS or maintenance media supplemented with TO901317. These conditions were chosen to represent efflux, baseline culture and induced conditions respectively. Following treatment human and murine Abcgl mRNA levels were measured by Q R T - P C R . Compared to no treatment, AcLDL incubation increased the expression of murine Abcgl in wildtype astrocytes 5.7 ± 1.1-fold (p=0.013, N=3) and murine Abcgl in ABCG1 BAC Tg astrocytes 12.2 ± 2.7-fold (p=0.016, N=3) and human Abcgl in BAC Tg astrocytes 7.6 ± 0.7-fold (p=0.0009, N=3) (Figure 2-8B). LXR treated cells were observed to have similar levels of transcription induction to AcLDL treated cells (Figure 2-8B). These data confirm that AcLDL treatment increases Abcgl expression in both wildtype and transgenic astrocytes in our efflux experiments and does not explain why Abcgl overexpression did not increase cholesterol efflux to HDL. 2.2.4.3. Steady-state lipid levels are not changed in ABCG1 BAC Tg brain An independently-generated ABCG1 BAC Tg model has been reported to have reduced levels of cholesterol and phospholipids in peripheral tissues when fed a high fat d ie t 1 1 8 . This finding was considered significant as it confirmed the established in vitro function of ABCG1 in vivo. W e hypothesized that A B C G 1 function in the brain and periphery would be similar, and because Abcgl is highly expressed in brain, the lipid content of the brain may be reduced by A B C G 1 overexpression. Because brain cholesterol levels are reported to be unaffected by dietary intake, chow-fed mice were analysed. To determine if ABCG1 regulates the levels of cellular lipids in the brain as it is reported to in peripheral tissues, total brain lipid levels were measured. Lipid extracts were prepared from adult brains from ABCG1 Tg mice and littermate controls. Total lipid content was assessed following chloroform:methanol extraction and separation on a positive phase H P L C column. Lipid species were identified by retention time on the H P L C column and were quantified using a standard curve method. The levels of cholesterol esters, free cholesterol, ceramide, phosphatidylinositol, phosphatidylethanolamine, phosphatidylserine, 58 phosphatidylcholine and sphingomyelin for ABCG1 BAC Tg and wildtype brains are summarized in Table 2-2. These results show that ABCG1 overexpression does not alter the abundance of these lipids in brain. 2.2.4.4. Cholesterol precursors and metabolites are reduced in A B C G 1 Tg brains The results describes in section 2.2.4.1 failed to detect A B C G 1 facilitated efflux from primary astrocytes as expected. Total lipid levels measured in section 2.2.4.3 also failed to find any effect on total lipid levels in brain. However, overexpression of a human ABCG1 BAC had been reported to impact cholesterol homeostasis in peripheral tissues and expression of a murine BAC was reported to facilitate cholesterol efflux from murine macrophages 1 1 8 , 1 2 2 , suggesting that physiological overexpression of ABCG1 facilitates lipid transport in vivo. The brain, however, is reported to be more resistant to disturbances in cholesterol homeostasis than the periphery, as it is insulated from the fluctuating intake of dietary cholesterol by the BBB and has been shown to tightly regulate cholesterol biosynthesis to maintain physiological levels despite pharmacological or genetic manipulations 1 4 2 , 1 6 2 Additionally, 70-80% of the cholesterol in the brain is present in myelin sheaths that is very stable and may obscure relatively small changes in membrane cholesterol levels in the pool of cholesterol that is metabolically active in the brain. To overcome the unique characteristics of cholesterol homeostasis in the brain, we performed cholesterol metabolite analysis that had previously been reported to detect small changes in cholesterol synthesis in the brain 1 4 2 , 1 6 3 . Cholesterol metabolite analysis takes advantage of the fact that the brain synthesizes all of its cholesterol 1 6 4 and disposes of excess cholesterol through enzymatic conversion by Cyp46A1 to the oxysterol 24S-hydroxycholesterol that passively diffuses across the BBB 1 6 5 . Because of this closed-system physiology, a thorough analysis of cholesterol precursors and metabolites can detect small changes in cholesterol synthesis or elimination, in addition to measuring total cholesterol levels. Because Abcgl expression is highest in murine embryos and juveniles and declines into adulthood and cholesterol synthesis and because cholesterol turnover in the Total Brain Lipids WT ABCG1+ L i p i d (ug/mg) sd (ug/mg) sd p val Cholestero l Ester 0.43 0.02 0.36 0.06 ns Free Cholestero l 1.5 0.11 1.5 0.16 ns Free Fatty A c i d s 1.14 0.18 1.19 0.13 ns Ceramide 1.61 0.11 1.66 0.18 ns Phosphat idy lethanolamine 5.55 0.19 5.37 0.32 ns Phosphat idyl inosi tol 0.33 0.03 0.30 0.03 ns Phosphat idy lser ine 1.32 0.27 1.53 0.13 ns Phosphat idy lchol ine 5.25 0.33 5.23 0.40 ns Sphingomyel in 0.22 0.00 0.23 0.02 ns Lysophosphocho l ine 0.38 0.09 0.38 0.04 ns Total Phosphol ip id 13.0 0.73 13.0 0.75 ns Table 2-2: Total lipid levels in brain are not significantly affected by A B C G 1 overexpression. Total lipid extracts were prepared from adult A B C G 1 BAC transgenic and wildtype brains, quantified by H P L C and normalized to the protein content of the starting material. The presence of the ABCG1 BAC was not associated with any significant differences in lipid levels (p>0.05, N=4). mouse brain is estimated to peak at 14 days 1 6 6 , 1 6 7 , we hypothesized that cholesterol metabolism in the developing mouse brain would be highly sensitive to ABCG1 overexpression. To thus obtain the most sensitive measure possible, whole brains of 18 day and adult mice were subjected to cholesterol metabolite analysis. Whole brains were harvested, total lipid extracts were prepared by chloroform:methanol extraction and metabolite levels were determined by G C : M S . Compared to wildtype adult animals, cholesterol precursors in ABCG1 60 BAC transgenic brain were reduced. Notably, the precursor lathosterol was reduced by 21%(p=0.007, N=8), lanosterol was reduced by 14%(p=0.002 N=8) and desmosterol was reduced by 27% (p=0.009, N=8) and levels of cholesterol catabolite 24S-hydroxycholesterol was reduced by 15% (p=0.001, N=8) (Table 2-3). The plant-derived phytosterols Stigmasterol, Sitosterol, Sitostanol, Brassosterol, Campesterol and Campestanol and peripherally derived oxysterols, 7a-hydroxycholesterol, 27-hydroxycholesterol, were included in this experiment as controls for BBB integrity and plasma contamination. The levels of phytosterols and cholesterol catabolites in ABCG1 transgenic brain did not differ significantly from wildtype mice, confirming the validity of our observations (Table 2-3). Contrary to findings in the adult brain, no significant changes in cholesterol metabolite or sterols levels were observed between 18 day old wildtype and ABCG1 transgenic mice (Table 2-3). These results provide the first evidence that ABCG1 BAC is functional in vivo. Secondly these data suggest a previously unreported result that ABCG1 may modulate the levels of cholesterol precursors in vivo. That total cholesterol levels remain stable, despite a decrease in cholesterol precursors, may be explained by the decrease conversion of cholesterol to 24S-hydroxycholesterol. It was not clear from this experiment why ABCG1 influenced metabolite levels in the adult brain but not in very young brains. 2.2.4.5. Apolipoprotein E levels are unchanged in brain of ABCG1 transgenic mice Numerous lines of evidence suggest that apoE in the brain is central to the progression of AD neuropathology 6 6 , 1 6 8 , 1 6 9 . Furthermore, it has been reported that ABCG1 facilitates apoE secretion from cells and may influence total apoE levels 1 7 0 . To determine if A B C G 1 overexpression in vivo could impact apoE metabolism, we measured total and secreted apoE levels in brain of ABCG1 BAC mice. Whole brain levels of apoE in ABCG1 transgenic mice were unchanged compared to wildtype animals as determined by Western blot (Figure 2-9A). Secreted apoE in C S F was also measured by ELISA and was not significantly different between wildtype mice (1.34ug/mL ± 0.18) and ABCG1 BAC Tg mice (1.20u,g/mL ± 0.12) 61 18 day 100 day WT ABCG1 + WT ABCG1 + ng/mg sd ng/mg sd p val ng/mg sd ng/mg sd p val Total Cholesterol 5980 1643 5800 2944 ns 105000 9805 96000 9381 0.100 Lathosterol 398.8 42.6 435.4 61.8 ns 139 18.2 111 14.1 0.007 Lanosterol 32.6 2.4 34.1 3.9 ns 22.8 1.98 19.2 1.32 0.002 Desmosterol 1320 260 1449 282 ns 498 80.1 366 80.0 0.009 7a-hydroxy cholesterol 9.8 7.0 12.4 4.1 ns 7.22 2.44 8.11 4.37 ns 24S-hydroxy cholesterol 193.7 14.6 175.5 8.6 ns 293 28.6 251 25.0 0.001 27-hydroxy cholesterol 2.2 0.5 2.2 0.3 ns 2.88 0.59 2.78 0.62 ns Campesterol 20.7 2.7 24.1 4.5 ns 87.8 13.5 81.9 15.8 ns Campestanol n.d n.d n.d n.d ns 5.07 0.30 4.86 0.34 ns Stigmasterol 5.3 1.0 5.1 0.1 ns 5.45 0.25 5.38 0.45 ns Sitosterol 15.3 2.1 16.3 0.7 ns 27.7 3.15 27.0 4.39 ns Sitostanol 2.3 0.7 2.2 0.1 ns 2.37 0.20 2.21 0.23 ns Brassosterol 8.4 0.7 8.6 0.2 ns 9.85 0.57 9.40 0.64 ns Table 2-3: Cholesterol precursors and 24S-hydroxycholesterol levels are reduced in adult ABCG1 overexpressing brains but not very young mice. Whole lipid extracts were prepared from brains of 100 day old chow-fed male mice or 18 day old male mice prior to wean. Sterol levels were determined by G C : M S and normalized to dry weight of starting material, (ns = p>0.05) N=8/genotype for all ages. 62 Total Bra in A p o E B A C A B C G 1 G A P D H B S e c r e t e d A p o E 2.0' 1 H 1.0 3 0.5-0.0' Wt ABCG1+ Figure 2-9: A p o E levels are not affected by A B C G 1 overexpress ion. (A) Western blots prepared from whole brain lysates of ABCG1 BAC Tg mice and wildtype littermates were probed with antibodies specific ApoE (N=2). (B) ApoE levels in murine C S F collected from ABCG1 BAC Tg mice and wildtype littermates controls was determined by ELISA in triplicate (p>0.05, N=8). Both experiments are representative of 2 individual experiments. (p=0.5, N=8) (Figure 2-9B). These results indicate that ABCG1 overexpression does not influence secreted or total levels of apoE in brain. 