Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Overexpression of ABCG1 does not contribute to cognitive deficits in Down syndrome-related Alzheimer's… Parkinson, Pamela Faye 2008

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


24-ubc_2008_spring_parkinson_pamela_faye.pdf [ 2.65MB ]
JSON: 24-1.0066730.json
JSON-LD: 24-1.0066730-ld.json
RDF/XML (Pretty): 24-1.0066730-rdf.xml
RDF/JSON: 24-1.0066730-rdf.json
Turtle: 24-1.0066730-turtle.txt
N-Triples: 24-1.0066730-rdf-ntriples.txt
Original Record: 24-1.0066730-source.json
Full Text

Full Text

OVEREXPRESSION OF ABCG1 DOES NOT CONTRIBUTE TO COGNITIVE DEFICITS IN DOWN SYNDROME-RELATED ALZHEIMER'S DISEASE by PAMELA FAYE PARKINSON B.Sc., The University of Alberta, 2005 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 (Vancouver) April 2008 © Pamela Faye Parkinson, 2008 ABSTRACT Cognitive deficits are a hallmark feature of both Down Syndrome (DS) and Alzheimer's Disease (AD). Individuals with DS exhibit a very early onset of AD neuropathology, by their mid to late 30's. Extra copies of the genes on chromosome 21 may play an important role in this accelerated onset of AD in DS individuals. The amyloid precursor protein (APP) is located on chromosome 21, and among its cleavage products is amyloid-beta (Am, a component of amyloid plaques. The presence of A13 and amyloid in the brain is a key pathogenic factor, and is considered the central and causative neuropathology in AD by the amyloid cascade hypothesis. Growing evidence suggests an important role for cholesterol in the pathogenesis of AD, particularly in APP metabolism and production of A13 peptides. The ATP-Binding Cassette-G1 (ABCG1) transporter is located on chromsome 21, and is believed to participate in the maintenance of cholesterol homeostasis. The effects of ABCG1 expression on the production of Ar3 have proved inconclusive in in vitro studies, demanding an in vivo resolution where appropriate physiology is maintained. To test the hypothesis that overexpression of ABCG1 will accelerate the onset or progression of AD in vivo, we evaluated the cognitive performance of ABCG1-overexpressing mice before and after crossing to the PDAPP mouse model of AD. Both normal and AD mice overexpressing ABCG1 showed no significant deficits on several cognitive tests, including reference and working memory task variations of the Morris Water Maze. Golgi analysis of neuronal structure revealed significantly reduced dendritic complexity in both normal and PDAPP mice overexpressing ABCG1, suggesting that the cholesterol- related functions of ABCG1 have a potentially important role in dendrite development. Interestingly, behavioural analysis of ABCG1-deficient mice revealed a gene- dose dependent trend toward worsened performance on the water maze probe trial, ii suggesting that the pathways that may compensate for ABCG1 overexpression could be unable to offset a complete deficiency. These experiments suggest an important role for ABCG1 in maintaining cellular cholesterol homeostasis, but do not support the hypothesis that ABCG 1 expression contributes to the accelerated onset of AD pathology in DS individuals. iii TABLE OF CONTENTS ABSTRACT^ ii TABLE OF CONTENTS^ iv LIST OF TABLES vii LIST OF FIGURES^ viii LIST OF ABBREVIATIONS AND ACRONYMS^ ix ACKNOWLEDGEMENTS^ xi 1. BACKGROUND 1 1.1. Alzheimer's Disease^ 1 1.1.1. Alzheimer's Disease — epidemiology and impact^ 1 1.1.2. Diagnosis and clinical profile^  1 1.1.3. AD and cognition^  2 1.1.4. AD neuropathology 3 Neurofibrillary tangles (NFTs)^ 3 Amyloid plaques^ 4 1.1.5. Neuroinflammation 4 1.1.6. Amyloid cascade hypothesis 5 1.1.7. APP processing^ 6 1.1.8. Mouse models of AD 7 PDAPP mouse model of AD^ 8 1.2. Cholesterol^ 9 1.2.1. Cholesterol - function, regulation and transport in cells^ 9 1.2.2. Cholesterol in the