2.2.5. A B C G 1 overexpress ion in v ivo, AB product ion and amyloid To determine whether ABCG1 overexpression could accelerate amyloid pathology in vivo, ABCG1 BAC Tg mice were crossed to the P D A P P mouse model of AD. P D A P P mice express human A P P bearing the Indiana (V717F) FAD mutation in hippocampal and cortical neurons 63 and develop primarily hippocampal amyloid deposits by 12 months of age . P D A P P mice are an established model for studying the impact of therapeutic compounds or genetic influences on amyloid formation 1 4 4 , 1 7 1 . To generate the experimental cohorts, A B C G 1 hemizygotes were influences on amyloid formation 1 4 4 , 1 7 1 . To generate the experimental cohorts, ABCG1 hemizygotes were crossed to human A P P homozygotes to produce A B C G 1 BAC transgenic progeny and wildtype controls that were hemizygous for the P D A P P A P P transgene. A S production and amyloid formation were then measured in our experimental cohorts. Because newly generated AB is PBS-soluble and has a halflife of 2 hours in brain 1 7 2 , the P B S -soluble AB content of hippocampus was measured to estimate AB production. Amyloid is primarily composed of PBS-insoluble AB fibrils, so the AB content of guanidine soluble hippocampus extracts was used to estimate amyloid load. Although P B S extraction removes soluble AB species, the mass of fibrillar AB can be many orders of magnitude higher than the mass of soluble AB and can easily contaminate soluble AB measurements. To reduce the effects of insoluble amyloid on soluble AB measurements, mice were analyzed before amyloid deposition at 3 months of age and at 13 months of age, after amyloid deposits are reported in P D A P P brain. 2.2.5.1. A B C G 1 overexpression does not influence amyloid deposition in vivo AB species were serially extracted from hippocampi, first by P B S to remove soluble species, and secondly by guanidine hydrochloride to solubilize fibrillar amyloid into AB monomers. The levels of AB40 and 42 in PBS-soluble and guanidine-insoluble extracts were determined by ELISA and normalized to wet weight of starting material. At 3 months of age, P B S soluble AB42 levels in hippocampus of 3 monthrold A B C G 1 transgenic mice were increased 1.19-fold (p=0.039, N=18) compared to littermate controls (Figure 2-1 OB). However, PBS-soluble AB40 levels were not significantly affected by A B C G 1 overexpression (p= 0.30, N=18) (Figure 2-1 OA). Furthermore, in 13 month old samples no significant differences were observed in PBS-soluble AB40 (p= 0.71, N=22) (Figure 2-10C) or AB42 (p=0.90, N=24) (Figure 64 A A640 Hippocampus 3 months B AB42 E «H WT ABCG1+ WT ABCG1+ .Il WT ABCG1+ WT ABCG1 + PBS GuHCI PBS GuHCI Hippocampus 13 months AB40 D AS42 ii .11 ni 4000-1 S 25-r 20-15-10-5- 1 PBS GuHCI PBS WT ABCG1+ GuHCI Figure 2-10: So lub le and insoluble human AB levels in the h ippocampus of P D A P P mice either wildtype or overexpressing A B C G L (A, B) Hippocampal AB levels from 3 month old mice prior to amyloid formation (N=18). (C,D) Hippocampal AB levels in 13 month old animals with amyloid deposits (N=25). Samples were serially extracted first in P B S to measure soluble AB species followed by Guanidine hydrocholoride (GuHCI) extraction to solubilize fibrillar amyloid. AB species were measured by human specific ELISAs. (A,C) AB40 and (B,D) AB42 . All results are p>0.05 unless noted. * = p<0.05 2-10D). Additionally, no significant changes in guanidine-soluble AB40 was observed at 3 months (p=.99, N=18) or 13 months (p= 0.71, N=22) (Figure 2-1 OA), nor did ABCG1 overexpression impact guanidine-soluble levels of AB42 at 3 months of age (p=0.21, N=18) (Figure 2-1 OB) or 13 months of age (p=0.87, N=23) (Figure 2-10D). These results show that 13 month old animals have ~50-fold more insoluble AB40 and 900-fold more insoluble AB42 than 65 3 month old animals, confirming the presence of amyloid at 13 months as expected. These data do not find a correlation between ABCG1 genotype and PBS-soluble or guanidine-soluble A S species with the exception of a 19% (p=0.039, N=18) increase of soluble Aft42 in ABCG1 transgenic mice compared to controls at 3 months of age. Given the small magnitude of the change at 3 months and the failure to observe a corresponding increase in insoluble amyloid levels at 13 months, these data suggest that ABCG1 overexpression does not substantially impact amyloid burden in the P D A P P mouse model of A D at 13 months of age. 2.2.5.2. Total A P P and C99 levels are not affected by ABCG1 overexpression To determine if ABCG1 overexpression in vivo increased A P P processing by ft and a-secretase as observed in H E K : A P P cells (Section 2.2.1.2, 2.2.1.3 and Figure 2-1,2-2), C99 and C83 levels were measured in brain of ABCG1 BAC Tg mice. RIPA buffer tissue homogenates were prepared from 3 and 13 month old cortex and hippocampus of ABCG1 BAC Tg / P D A P P cohort described in section 2.2.5 and C83 and C99 levels were measured by Western blot. Because P D A P P mice are known to have variable levels of A P P expression that could directly impact CTF levels, CTF levels were normalized to A P P protein. In contrast to our in vitro results, A P P and C99 levels in cortex and hippocampus were not significantly affected by ABCG1 over-expression at either 3 or 13 months of age (Figure 2-11A-D). Because the C83 band detected in these blots may be comigrating with other smaller CTF species (data not shown), an accurate determination of C83 levels cannot be made from this experiment. These results show that C99 production is not increased by the overexpression of ABCG1 in vivo. These findings largely confirm previous findings in section 2.2.5.1 that suggest Aft production and amyloid formation are not substantially affected by the overexpression of A B C G 1 . 2.2.5.3. Murine Aft levels are not affected by ABCG1 protein levels To obtain an independent measure of Aft production in brain, the levels of endogenous murine Aft were also determined in ABCG1 BAC Tg mice in the absence of the P D A P P A P P 66 A P P C99 G A P D H 3 month Hippocampus W T A B C G 1 + ctn WHi flflit mmt l^K V fli 2.0-1.5-1.0-0.5-0.O A P P C 9 9 / A P P WT ABCG1+ WT ABCG1+ A B C G 1 A P P C99-G A P D H 3 month Cortex - _ + . + . + + » i A P P C 9 9 / A P P 1.5-1.0-0.5-0.O III WT ABCG1 + WT ABCG1+ C 13 month Hippocampus D 13 month Cortex WT ABCG1+ WT ABCG1 + Figure 2-11: C99 and total A P P levels are not affected by A B C G 1 ove rexp ress i on . Tissue homogenates from hippocampus and cortex of P D A P P crossed A B C G 1 B A C Tg mice were prepared by tris-tricine S D S - P A G E for Western blotting. Membranes were probed with antibodies specific to the C-terminus of human A P P and the results were quantified by densitometry from duplicate gels for comparison. C99 levels are expressed as a ratio of total A P P . transgene. Murine A P P is cleaved through similar mechanisms as human A P P , however, because murine AB does not form amyloid deposits 1 7 3 , 1 7 4 the levels of murine AB may more accurately reflect production of soluble AB species than the level of human AB observed in P D A P P mice. Furthermore, to thoroughly understand the role of ABGC1 in AB production, both ABCG1 deficient and ABCG1 Tg mice were evaluated in this experiment. To measure soluble murine AB levels, P B S homogenates of total brain from A B C G 1 B A C Tg mice and wildtype littermates as well as A B C G 1 +/+, +/- and -/- littermates were prepared and murine AB40 and 42 levels were determined by ELISA. Compared to wildtype littermates, ABCG1 BAC transgenice mice did not have significantly different levels of murine AB40 (p=0.25, N=6 wildtype, N=4 Tg) or murine AB42 p=0.34, N=6wildtype, N=4 Tg) (Figure 2-12A). Furthermore, compared to wildtype (+/+) littermates, we did not observe statistically significant changes in either murine AB40 levels in ABCG1 hemizygotes (+/-) (P<0.40, N=3) or A B C G 1 deficient mice (-/-) (p<0.22, N=3) or changes in murine AB42 in hemizygotes (p=0.27, N=3) or A B C G 1 deficient mice (p=0.18, N=3) (Figure 2-12B). These data suggest that neither A B C G 1 overexpression, nor deficiency, has a significant impact on murine AB production in vivo. 68 M u r i n e A f t A 10.0-1 AG 40 AU 42 ns 7.5H A B C G 1 WT BAC + WT BAC + B 10.0-1 AB 40 A B 42 A B C G 1 +/+ +/- -i 1 i r +/+ +/. ./-«n n ns Figure 2-12: A B C G 1 overexpress ion or def ic iency do not affect murine AB levels in brain. (A) Whole brains of ABCG1 BAC Tg mice and wildtype littermate controls (WT N= 6, Tg N=4) or were homogenized in P B S and the level of P B S soluble murine AB40 and AB42 were determined by ELISA. (B) Murine AB levels in brain of wildtype(+/+), ABCG1 hemizygous(+/-) and ABCG1 deficient(-/-) littermates (N=3). ns = p>0.05 3. DISCUSSION AND FUTURE DIRECTIONS 3.1. Introduction The overall goal of this research was to test the hypothesis that A B C G 1 plays an important role in the pathogenesis of Alzheimer's Disease. The research hypothesis is driven by the convergent observations that 1: Membrane cholesterol levels modulate AG. production in neurons, a key event in amyloid formation. 2: The cholesterol transporter ABCG1 is highly expressed in brain and regulates cholesterol distribution and transport from cells suggesting that overexpression or deficiency of ABCG1 might impact the production of AG through a cholesterol-related mechanism. 3: Because ABCG1 overexpression occurs in DS, concomitant with accelerated amyloid deposition, we hypothesize that A B C G 1 may be one of the genes on chromosome 21 capable of promoting amyloid pathology when overexpressed. Because of the presence of amyloid in individuals with DS, our experiments focused primarily on testing the hypothesis that ABCG1 overexpression would promote AG production from cells and promote amyloid formation in brain. The hypothesis was tested both in vitro using a cell-based assay and in vivo by crossing a murine model of AD with an ABCG1 overexpressing mouse developed in this thesis. Given that the ABCG1 mouse was a novel animal model the expression of the human ABCG1 transgene was characterized and lipid-related phenotypes were also assessed. Finally, the impact of ABCG1 deficiency on AG production was explored. 3.2. Discussion 3.2.1. A B C G 1 overexpression and AG production in vitro Using a cell-based assay, we showed that ABCG1 overexpressing cells produced more AG peptides of 40 and 42 amino acids in length, suggesting that an increase in G-secretase-mediated processing of A P P may be the most likely underlying mechanism. The fact that only efflux-active A B C G 1 was able to promote AG production suggests that the cholesterol transporting-activity of ABCG1 is essential for this process. Given a recent report that shows 70 A B C G 1 overexpression increases the abundance of free cholesterol in the plasma membrane 1 1 7 , this result is consistent with previous studies linking increased membrane cholesterol content with increased AB production in vitro 9 1 . Because an increase in AB production can directly correlate to an increased amyloid burden in humans 3 3 , our results are consistent with the hypothesis that ABCG1 overexpression accelerates amyloid formation. Unexpectedly, an increase in a-secretase activity was also observed. Because a and B secretase activity occur primarily at the cell surface, these phenomena are likely explained by the observation that A B C G 1 overexpression increases the amount of A P P present at the cell surface 1 2 7 . These experiments demonstrate that A B C G 1 overexpression in HEK cells increases the processing of A P P by B-secretase into C99 and a corresponding increase in AB production. These findings are in contrast with a recently published report that found A B C G 1 overexpression reduced the secretion of AB in vi t ro 1 7 5 . Both studies employ transient transfection assays however, the study by Kim and colleagues does not normalize secreted AB levels to control for cell number nor did the study measure CTF levels to verify increased A P P processing. Given these shortcomings, Kim and authors cannot conclude that A B C G 1 decreases AB production. Furthermore, Kim and colleagues used C H O - A P P cells as opposed to H E K - A P P cells that also may account for, but not explain, the observed differences in experimental outcomes. One must use caution when interpreting these in vitro results however, as similar cell culture experiments analyzing the transporter ABCA1 reported that expression levels correlate with AB production, either positively or negatively depending on the study 1 7 6 1 7 7 , 1 7 8 . Since the completion of the cell culture studies however, three further reports have been conducted that show ABCA1 deficiency in fact has no impact on AB production in vivo 1 6 9 , 1 7 9 , 1 8 0 . These contradictory reports highlight the need to validate any results based on cell-models in vivo. 71 3.2.2. The A B C G 1 BAC transgenic mouse The ABCG1 overexpressing mouse carries a stably integrated BAC containing an insert of genomic DNA corresponding to the human Abcgl gene and flanking regulatory sequence. The use of a BAC is necessitated by the size of the 141 kb transgene that is too large to be contained in traditional bacterial vectors. The presence of cis-regulatory elements and preservation of the exon-intronic structure should ensure the natural protein isoforms are expressed in a physiologically regulated way, provided that the transcriptional machinery of the host organism is capable of recognizing and accurately binding the transgenic DNA. A very successful precedent for the use of genomic DNA is the generation of the ABCA1 transgenic mouse, which worked exceptionally well for the analysis of the effect of excess ABCA1 on HDL metabolism and atherosclerosis 1 0 5 , 1 5 5 , as well as reports showing that several BAC transgenes were expressed in a coordinated way resembling the physiologically regulated gene, reviewed in 1 5 4 . The caveats of the BAC approach are 1: high-degrees of overexpression can be difficult to obtain, as gene over-expression can lead to a compensatory down-regulation of the transgene and 2: The pattern of transgene expression will depend to some degree on host-specific expression of transcription factors. For instance, in the case of the A B C G 1 BAC, the transgene is expected to be expressed in part through the binding of murine L X R / R X R and PPAR -y transcription factors that may not have identical expression patterns in mice and h u m a n s 1 0 9 ' 1 1 2 . Because the goal of these experiments was to determine the impact of increased Abcgl gene dose on amyloid formation, we wanted to ensure the A B C G 1 B A C was expressed in a physiological relevant way and that robust protein overexpression was achieved to facilitate detection of a phenotype. Physiological regulation of transgene expression was confirmed by 1: A survey of different tissues revealing that human and murine Abcgl mRNA levels were similar in most tissues. 2: The observation that human and murine Abcgl mRNA expression in primary neurons was induced to a similar degree by a synthetic LXR agonist and 3: Cholesterol loading of primary astrocytes induced the expression of both species of Abcgl mRNAs to 72 similar degrees. Western blots that show ABCG1 protein is overexpressed 3-6-fold in peripheral tissues and brain regions of A B C G 1 BAC Tg mice, confirming that human Abcgl mRNA is translated into excess protein. In comparison, individuals with DS are reported to overexpress Abcgl mRNA 1.5-2-fold and 2-fold at the protein l e v e l 1 2 7 1 1 8 1 . We also observed that the ABCG1 BAC remains robustly expressed in A P P expressing mice with amyloid, which is an observation in contrast to ABCA1 BAC transgenic mice that subtly overexpress ABCA1 in brain in the absence of human A P P , but lose this overexpression in the presence of amyloid 1 5 8 . Taken together these experiments confirmed that the human ABCG1 BAC transgenic mouse is a good model for studying the in vivo role of A B C G 1 . 3.2.3. Human and murine Abcgl expression A survey of murine Abcgl expression was performed in 12 separate tissues. A comparison of the murine mRNA expression data with a previously published report of human Abcgl mRNA abundance suggests that Abcgl expression differs between mice and humans 1 0 7 . Specifically, the abundance of Abcgl mRNA in human liver was similar to most tissues, whereas we observed that Abcgl was poorly expressed in murine liver. This observation, however, may reflect differences between the average human diet and the chow fed mice in this study as cholesterol transporter expression is reported to be greatly increased in hepatic cells of mice fed a diet high in cholesterol and fat 1 8 2. Additionally, in human brain, Abcgl expression was reported to be lower than kidney, spleen, liver and lung, whereas we observed that murine brain had one of the highest levels of Abcgl expression. These comparisons suggest that humans and mice may express Abcgl at different relative levels in specific tissues. 3.2.4. Cell-specific expression of Abcgl in brain The knockout strategy employed to generate the Abcgl knockout mice places a LacZ reporter construct under the control of Abcgl regulatory sequence. Staining for the presence 73 of the reporter in Abcgl +/- brain sections clearly indicated that Abcgl was expressed in hippocampal layers primarily composed of neurons and cortical layers. Diffuse staining was also observed throughout the rest of the brain section shown. These findings are consistent with previous reports of Abcgl expression in neuronal cell lines and embryonic neurons 1 0 6 ' 1 7 5 . In our experiments, LacZ staining appears to be most intense in hippocampal neurons, however, care must be taken to discriminate the intensity of staining from density of stained cells as densely packed cells may create the illusion of high expression. Astrocytes are reported to express Abcgl153 and may represent some or all of the non-hippocampal staining however, no reliable comparison can be made between the level of neuronal and non-neuronal Abcgl expression from these data. 3.2.5. The in vivo function of ABCG1 3.2.5.1. Cholesterol efflux: The established function of A B C G 1 is to facilitate cholesterol efflux to lipidated acceptors 1 0 6 , 1 0 7 , ne, n s T o ^ 1 ^ 3 ^ the functionality of the ABCG1 BAC in a cell type known to play an important role in cholesterol homeostasis of the brain, we measured the cholesterol efflux capacity of primary astrocytes. In these experiments ABCA1-specif ic controls confirmed ABCA1-mediated efflux, suggesting that the experimental conditions employed were sufficient to observe the export of cholesterol from cells 1 5 5 . However, in contrast to previous reports that suggest ABCG1 is more important than ABCA1 for astrocyte cholesterol efflux 1 5 3 , excess A B C G 1 did not promote a statistically significant increase in cholesterol efflux to HDL. This result suggests three possible explanations. Firstly, in primary cells expressing endogenous A B C G 1 , cholesterol efflux to HDL may not be a sensitive measure of A B C G 1 function or ABCG1 may not be a rate-limiting step in cholesterol efflux. For example, complete knockout of ABCG1 protein in macrophages suggests that A B C G 1 only accounts for 30-50% of the cholesterol efflux to HDL overa l l 1 0 7 . Furthermore, very high levels of ABCG1 overexpression in immortalized cell types that lack 74 endogenous ABCG1 typically produce a 30-40% increase in cholesterol efflux to HDL suggesting that the relatively modest 3-5-fold overexpression of A B C G 1 obtained in this experiment may not be easily detectable, especially in