brain ^  10 1.2.3. Cholesterol - connection to AD^ 12 Apolipoprotein E  12 Plasma Cholesterol Levels 13 Statins^  14 1.3. Down Syndrome and AD 15 1.4. ATP-Binding-Cassette G1 17 1.4.1. ATP-Binding-Cassette superfamily^  17 1.4.2. The ABCG1 Transporter^  18 1.4.3. ABCG1 function and regulation  18 1.4.4. Human disease and ABCG1 20 1.5. Research Rationale and Hypothesis^ 21 1.5.1 Rationale^ 21 1.5.2. Experimental hypothesis and objectives^ 22 Specific objectives^ 22 iv 2. MATERIALS AND METHODS^ 23 2.1. Transgenic Animals^ 23 2.1.1. Generation of ABCG1 BAC transgenic mice^ 23 2.1.2. PDAPP Mice 23 2.1.3. ABCG1-deficient mice^ 24 2.1.4. Diet^ 24 2.1.5. Acclimatization, enrichment and light cycle^ 25 2.2. Behavioural Testing^ 25 2.2.1. Testing conditions^ 25 2.2.2. SHIRPA 26 2.2.3. Open Field Test 26 2.2.4. Elevated Plus Maze^ 27 2.2.5. Morris Water Maze 27 Visible platform task^ 28 Spatial reference memory task^ 28 Spatial working memory task 29 2.3. Golgi-Cox Staining^ 29 2.3.1. Preparing Golgi-Cox solution ^ 29 2.3.2. Sacrifice and brain collection 30 2.3.3. Brain sectioning^ 30 2.3.4. Slide preparation and staining^ 31 2.3.5. Dendritic analysis 32 2.4. Statistical analysis^ 32 3. RESULTS^ 33 3.1. ABCG1 BAC transgenic mice^ 33 3.1.1. ABCG1 is overexpressed in brain^ 33 3.1.2. ABCG1 BAC Tg mice show normal baseline behaviour^ 33 3.1.3. ABCG1 overexpression has no effect on anxiety or general locomotor activity ^ 34 3.1.4. Overexpression of ABCG1 alone does not affect learning and memory^ 35 3.2. ABCG1 and Alzheimer's Disease^ 35 3.2.1. ABCG1 overexpression does not alter anxiety or general locomotor activity in PDAPP mice^ 36 3.2.2. Determination of a suitable cohort for water maze task^ 38 3.2.3. Overexpression of ABCG1 did not affect performance on the spatial reference memory water maze task 38 3.2.4. Performance on the spatial working memory water maze task was not affected by overexpression of ABCG1 40 3.2.5. ABCG1 overexpression does not influence performance of PDAPP mice on behavioural tests^ 41 3.3. ABCG1-deficient mice 41 3.3.1. ABCG1-deficient mice do not express ABCG1 protein^ 43 3.3.2. Loss of ABCG1 expression tends to worsen probe trial performance^ 43 3.4. Morphological analysis of Golgi-stained neurons^ 44 3.4.1. ABCG1 overexpression does not affect number of primary dendrites, branch order analysis or length to branch point^ 46 3.4.2. Overexpression of ABCG1 results in less complex dendritic branching patterns ^ 47 v 4. DISCUSSION AND FUTURE DIRECTIONS^ 49 4.1. Discussion^ 49 4.1.1. ABCG1 overexpression and cognition^ 49 4.1.2. Overexpression of ABCG1 in PDAPP mice 50 4.1.3. Benefits and limitations of behavioural testing^ 52 4.1.4. ABCG1 and dendritogenesis^ 53 4.1.5. ABCG1 deficiency and cognition 56 4.2. Conclusions^ 58 4.3. Future Directions 58 4.3.1. Behavioural characterization of ABCG1-deficient mice^ 59 4.3.2. Investigate the effects of ABCG4 and ABCG1 on cognition 60 4.3.3. Conduct morphological analysis of ABCG1- and ABCG4-deficient dendritic trees^ 60 REFERENCES 78 APPENDIX 1 — Supplementary Data^ 92 APPENDIX 2 — UBC Animal Care Certificate^  93 vi LIST OF TABLES Table 1. Protocol for preparation of Golgi-stained slides^  31 vii LIST OF FIGURES Figure 1.^Processing of the amyloid precursor protein^  61 Figure 2.^Summary of the cholesterol biosynthetic pathway  62 Figure 3.^Predicted topology of the ABCG1 monomer   63 Figure 4.^Behavioural testing apparatus and methodology^  64 Figure 5.^Cell drawings of Golgi-stained neurons and dendrites  65 Figure 6.^ABCG1 mice show normal baseline phenotype  66 Figure 7.^ABCG1 overexpression does not alter anxiety or general locomotor activity   67 Figure 8.^Performance of 12-month old ABCG1 BAC Tg mice on the spatial reference memory water maze task   68 Figure 9.