the context of endogenous ABCG1 1 expression . The second explanation for the absence of efflux activity in primary astrocytes may reflect the in vivo function of A B C G 1 . The vast majority of overexpression studies, including the HEK studies described here, use a human cDNA corresponding to an mRNA isoform containing exons 5,7, 11-23. This isoform is highly induced by oxysterol treatment, however, not all Abcgl isoforms have been conclusively shown to promote cholesterol efflux 1 0 6 . As the Abcgl isoforms expressed in brain have not been described here or elsewhere, it is possible that the Abcgl isoforms expressed from the physiologically regulated ABCG1 BAC are either less effective or ineffective at promoting cholesterol efflux from astrocytes. It is unlikely that the lack of A B C G 1 mediated efflux in these experiments is entirely related to the use of astrocytes as the A B C G 1 BAC also fails to promote cholesterol efflux from primary macrophages that are known to utilize A B C G 1 mediated pathways for cholesterol efflux (Burgess BL., et al. manuscript submitted to J . Lipid Res; Edwards PA., Principal Investigator, personal communicat ion) 1 0 7 . The third explanation for the failure to observe ABCG1 facilitated efflux from either ABCG1 overexpressing astrocytes or macrophages (Burgess BL., et al. manuscript submitted to J . Lipid Res) may represent differences or incompatibilities between murine and human A B C G L This is supported by the observation that macrophages harvested from BAC Tg mice generated with murine Abcgl have elevated cholesterol efflux to HDL 1 2 2 . However two groups, including ours, that generated BAC Tg mice with human Abcgl failed to observe an increase of cholesterol efflux to HDL from macrophages or astrocytes. This suggests that human A B C G 1 may not function properly in mice (Burgess BL., et al. manuscript submitted to J . Lipid Res; Edwards PA., Principal Investigator, personal communication). Although murine and human A B C G 1 are 97-98% identical at the protein level depending on the particular protein isoform, the 2-3% difference accounts for ~10-25 non-identical amino acids between the species. The incompatibility could be a result of improper dimerization of murine and human ABCG1 or improper trafficking of human A B C G 1 in murine cells. For example, single missense mutations in A B C G 5 or A B C G 8 are reported to prevent the assembly of functional G5/G8 heterodimers and impair proper sub-cellular trafficking 1 8 4 , suggesting that small differences between human and murine ABCG1 may prevent, or impair the assembly of functional human/murine A B C G 1 dimers. Although human Abcgl cDNAs have been shown to facilitate efflux in murine cell types, these experiments are typically conducted under conditions where human ABCG1 is greatly overexpressed and would be largely present in human/human homodimers with human/murine dimers forming a minor component. As we do observe a gain of function in A B C G 1 overexpressing mice with respect to the levels of cholesterol precursors, that activity may be preserved in murine/human ABCG1 dimers, or reflect the contribution of human/human dimers. 3.2.5.2. Brain lipid levels As the A B C G 1 protein is reported to protect against diet-induced accumulation of cholesterol and phospholipids in lung and liver, we investigated the lipid content of ABCG1 overexpressing brain 1 1 8 . Our experiments did not involve any dietary manipulation and the BBB would prevent dietary lipids from accumulating in the brain. However, we hypothesized that because murine Abcgl expression in brain is 2-6-fold higher than in lung and liver respectively, the steady-state level of lipids in ABCG1 overexpressing brain may be reduced. Despite the high level of expression in brain, the total levels of cholesterol ester, free cholesterol, free-fatty acids, sphingomyelin, ceramide and all phospholipids species examined were not changed, despite reports that ABCG1 can facilitate efflux of cholesterol, phosphatidylcholine and sphingomyelin from c e l l s 1 0 7 , 1 8 5 . The possibility remains that the cellular distribution within the brain of one or more of these lipids is affected. 76 3.2.5.3. Cholesterol metabolites in brain of ABCG1 BAC transgenic mice We observed a reduction in cholesterol precursors and 24S-hydroxycholesterol in A B C G 1 overexpressing adult mice, however, total cholesterol levels were not significantly changed. Specifically, a reduction in lanosterol, desmosterol and lathosterol precursors and 24S-hydroxycholesterol was observed. The cholesterol biosynthetic pathway is reviewed in Figure 1-2. Intriguingly and consistent with this finding, Abcgl -I- mice are reported to accumulate desmosterol and lathosterol in brain (Wang N., Associate Research Scientist, personal communication). One interpretation of these results is that the activity of ABCG1 suppresses cholesterol biosynthesis through a mechanism related to the intracellular distribution of cholesterol. HMGR transcription is regulated in part by the sterol-sensing protein S C A P in the ER 1 8 6 , suggesting that A B C G 1 could regulate cholesterol biosynthesis by controlling the cholesterol content of the ER membrane. Alternatively, as small quantities of non-hydroxylated cholesterol are capable of being cleared across the BBB into the peripheral circulation through an unknown mechanism 5 0 , A B C G 1 may promote the clearance of one or more cholesterol precursors from the brain. In both of these models cholesterol levels in the brain could remain unchanged through a compensatory decrease in enzymatic conversion of cholesterol into 24S-hydroxycholesterol for elimination, which is supported by the data. The failure to observe changes in cholesterol precursor or metabolite levels in very young mice may provide additional insights into the underlying mechanism. In very young mice cholesterol synthesis is occurring rapidly, and although the relative contributions of glia and neurons are not well understood, it has been suggested that astrocytes may play a smaller role in cholesterol synthesis in the developing brain than in the adult brain 1 8 7 , 1 8 8 . Other studies have reported that the majority of cholesterol is synthesized through the desmosterol pathway in the developing brain whereas the lathosterol branch is primarily used in the adult brain 1 8 9 . These studies suggest that age-related differences in cholesterol metabolism or age related differences in cell-specific expression patterns of Abcgl may explain this discrepancy. 77 3.2.6. A B C G 1 and ApoE The level of total-brain apoE and secreted apoE in C S F were not significantly changed by overexpression of the A B C G 1 BAC. These findings fail to support one study that reported s iRNA inhibition of ABCG1 reduced apoE secretion from human monocyte derived macrophages 1 7 0 . It is not clear from our experiments whether the discrepancy reflects differences between macrophages and glia, or the effect of overexpression versus knockdown. 3.2.7. A B C G 1 and A D 3.2.7.1. A B C G 1 and amyloid To test whether ABCG1 overexpression could influence the production of AB and produce a corresponding increase in amyloid formation, the A B C G 1 overexpressing mice were crossed to the P D A P P mouse model of A D that express human A P P and develop amyloid deposits. We measured PBS-soluble AB to estimate newly synthesized AB and the P B S -insoluble-guanidine-soluble pool to measure AB deposited as amyloid fibrils 1 9 0 . We expected to observe an increase in AB production in ABCG1 overexpressing mice and a corresponding increase in amyloid in older mice consistent with our hypothesis and our in vitro findings. However, our results show that A B C G 1 overexpression was only associated with a 19% increase in soluble AB42 levels and this was specifically observed in amyloid-free hippocampus. No significant effects were seen on AB40. Furthermore, despite the increase in soluble AB42 at three months of age, a corresponding increase in amyloid AB42 was not observed in older animals, suggesting that the small increase is not sufficient to affect murine amyloid formation over this timeframe. In our HEK experiments, we observed that the increase in AB production was coupled with an increase of A P P processing into C99 by B-secretase. However, in vivo, ABCG1 overexpression did significantly impact C99 levels in cortex or hippocampus. This would suggest that subtle changes in C99 levels were below our sensitivity threshold, or that the increase in AB42 in young mice may occur through an alternate pathway, possibly involving y-secretase. Because we observed no change in CTF levels, these results 78 suggest that overexpression of ABCG1 in the brain is not sufficient to increase AU production, indicating that much more A B C G 1 may be required to affect AU production as in the experiments involving H E K - A P P cells. It should be noted that difficulties in detecting changes in steady state AU levels could also be explained by increased AU clearance matching increased AU production, resulting in no net change. Taken together, ABCG1 overexpression may slightly increase AU42 production through a non-U-secretase mediated pathway, however these effects are not sufficient to impact amyloid levels in vivo. 3.2.7.2. ABCG1 and Murine AU Murine AU levels were measured in the absence of human A P P as an alternative measure of AU production. As murine A P P lacks FAD mutations, does not form amyloid, and is endogenously regulated in contrast to P D A P P mice that express A P P under the PDGF-U promoter, murine AU levels may provide a more accurate measure of AU production. A trend towards lower murine AU levels was observed in ABCG1 deficient mice that may achieve statistical significance in an experiment comprised of more than 3 animals per group. The relevance of this effect on amyloid formation cannot be determined by this experiment as murine AU does not accumulate as amyloid over the lifespan of the mouse. Consistent with the overall findings from the previous experiment in P D A P P mice, ABCG1 overexpression failed to make a significant impact on murine AU levels. These experiments further suggest that A B C G 1 does not significantly impact AU production. 3.2.8. Conc lus ion : The funct ion of A B C G 1 in brain This research described the generation of a novel A B C G 1 overexpressing murine research model. The ABCG1 BAC transgenic mouse possesses human genomic Abcgl under the control of endogenous cis-regulatory sequence and preserves intronic/exonic structure. The transgene was expressed in a similar tissue specific pattern as murine Abcgl, which we have shown may differ from Abcgl expression in humans. Additionally, this research provides the first evidence that Abcgl is well expressed in adult hippocampal neurons and confirms A B C G 1 expression in cultured embryonic neurons and post-natal astrocytes. This research produced novel data suggesting that ABCG1 may modulate cholesterol precursor and metabolite levels in brain, although the underlying mechanism has yet to be fully determined. Furthermore, A B C G 1 overexpression did not facilitate cholesterol efflux from astrocytes, suggesting that physiological overexpression of ABCG1 may have different consequences on cholesterol metabolism compared to high-level constitutive overexpression in cell-lines. A B C G 1 overexpression also did not influence steady state levels of lipids in brain. In conclusion, this thesis describes the generation of a novel transgenic mouse overexpressing human A B C G 1 and suggests a novel role for ABCG1 in cholesterol precursor metabolism in the brain. 3.2.9. Conc lus ion : The role of A B C G 1 in A D This research was based upon the proposed ability of A B C G 1 to influence cellular cholesterol homeostasis in neurons and the correlation between neuronal cholesterol levels and AU production. Because we observed reduced 24S-hydroxycholesterol levels in ABCG1 overexpressing brain and the production of 24S-hydroxycholesterol is reported to occur specifically in neurons, it can be inferred that A B C G 1 overexpression influences cholesterol metabolism to some extent in neurons. However, the observations that murine AU levels are not affected by ABCG1 overexpression or deficiency, and that A B C G 1 overexpression does not affect amyloid deposition or U-secretase cleavage of A P P , provides compelling evidence that ABCG1 protein levels do not substantially affect AU production or amyloid pathology. This research concludes that ABCG1 is unlikely to participate in the rapid onset of AD pathology in DS. 80 3.3. Future Directions The conclusion of this thesis is that A B C G 1 overexpression is unlikely to play an important role in amyloid formation in AD, however, its potential influence on cholesterol metabolism remains important to AD, brain injury and other forms of neurodegeneration 6 3 , warranting further study of the role of ABCG1 and other cholesterol related genes in the brain. 3.3.1. Determine how ABCG1 regulates cholesterol metabolite levels The underlying mechanism through which ABCG1 overexpression reduces cholesterol precursors and 24S-hydroxycholesterol was not determined in this thesis. As cholesterol synthesis is primarily regulated through the abundance of HMGR protein, Western blots of brain tissue would determine whether ABCG1 modulates H M G R activity and is an appropriate next step for this line of investigation. Western blots from both mice and individuals with DS samples could be easily analyzed to confirm and extend the findings to humans. 3.3.2. Determine the cell-specific patterns of expression in brain The primary cell types that express Abcgl in the adult brain were not rigorously explored in this thesis. Immunofluorescence studies that co-label neurons, astrocytes or microglia in addition to ABCG1 would allow the cell-specific patterns of ABCG1 expression to be determined, however this was impossible prior to obtaining the A B C G 1 knockout mice from Deltagen as all commercially available antibodies to A B C G 1 perform poorly in immunofluorescence applications. In this thesis the LacZ reporter gene was detected by utilizing the catalytic activity of the B-galactosidase enzyme to convert X-gal to a colored precipitate, however, immunological detection of the LacZ reporter gene is also possible. By co-staining with monoclonal antibodies to the LacZ reporter gene and cell-type specific antibodies the regional and cell-specific patterns of A B C G 1 expression could be determined in detail. 81 3.3.3. Determine the relevant Abcgl isoforms in brain and their function This thesis raised questions about the in vivo function of A B C G 1 in the brain. As not all A B C G 1 protein isoforms have been shown to induce cholesterol efflux and this study failed to observe A B C G 1 mediated efflux from astrocytes, this discrepancy may be explained by determining the major Abcgl isoforms in brain. It is likely that neurons and glia express different isoforms and the relevant isoforms may vary throughout development or in response to oxysterols and PPAR -y ligands. The transgenic ABCG1 model developed in this study offers the unique potential to study ABCG1 mRNA isoform expression during development and in response to treatment by various agonists, using both in vivo and primary cell culture approaches. Once cloned, these ABCG1 mRNA isoforms, which are predicted to differ primarily at the N-terminus, could be screened for an ability to promote efflux or redistribution of cholesterol and other lipids reported to be transported by A B C G 1 such as sphingomyelin and phosphatidylcholine. It is possible that alternative splicing could confer alternative substrate specificity, or could affect intracellular-trafficking of A B C G L 3.3.4. Test the functionality of the human A B C G 1 BAC in A B C G 1 knockout mice The hypothesis suggested in section 3.2.5.1 that human A B C G 1 may be functionally impaired in mice as a result of improper dimerization with murine A B C G 1 , or incorrect intracellular trafficking could be tested. One approach to testing the hypothesis involves backcrossing the ABCG1 BAC onto the A B C G 1 - / - mice to determine if the A B C G 1 BAC could rescue the impaired efflux-phenotype reported in ABCG1 knockout mice. A comparison of A B C G 1 BAC hemizygotes and homozygotes in conjunction with Abcgl -/-, -/+ and +/+ genotypes would determine whether human ABCG1 BAC is functional in mice, both in the presence and absence of murine A B C G 1 . 82 REFERENCES: 1. Price DL, Tanzi RE, Borchelt DR, Sisodia SS . Alzheimer's disease: Genetic studies and transgenic models. Annu Rev Genet. 1998;32:461-493. 2. Larson EB, Shadlen MF, Wang L, et al. Survival after initial diagnosis of alzheimer disease. Ann Intern Med. 2004;140:501-509. 3. Hebert LE, Scherr PA, Bienias JL, Bennett DA, Evans DA. Alzheimer disease in the US population: Prevalence estimates using the 2000 census. Arch Neurol. 2003;60:1119-1122. 4. Li Y J , Scott WK, Hedges DJ, et al. Age at onset in two common neurodegenerative diseases is genetically controlled. Am d Hum Genet. 2002;70:985-993. 5. Hull M, Berger M, Heneka M. Disease-modifying therapies in alzheimer's disease: How far have we come? Drugs. 2006;66:2075-2093. 6. Birks J . Cholinesterase inhibitors for alzheimer's disease. Cochrane Database Syst Rev. 2006;(1):CD005593. 7. Brookmeyer R, Gray S, Kawas C. Projections of alzheimer's disease in the united states and the public health impact of delaying disease onset. Am J Public Health. 1998;88:1337-1342. 8. Schneider LS, Dagerman KS. Psychosis of alzheimer's disease: Clinical characteristics and history, d Psychiatr Res. 2004;38:105-111. 9. McKhann Gea. Criteria for the clinical diagnosis of alzheimer's disease, excerpts from the NINCDS-ADRDA work group report, d Am Geriatr Soc. 1985;33:2-3. 10. Bigler ED. The clinical significance of cerebral atrophy in dementia. Arch Clin Neuropsychol. 1987;2:177-190. 83 11. Klunk WE, Engler H, Nordberg A, et al. Imaging brain amyloid in alzheimer's disease with Pittsburgh compound-B. Ann Neurol. 2004;55:306-319. 12. Fagan A M , Mintun MA, Mach RH, et al. Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neurol. 2006;59:512-519. 13. Grundke-lqbal I, Iqbal K, Quinlan M, Tung Y C , Zaidi MS, Wisniewski HM. Microtubule-associated protein tau. A component of alzheimer paired helical filaments. J Biol Chem. 1986;261:6084-6089. 14. Grundke-lqbal I, Iqbal K, Tung YC, Quinlan M, Wisniewski HM, Binder LI. Abnormal phosphorylation of the microtubule-associated protein tau (tau) in alzheimer cytoskeletal pathology. Proc Natl Acad Sci USA. 1986;83:4913-4917. 15. Goedert M, Jakes R. Mutations causing neurodegenerative tauopathies. Biochim Biophys Acta. 2005;1739:240-250. 16. Poorkaj P, Bird TD, Wijsman E, et al. Tau is a candidate gene for chromosome 17 frontotemporal dementia. Ann Neurol. 1998;43:815-825. 17. Dickson DW, Crystal HA, Bevona C, Honer W, Vincent I, Davies P. Correlations of synaptic and pathological markers with cognition of the elderly. Neurobiol Aging. 1995;16:285-98; discussion 298-304. 18. Cras. Extracellular neurofibrillary tangles reflect neuronal loss and provide further evidence of extensive protein cross-linking in alzheimer disease. Acta Neuropathol. 1995;89:291. 19. Glenner G G , Wong CW. Alzheimer's disease: Initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun. 1984;120:885-890. 20. Gellermann G P , Appel TR, Tannert A, et al. Raft lipids as common components of human extracellular amyloid fibrils. Proc Natl Acad Sci USA. 2005;102:6297-6302. 84 21. Namba Y, Tomonaga M, Kawasaki H, Otomo E, Ikeda K. Apolipoprotein E immunoreactivity in cerebral amyloid deposits and neurofibrillary tangles in alzheimer's disease and kuru plaque amyloid in creutzfeldt-jakob disease. Brain Res. 1991;541:163-166. 22. Choi-Miura NH, lhara Y, Fukuchi K, et al. SP-40,40 is a constituent of alzheimer's amyloid. Acta Neuropathol (Bert). 1992;83:260-264. 23. Iwatsubo T, Odaka A, Suzuki N, Mizusawa H, Nukina N, lhara Y. Visualization of A beta 42(43) and A beta 40 in senile plaques with end-specific A beta monoclonals: Evidence that an initially deposited species is A beta 42(43). Neuron. 1994;13:45-53. 24. Lorenzo A, Yankner BA. Beta-amyloid neurotoxicity requires fibril formation and is inhibited by congo red. Proc Natl Acad Sci US A. 1994;91:12243-12247. 25. Walsh DM, Klyubin I, Fadeeva JV, et al. Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term potentiation in vivo. Nature. 2002;416:535-539. 26. McGeer PL, Itagaki S, Boyes BE, McGeer EG. Reactive microglia are positive for HLA-DR in the substantia nigra of parkinson's and alzheimer's disease brains. Neurology. 1988;38:1285-1291. 27. Rogers J , Luber-Narod J , Styren SD, Civin WH. Expression of immune system-associated antigens by cells of the human central nervous system: Relationship to the pathology of alzheimer's disease. Neurobiol Aging. 1988;9:339-349. 28. McGeer PL, McGeer E G . Inflammation, autotoxicity and alzheimer disease. Neurobiol Aging. 2001;22:799-809. 29. Webster S, Lue LF, Brachova L, et al. Molecular and cellular characterization of the membrane attack complex, C5b-9, in alzheimer's disease. Neurobiol Aging. 1997;18:415-421. 30. McGeer PL, McGeer E, Rogers J , Sibley J . Anti-inflammatory drugs and alzheimer disease. Lancet. 1990;335:1037. 85 31. Wyss-Coray T. Inflammation in alzheimer disease: Driving force, bystander or beneficial response? Nat Med. 2006;12:1005-1015. 32. Tanzi RE, Bertram L. Twenty years of the alzheimer's disease amyloid hypothesis: A genetic perspective. Cell. 2005;120:545-555. 33. Citron M, Oltersdorf T, Haass C, et al. Mutation of the beta-amyloid precursor protein in familial alzheimer's disease increases beta-protein production. Nature. 1992;360:672-674. 34. Cai XD, Golde TE, Younkin S G . Release of excess amyloid beta protein from a mutant amyloid beta protein precursor. Science. 1993;259:514-516. 35. Scheuner D, Eckman C, Jensen M, et al. 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 A P P mutations linked to familial alzheimer's disease. Nat Med. 1996;2:864-870. 36. Citron M, Westaway D, Xia W, et al. Mutant presenilins of alzheimer's disease increase production of 42-residue amyloid beta-protein in both transfected cells and transgenic mice. Nat Med. 1997;3:67-72. 37. Haass C, Hung AY, Selkoe DJ, Teplow DB. Mutations associated with a locus for familial alzheimer's disease result in alternative processing of amyloid beta-protein precursor. J Biol Chem. 1994;269:17741-17748. 38. Kumar-Singh S, Theuns J , Van Broeck B, et al. Mean age-of-onset of familial alzheimer disease caused by presenilin mutations correlates with both increased Abeta42 and decreased Abeta40. Hum Mutat. 2006;27:686-695. 39. Tanzi RE. Molecular genetics of alzheimer's disease and the amyloid beta peptide precursor gene. Ann Med. 1989;21:91-94. 40. Kang J , Lemaire HG, Unterbeck A, et al. The precursor of alzheimer's disease amyloid A4 protein resembles a cell-surface receptor. Nature. 1987;325:733-736. 41. Moya KL, Benowitz LI, Schneider GE, Allinquant B. The amyloid precursor protein is developmental^ regulated and correlated with synaptogenesis. Dev Biol. 1994;161:597-603. 42. Nishimoto I, Okamoto T, Matsuura Y, et al. Alzheimer amyloid protein precursor complexes with brain GTP-binding protein G(o). Nature. 1993;362:75-79. 43. Haass C, Schlossmacher MG, Hung AY, et al. Amyloid beta-peptide is produced by cultured cells during normal metabolism. Nature. 1992;359:322-325. 44. Zheng H, Koo EH. The amyloid precursor protein: Beyond amyloid. Mol Neurodegener. 2006;1:5. 45. Haass C. Take f ive-BACE and the gamma-secretase quartet conduct alzheimer's amyloid beta-peptide generation. EMBO d. 2004;23:483-488. 46. Ehehalt R, Keller P, Haass C, Thiele C, Simons K. Amyloidogenic processing of the alzheimer beta-amyloid precursor protein depends on lipid rafts. J Cell Biol. 2003;160:113-123. 47. Wolozin B. A fluid connection: Cholesterol and abeta. Proc Natl Acad Sci USA. 2001 ;98:5371-5373. 48. Luo Y, Bolon B, Kahn S, et al. Mice deficient in BACE1, the alzheimer's beta-secretase, have normal phenotype and abolished beta-amyloid generation. Nat Neurosci. 2001;4:231-232. 49. Cai H, Wang Y, McCarthy D, et al. BACE1 is the major beta-secretase for generation of abeta peptides by neurons. Nat Neurosci. 2001;4:233-234. 50. Dietschy JM, Turley SD. Cholesterol metabolism in the brain. Curr Opin Lipidol. 2001 ;12:105-112. 51. Beffert U, Stolt PC, Herz J . Functions of lipoprotein receptors in neurons, d Lipid Res. 2004;45:403-409. 87 52. Horton JD, Goldstein JL, Brown MS. SREBPs : Activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest. 2002;109:1125-1131. 53. Goldstein JL, Brown MS. Regulation of the mevalonate pathway. Nature. 1990;343:425-430. 54. Soccio RE, Breslow JL. Intracellular cholesterol transport. Arterioscler Thromb Vase Biol. 2004;24:1150-1160. 55. Snipes G J , Suter U. Cholesterol and myelin. Subcell Biochem. 1997;28:173-204. 56. Wilson JD. The measurement of the exchangeable pools of cholesterol in the baboon. J Clin Invest. 1970;49:655-665. 57. Quan G, Xie C, Dietschy JM, Turley SD. Ontogenesis and regulation of cholesterol metabolism in the central nervous system of the mouse. Brain Res Dev Brain Res. 2003;146:87-98. 58. Rubin LL, Staddon JM. The cell biology of the blood-brain barrier. Annu Rev Neurosci. 1999;22:11-28. 59. Bjorkhem I, Diczfalusy U, Lutjohann D. Removal of cholesterol from extrahepatic sources by oxidative mechanisms. CurrOpin Lipidol. 1999;10:161-165. 60. Xie C, Lund EG, Turley SD, Russell DW, Dietschy JM. Quantitation of two pathways for cholesterol excretion from the brain in normal mice and mice with neurodegeneration. J Lipid Res. 2003;44:1780-1789. 61. Kolsch H, Lutjohann D, Ludwig M, et al. Polymorphism in the cholesterol 24S-hydroxylase gene is associated with alzheimer's disease. Mol Psychiatry. 2002;7:899-902. 62. Wolozin B, Manger J , Bryant R, Cordy J , Green RC, McKee A. Re-assessing the relationship between cholesterol, statins and alzheimer's disease. Acta Neurol Scand Suppl. 2006;185:63-70. 63. Vance JE , Hayashi H, Karten B. Cholesterol homeostasis in neurons and glial cells. Semin Cell Dev Biol. 2005;16:193-212. 64. Posse De Chaves El , Vance DE, Campenot RB, Kiss RS, Vance JE . Uptake of lipoproteins for axonal growth of sympathetic neurons. J Biol Chem. 2000;275:19883-19890. 65. de Chaves El , Rusinol AE , Vance DE, Campenot RB, Vance JE. Role of lipoproteins in the delivery of lipids to axons during axonal regeneration. J Biol Chem. 1997;272:30766-30773. 66. Corder EH, Saunders AM, Strittmatter W J , et al. Gene dose of apolipoprotein E type 4 allele and the risk of alzheimer's disease in late onset families. Science. 1993;261:921-923. 67. Corder EH, Saunders AM, Risch NJ, et al. Protective effect of apolipoprotein E type 2 allele for late onset alzheimer disease. Nat Genet. 1994;7:180-184. 68. Strittmatter W J , Roses AD. Apolipoprotein E and alzheimer disease. Proc Natl Acad Sci USA. 1995;92:4725-4727. 69. Rapp A, Gmeiner B, Huttinger M. Implication of apoE isoforms in cholesterol metabolism by primary rat hippocampal neurons and astrocytes. Biochimie. 2006;88:473-483. 70. DeMattos RB, Curtiss LK, Williams DL. A minimally lipidated form of cell-derived apolipoprotein E exhibits isoform-specific stimulation of neurite outgrowth in the absence of exogenous lipids or lipoproteins. J Biol Chem. 1998;273:4206-4212. 71. LaDu MJ, Falduto MT, Manelli AM, Reardon CA, Getz G S , Frail DE. Isoform-specific binding of apolipoprotein E to beta-amyloid. J Biol Chem. 1994;269:23403-23406. 72. Zhou Z, Relkin N, Ghiso J , Smith JD, Gandy S. Human cerebrospinal fluid apolipoprotein E isoforms are apparently inefficient at complexing with synthetic alzheimer's amyloid-[beta] peptide (A[beta] 1-40 ) in vitro. Mol Med. 2002;8:376-381. 73. Bell RD, Sagare AP , Friedman AE, et al. Transport pathways for clearance of human alzheimer's amyloid beta-peptide and apolipoproteins E and J in the mouse central nervous system. J Cereb Blood Flow Metab. 2006. 89 74. Bales KR, Verina T, Cummins DJ, et al. Apolipoprotein E is essential for amyloid deposition in the APP(V717F) transgenic mouse model of alzheimer's disease. Proc Natl Acad Sci USA. 1999;96:15233-15238. 75. Kivipelto M, Ngandu T, Fratiglioni L, et al. Obesity and vascular risk factors at midlife and the risk of dementia and alzheimer disease. Arch Neurol. 2005;62:1556-1560. 76. Kivipelto M, Helkala EL, Laakso MP, et al. Apolipoprotein E epsilon4 allele, elevated midlife total cholesterol level, and high midlife systolic blood pressure are independent risk factors for late-life alzheimer disease. Ann Intern Med. 2002;137:149-155. 77. Notkola IL, Sulkava R, Pekkanen J , et al. Serum total cholesterol, apolipoprotein E epsilon 4 allele, and alzheimer's disease. Neuroepidemiology. 1998;17:14-20. 78. Kalmijn S, Foley D, White L, et al. Metabolic cardiovascular syndrome and risk of dementia in japanese-american elderly men. the honolulu-asia aging study. Arterioscler Thromb Vase Biol. 2000;20:2255-2260. 79. Tan ZS, Seshadri S, Beiser A, et al. Plasma total cholesterol level as a risk factor for alzheimer disease: The framingham study. Arch Intern Med. 2003;163:1053-1057. 80. Mielke MM, Zandi PP , Sjogren M, et al. High total cholesterol levels in late life associated with a reduced risk of dementia. Neurology. 2005;64:1689-1695. 81. Kivipelto M, Solomon A. Cholesterol as a risk factor for alzheimer's disease - epidemiological evidence. Acta Neurol Scand Suppl. 2006;185:50-57. 82. Sparks DL, Scheff SW, Hunsaker JC,3rd, Liu H, Landers T, Gross DR. Induction of alzheimer-like beta-amyloid immunoreactivity in the brains of rabbits with dietary cholesterol. Exp Neurol. 1994;126:88-94. 83. Refolo LM, Malester B, LaFrancois J , et al. Hypercholesterolemia accelerates the alzheimer's amyloid pathology in a transgenic mouse model. Neurobiol Dis. 2000;7:321-331. 90 84. Howland DS, Trusko SP, Savage MJ, et al. Modulation of secreted beta-amyloid precursor protein and amyloid beta-peptide in brain by cholesterol. J Biol Chem. 1998;273:16576-16582. 85. George AJ , Holsinger RM, McLean CA, et al. A P P intracellular domain is increased and soluble abeta is reduced with diet-induced hypercholesterolemia in a transgenic mouse model of alzheimer disease. Neurobiol Dis. 2004;16:124-132. 86. Wolozin B, Kellman W, Ruosseau P, Celesia G G , Siegel G. Decreased prevalence of alzheimer disease associated with 3-hydroxy-3-methyglutaryl coenzyme A reductase inhibitors. Arch Neurol. 2000;57:1439-1443. 87. Shepherd J , Blauw G J , Murphy MB, et al. Pravastatin in elderly individuals at risk of vascular disease (PROSPER): A randomised controlled trial. Lancet. 2002;360:1623-1630. 88. Heart Protection Study Collaborative Group. MRC/BHF heart protection study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: A randomised placebo-controlled trial. Lancet. 2002;360:7-22. 89. Sparks DL, Sabbagh M, Connor D, et al. Statin therapy in alzheimer's disease. Acta Neurol Scand Suppl. 2006;185:78-86. 90. Simons M, Schwarzler F, Lutjohann D, et al. Treatment with simvastatin in normocholesterolemic patients with alzheimer's disease: A 26-week randomized, placebo-controlled, double-blind trial. Ann Neurol. 2002;52:346-350. 91. Fassbender K, Simons M, Bergmann C, et al. Simvastatin strongly reduces levels of alzheimer's disease beta -amyloid peptides abeta 42 and abeta 40 in vitro and in vivo. Proc Natl Acad Sci U S A. 2001;98:5856-5861. 92. Kojro E, Gimpl G, Lammich S, Marz W, Fahrenholz F. Low cholesterol stimulates the nonamyloidogenic pathway by its effect on the alpha -secretase ADAM 10. Proc Natl Acad Sci U S A. 2001;98:5815-5820. 91 93. Sjogren M, Gustafsson K, Syversen S, et al. Treatment with simvastatin in patients with alzheimer's disease lowers both alpha- and beta-cleaved amyloid precursor protein. Dement Geriatr Cogn Disord. 2003;16:25-30. 94. Friedhoff LT, Cullen El , Geoghagen NS, Buxbaum JD. Treatment with controlled-release lovastatin decreases serum concentrations of human beta-amyloid (A beta) peptide. Int J Neuropsychopharmacol. 2001 ;4:127-130. 95. Cordle A, Landreth G. 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors attenuate beta-amyloid-induced microglial inflammatory responses. J Neurosci. 2005;25:299-307. 96. Bi X, Baudry M, Liu J , et al. Inhibition of geranylgeranylation mediates the effects of 3-hydroxy-3-methylglutaryl (HMG)-CoA reductase inhibitors on microglia, d Biol Chem. 2004;279:48238-48245. 97. Takemoto M, Liao JK. Pleiotropic effects of 3-hydroxy-3-methylglutaryl coenzyme a reductase inhibitors. Arterioscler Thromb Vase Biol. 2001;21:1712-1719. 98. Tanzi RE. Neuropathology in the down's syndrome brain. Nat Med. 1996;2:31-32. 99. Cabrejo L, Guyant-Marechal L, Laquerriere A, et al. Phenotype associated with APP duplication in five families. Brain. 2006;129:2966-2976. 100. Wisniewski KE, Wisniewski HM, Wen GY. Occurrence of neuropathological changes and dementia of alzheimer's disease in down's syndrome. Ann Neurol. 1985;17:278-282. 101. Head E, Lott IT. Down syndrome and beta-amyloid deposition. Curr Opin Neurol. 2004;17:95-100. 102. Dean M, Hamon Y, Chimini G. The human ATP-binding cassette (ABC) transporter superfamily. J Lipid Res. 2001;42:1007-1017. 92 103. Hyde SC, Emsley P, Hartshorn MJ, et al. Structural model of ATP-binding proteins associated with cystic fibrosis, multidrug resistance and bacterial transport. Nature. 1990;346:362-365. 104. Childs S, Ling V. The MDR superfamily of genes and its biological implications. Important Adv Oncol. 1994:21-36. 105. Artie AD, Kastelein JP, Hayden MR. Pivotal role of ABCA1 in reverse cholesterol transport influencing HDL levels and susceptibility to atherosclerosis. J Lipid Res. 2001;42:1717-1726. 106. Nakamura K, Kennedy MA, Baldan A, Bojanic DD, Lyons K, Edwards PA. Expression and regulation of multiple murine ATP-binding cassette transporter G1 mRNAs/isoforms that stimulate cellular cholesterol efflux to high density lipoprotein. J Biol Chem. 2004;279:45980-45989. 107. Klucken J , Buchler C, Orso E, et al. ABCG1 (ABC8), the human homolog of the drosophila white gene, is a regulator of macrophage cholesterol and phospholipid transport. Proc Natl Acad Sci USA. 2000;97:817-822. 108. Savary S, Denizot F, Luciani M, Mattei M, Chimini G. Molecular cloning of a mammalian ABC transporter homologous to drosophila white gene. Mamm Genome. 1996;7:673-676. 109. Kennedy MA, Venkateswaran A, Tarr PT, et al. Characterization of the human ABCG1 gene: Liver X receptor activates an internal promoter that produces a novel transcript encoding an alternative form of the protein. J Biol Chem. 2001 ;276:39438-39447. 110. Kalaany NY, Mangelsdorf DJ. LXRS and FXR: The yin and yang of cholesterol and fat metabolism. Annu Rev Physiol. 2006;68:159-191. 111. Venkateswaran A, Laffitte BA, Joseph SB, et al. Control of cellular cholesterol efflux by the nuclear oxysterol receptor LXR alpha. Proc Natl Acad Sci USA. 2000;97:12097-12102. 112. Li AC, Binder C J , Gutierrez A, et al. Differential inhibition of macrophage foam-cell formation and atherosclerosis in mice by PPARalpha, beta/delta, and gamma. J Clin Invest. 2004;114:1564-1576. 93 113. Sundararajan S, Jiang Q, Heneka M, Landreth G. PPARgamma as a therapeutic target in central nervous system diseases. Neurochem Int. 2006;49:136-144. 114. Sullivan DT, Bell LA, Paton DR, Sullivan MC. Genetic and functional analysis of tryptophan transport in malpighian tubules of drosophila. Biochem Genet. 1980;18:1109-1130. 115. Cserepes J , Szentpetery Z, Seres L, et al. Functional expression and characterization of the human ABCG1 and ABCG4 proteins: Indications for heterodimerization. Biochem Biophys Res Commun. 2004;320:860-867. 116. Wang N, Lan D, Chen W, Matsuura F, Tall AR. ATP-binding cassette transporters G1 and G4 mediate cellular cholesterol efflux to high-density lipoproteins. Proc Natl Acad Sci USA. 2004;101:9774-9779. 117. Vaughan AM, Oram JF. ABCG1 redistributes cell cholesterol to domains removable by high density lipoprotein but not by lipid-depleted apolipoproteins. J Biol Chem. 2005;280:30150-30157. 118. Kennedy MA, Barrera GC, Nakamura K, et al. ABCG1 has a critical role in mediating cholesterol efflux to HDL and preventing cellular lipid accumulation. Cell Metab. 2005;1:121-131. 119. Gelissen IC, Harris M, Rye KA, et al. ABCA1 and ABCG1 synergize to mediate cholesterol export to apoA-l. Artenoscler Thromb Vase Biol. 2006;26:534-540. 120. Vaughan AM, Oram JF. ABCA1 and ABCG1 or ABCG4 act sequentially to remove cellular cholesterol and generate cholesterol-rich HDL. J Lipid Res. 2006;47:2433-2443. 121. Oram JF, Vaughan AM. ATP-binding cassette cholesterol transporters and cardiovascular disease. CircRes. 2006;99:1031-1043. 122. Basso F, Amar MJ, Wagner EM, et al. Enhanced ABCG1 expression increases atherosclerosis in LDLr-KO mice on a western diet. Biochem Biophys Res Commun. 2006;351:398-404. 94 123. Ranalletta M, Wang N, Han S, Yvan-Charvet L, Welch C, Tall AR. Decreased atherosclerosis in low-density lipoprotein receptor knockout mice transplanted with Abcgl- / - bone marrow. Arterioscler Thromb Vase Biol. 2006;26:2308-2315. 124. Baldan A, Pei L, Lee R, et al. Impaired development of atherosclerosis in hyperlipidemic Idlr-/-and ApoE-/- mice transplanted with Abcgl- / - bone marrow. Arterioscler Thromb Vase Biol. 2006;26:2301-2307. 125. Out R, Hoekstra M, Meurs I, et al. Total body ABCG1 expression protects against early atherosclerotic lesion development in mice. Arterioscler Thromb Vase Biol. 2007;27:594-599. 126. Mao R, Zielke CL, Zielke HR, Pevsner J . Global up-regulation of chromosome 21 gene expression in the developing down syndrome brain. Genomics. 2003;81:457-467. 127. Tansley GH, Burgess BL, Bryan MT, et al. The cholesterol transporter ABCG1 modulates the subcellular distribution and proteolytic processing of beta -amyloid precursor protein. J Lipid Res. 2007. 128. Bocconi L, Nava S, Fogliani R, Nicolini U. Trisomy 21 is associated with hypercholesterolemia during intrauterine life. Am J Obstet Gynecol. 1997;176:540-543. 129. Toscano E, Trivellini V, Andria G. Cholelithiasis in down's syndrome. Arch Dis Child. 2001;85:242-243. 130. Admirand WH, Small DM. The physicochemical basis of cholesterol gallstone formation in man. J Clin Invest. 1968;47:1043-1052. 131. Salo MK, Solakivi-Jaakkola T, Kivimaki T, Nikkari T. Plasma lipids and lipoproteins in down's syndrome. ScandJClin Lab Invest. 1979;39:485-490. 132. Yla-Herttuala S, Luoma J , Nikkari T, Kivimaki T. Down's syndrome and atherosclerosis. Atherosclerosis. 1989;76:269-272. 95 133. Oddo S, Caccamo A, Shepherd JD, et al. Triple-transgenic model of alzheimer's disease with plaques and tangles: Intracellular abeta and synaptic dysfunction. Neuron. 2003;39:409-421. 134. Duff K, Suleman F. Transgenic mouse models of alzheimer's disease: How useful have they been for therapeutic development? Brief Funct Genomic Proteomic. 2004;3:47-59. 135. Games D, Adams D, Alessandrini R, et al. Alzheimer-type neuropathology in transgenic mice overexpressing V717F beta-amyloid precursor protein. Nature. 1995;373:523-527. 136. Irizarry MC, Soriano F, McNamara M, et al. Abeta deposition is associated with neuropil changes, but not with overt neuronal loss in the human amyloid precursor protein V717F (PDAPP) transgenic mouse. J Neurosci. 1997;17:7053-7059. 137. Huitron-Resendiz S, Sanchez-Alavez M, Gallegos R, et al. Age-independent and age-related deficits in visuospatial learning, sleep-wake states, thermoregulation and motor activity in PDAPP mice. Brain Res. 2002;928:126-137. 138. Holtzman DM, Bales KR, Tenkova T, et al. Apolipoprotein E isoform-dependent amyloid deposition and neuritic degeneration in a mouse model of alzheimer's disease. Proc Natl Acad Sci USA. 2000;97:2892-2897. 139. Oyama F, Cairns NJ, Shimada H, Oyama R, Titani K, lhara Y. Down's syndrome: Up-regulation of beta-amyloid protein precursor and tau mRNAs and their defective coordination. J Neurochem. 1994;62:1062-1066. 140. Masliah E, Sisk A, Mallory M, Mucke L, Schenk D, Games D. Comparison of neurodegenerative pathology in transgenic mice overexpressing V717F beta-amyloid precursor protein and alzheimer's disease. J Neurosci. 1996;16:5795-5811. 141. Innis S M , Dyer RA. Brain astrocyte synthesis of docosahexaenoic acid from n-3 fatty acids is limited at the elongation of docosapentaenoic acid. J Lipid Res. 2002;43:1529-1536. 96 142. Lutjohann D, Stroick M, Bertsch T, et al. High doses of simvastatin, pravastatin, and cholesterol reduce brain cholesterol synthesis in guinea pigs. Steroids. 2004;69:431-438. 143. Wellington CL, Yang YZ, Zhou S, et al. Truncation mutations in ABCA1 suppress normal upregulation of full-length ABCA1 by 9-cis-retinoic acid and 22-R-hydroxycholesterol. J Lipid Res. 2002;43:1939-1949. 144. DeMattos RB, Cirrito JR, Parsadanian M, et al. ApoE and clusterin cooperatively suppress abeta levels and deposition: Evidence that ApoE regulates extracellular abeta metabolism in vivo. Neuron. 2004;41:193-202. 145. Fryer JD, Demattos RB, McCormick LM, et al. The low density lipoprotein receptor regulates the level of central nervous system human and murine apolipoprotein E but does not modify amyloid plaque pathology in PDAPP mice. J Biol Chem. 2005;280:25754-25759. 146. Li Y, Zhou W, Tong Y, He G, Song W. Control of A P P processing and abeta generation level by BACE1 enzymatic activity and transcription. FASEB J. 2006;20:285-292. 147. Chen F, Yu G, Arawaka S, et al. Nicastrin binds to membrane-tethered notch. Nat Cell Biol. 2001;3:751-754. 148. Haass C, Selkoe DJ. Cellular processing of beta-amyloid precursor protein and the genesis of amyloid beta-peptide. Cell. 1993;75:1039-1042. 149. Evin G, Zhu A, Holsinger RM, Masters CL, Li QX. Proteolytic processing of the alzheimer's disease amyloid precursor protein in brain and platelets. J Neurosci Res. 2003;74:386-392. 150. Hyman BT, Van Hoesen GW, Damasio AR. Memory-related neural systems in alzheimer's disease: An anatomic study. Neurology. 1990;40:1721-1730. 151. Zhao J , Paganini L, Mucke L, et al. Beta-secretase processing of the beta-amyloid precursor protein in transgenic mice is efficient in neurons but inefficient in astrocytes. J Biol Chem. 1996;271:31407-31411. 97 152. Esch. Cleavage of amyloid beta peptide during constitutive processing of its precursor. Science. 1990;248:1122. 153. Karten B, Campenot RB, Vance DE, Vance JE . Expression of ABCG1, but not ABCA1, correlates with cholesterol release by cerebellar astroglia. J Biol Chem. 2006;281:4049-4057. 154. Giraldo P, Montoliu L. Size matters: Use of YACs, BACs and PACs in transgenic animals. Transgenic Res. 2001;10:83-103. 155. Singaraja RR, Bocher V, James ER, et al. Human ABCA1 BAC transgenic mice show increased high density lipoprotein cholesterol and ApoAI-dependent efflux stimulated by an internal promoter containing liver X receptor response elements in intron 1. J Biol Chem. 2001 ;276:33969-33979. 156. Maxwell KN, Soccio RE, Duncan EM, Sehayek E, Breslow JL. Novel putative S R E B P and LXR target genes identified by microarray analysis in liver of cholesterol-fed mice. J Lipid Res. 2003;44:2109-2119. 157. Gygi S P , Rochon Y, Franza BR, Aebersold R. Correlation between protein and mRNA abundance in yeast. Mol Cell Biol. 1999;19:1720-1730. 158. Hirsch-Reinshagen V, Chan JY, Wilkinson A, et al. Physiologically regulated transgenic ABCA1 does not reduce amyloid burden or abeta levels in vivo. J Lipid Res. 2007. 159. Wahrle S E , Jiang H, Parsadanian M, et al. ABCA1 is required for normal central nervous system ApoE levels and for lipidation of astrocyte-secreted apoE. J Biol Chem. 2004;279:40987-40993. 160. Hirsch-Reinshagen V, Zhou S, Burgess BL, et al. Deficiency of ABCA1 impairs apolipoprotein E metabolism in brain. J Biol Chem. 2004;279:41197-41207. 161. Berg KA, Petty HR. Cholesteryl ester handling by RAW264 macrophages: Response to native and acetylated low density lipoprotein. Mol Cell Biochem. 1988;84:29-40. 98 162. Lund EG, Xie C, Kotti T, Turley SD, Dietschy JM, Russell DW. Knockout of the cholesterol 24-hydroxylase gene in mice reveals a brain-specific mechanism of cholesterol turnover. J Biol Chem. 2003;278:22980-22988. 163. Lutjohann D, Breuer O, Ahlborg G, et al. Cholesterol homeostasis in human brain: Evidence for an age-dependent flux of 24S-hydroxycholesterol from the brain into the circulation. Proc Natl Acad Sci USA. 1996;93:9799-9804. 164. CHOBANIAN AV, HOLLANDER W. Body cholesterol metabolism in man. I. the equilibration of serum and tissue cholesterol. J Clin Invest. 1962;41:1732-1737. 165. Meaney S, Hassan M, Sakinis A, et al. Evidence that the major oxysterols in human circulation originate from distinct pools of cholesterol: A stable isotope study. J Lipid Res. 2001 ;42:70-78. 166. Tachikawa M, Watanabe M, Hori S, et al. Distinct spatio-temporal expression of ABCA and A B C G transporters in the developing and adult mouse brain. J Neurochem. 2005;95:294-304. 167. Lund E G , Guileyardo JM, Russell DW. cDNA cloning of cholesterol 24-hydroxylase, a mediator of cholesterol homeostasis in the brain. Proc Natl Acad Sci USA. 1999;96:7238-7243. 168. Mahley RW, Weisgraber KH, Huang Y. Apolipoprotein E4: A causative factor and therapeutic target in neuropathology, including alzheimer's disease. Proc Natl Acad Sci USA. 2006; 103:5644-5651. 169. Wahrle S E , Jiang H, Parsadanian M, et al. Deletion of Abcal increases abeta deposition in the PDAPP transgenic mouse model of alzheimer disease. J Biol Chem. 2005;280:43236-43242. 170. Von Eckardstein A, Langer C, Engel T, et al. ATP binding cassette transporter ABCA1 modulates the secretion of apolipoprotein E from human monocyte-derived macrophages. FASEB J. 2001;15:1555-1561. 99 171. Weiner HL, Lemere CA, Maron R, et al. Nasal administration of amyloid-beta peptide decreases cerebral amyloid burden in a mouse model of alzheimer's disease. Ann Neurol. 2000;48:567-579. 172. Cirrito JR, May PC, O'Dell MA, et al. In vivo assessment of brain interstitial fluid with microdialysis reveals plaque-associated changes in amyloid-beta metabolism and half-life. J Neurosci. 2003;23:8844-8853. 173. Mathews PM, Jiang Y, Schmidt SD, Grbovic OM, Mercken M, Nixon RA. Calpain activity regulates the cell surface distribution of amyloid precursor protein, inhibition of clapains enhances endosomal generation of beta-cleaved C-terminal A P P fragments. J Biol Chem. 2002;277:36415-36424. 174. Yao J , Petanceska SS, Montine TJ, et al. Aging, gender and A P O E isotype modulate metabolism of alzheimer's abeta peptides and F-isoprostanes in the absence of detectable amyloid deposits. J Neurochem. 2004;90:1011-1018. 175. Kim WS, Rahmanto AS , Kamili A, et al. Role of ABCG1 and ABCA1 in regulation of neuronal cholesterol efflux to apolipoprotein E discs and suppression of amyloid-beta peptide generation. J Biol Chem. 2007;282:2851-2861. 176. Fukumoto H, Deng A, Irizarry MC, Fitzgerald ML, Rebeck GW. Induction of the cholesterol transporter ABCA1 in central nervous system cells by liver X receptor agonists increases secreted abeta levels. J Biol Chem. 2002;277:48508-48513. 177. Sun Y, Yao J , Kim TW, Tall AR. Expression of liver X receptor target genes decreases cellular amyloid beta peptide secretion. J Biol Chem. 2003;278:27688-27694. 178. Koldamova RP, Lefterov IM, Staufenbiel M, etal. The liver X receptor ligand T0901317 decreases amyloid beta production in vitro and in a mouse model of alzheimer's disease. J Biol Chem. 2005;280:4079-4088. 100 179. Hirsch-Reinshagen V, Maia LF, Burgess BL, et al. The absence of ABCA1 decreases soluble ApoE levels but does not diminish amyloid deposition in two murine models of alzheimer disease. J Biol Chem. 2005;280:43243-43256. 180. Koldamova R, Staufenbiel M, Lefterov I. Lack of ABCA1 considerably decreases brain ApoE level and increases amyloid deposition in APP23 mice. J Biol Chem. 2005;280:43224-43235. 181. FitzPatrick DR, Ramsay J , McGill NI, Shade M, Carothers AD, Hastie ND. Transcriptome analysis of human autosomal trisomy. Hum Mol Genet. 2002;11:3249-3256. 182. Hoekstra M, Out R, Kruijt JK, Van Eck M, Van Berkel TJ. Diet induced regulation of genes involved in cholesterol metabolism in rat liver parenchymal and kupffer cells. J Hepatol. 2005;42:400-407. 183. Wang N, Ranalletta M, Matsuura F, Peng F, Tall AR. LXR-induced redistribution of ABCG1 to plasma membrane in macrophages enhances cholesterol mass efflux to HDL. Arterioscler Thromb Vase Biol. 2006;26:1310-1316. 184. Graf GA, Cohen JC, Hobbs HH. Missense mutations in ABCG5 and ABCG8 disrupt heterodimerization and trafficking. J Biol Chem. 2004;279:24881-24888. 185. Kobayashi A, Takanezawa Y, Hirata T, et al. Efflux of sphingomyelin, cholesterol, and phosphatidylcholine by ABCG1. J Lipid Res. 2006;47:1791-1802. 186. Adams C M , Goldstein JL, Brown MS. Cholesterol-induced conformational change in S C A P enhanced by insig proteins and mimicked by cationic amphiphiles. Proc Natl Acad Sci USA. 2003;100:10647-10652. 187. Funfschilling U, SaherG, Xiao L, Mobius W, Nave KA. Survival of adult neurons lacking cholesterol synthesis in vivo. BMC Neurosci. 2007;8:1. 188. Pfrieger FW. Outsourcing in the brain: Do neurons depend on cholesterol delivery by astrocytes? Bioessays. 2003;25:72-78. 101 189. Lutjohann D, Brzezinka A, Barth E, et al. Profile of cholesterol-related sterols in aged amyloid precursor protein transgenic mouse brain. J Lipid Res. 2002;43:1078-1085. 190. Fezoui Y, Teplow DB. Kinetic studies of amyloid beta-protein fibril assembly, differential effects of alpha-helix stabilization. J Biol Chem. 2002;277:36948-36954. 102 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0100915/manifest

Comment

Related Items