^PDAPP mice overexpressing ABCG1 show no differences in anxiety or Icomotor activity^  69 Figure 10.^No visual deficits revealed by water maze visual task   70 Figure 11. Overexpression of ABCG1 in PDAPP mice did not affect performance on the spatial reference memory water maze task   71 Figure 12. Performance of PDAPP mice on spatial working memory water maze task was not affected by overexpression of ABCG1 ^ 72 Figure 13.^Loss of ABCG1 expression tends to worsen water maze probe trial 73 performance^ Figure 14.^Hippocampal anatomy^  74 Figure 15. ABCG1 overexpression does not affect number of primary dendrites or length to branch point  75 Figure 16.^Branch order analysis is not significantly affected by overexpression of ABCG1    76 Figure 17.^Overexpression of ABCG1 results in less complex dendritic branching patterns   77 viii LIST OF ABBREVIATIONS AND ACRONYMS AR^Amyloid-beta peptide ABCA1^ATP-binding cassette Al ABCG 1^ATP-binding cassette G1 ABCG4^ATP-binding cassette G4 AD Alzheimer's Disease ANOVA^Analysis of variance APP Amyloid precursor protein ATP^Adenosine triphosphate BAC Bacterial artificial chromosome BACE-1^Beta-site APP cleaving enzyme-1 BBB Blood brain barrier CHO^Chinese hamster ovary CNS Central nervous system CSF^Cerebrospinal fluid DG Dentate gyrus DS^Down Syndrome EPM Elevated plus maze FAD^Familial Alzheimer's Disease fEPSP Excitatory postsynaptic field potentials HDL^High-density lipoprotein HEK Human embryonic kidney HMGR^3-hydroxy-3-methylglutaryl-CoA reductase LTP Long term potentiation LXR^Liver X receptor ix LXRE^Liver-X-receptor response element MAC Membrane attack complex mRNA^Messenger ribonucleic acid MTL Medial temporal lobe MWM^Morris Water Maze NFT Neurofibrillary tangle NI NDS-ADRDA^National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorder Association NMDA^N-methyl-D-aspartic acid OFT Open field test ORT^Object recognition test PDGF-13^Platelet derived growth factor-beta SHIRPA^Smithkline Beecham Pharmaceuticals, Harwell, MRC Mouse Genome Centre and Mammalian Genetics Unit, Imperial College School of Medicine at St Mary's, Royal London Hospital, St Bartholemew's and the Royal London School of Medicine Phenotype Assessment SREBP-2^Sterol response element binding protein-2 Tg^Transgenic x ACKNOWLEDGEMENTS I wish to thank my supervisor, Dr Cheryl Wellington, for entrusting me with a project unlike any other in her lab. Throughout this process, Cheryl has allowed me the independence necessary to complete my unique project, yet she has always been there to provide support, direction and ideas when I needed them. I also wish to thank my co- supervisor, Dr Brian Christie, who generously allowed me to 'invade' his lab space to carry out my research. Brian's involvement made this project possible, and I am grateful to him for the guidance and support. This experience has challenged me in ways I didn't expect, and my personal growth and success would not have been possible without both Cheryl and Brian. I would also like to acknowledge the members of my supervisory committee, who have provided guidance, feedback and insight on this project since the beginning. The members of the Wellington and Christie labs have contributed immeasurably to these last few years; thank you for your support and ideas, and particularly for the laughs. Of note are Gavin, Jen, Kate, Sean, Vero, James and Julie. I would like to acknowledge Brady Burgess in particular for all his help with this project, and for always being so patient with me. I also wish to thank Timal Kannangara for his support and willingness to help me out, and for making long days in the lab enjoyable with his crazy sense of humour. I wish to thank my best friend, Fiona Zeeb, for lending me her neuroscience expertise on so many occasions and for keeping me sane these last few years. I have to thank Tyler, for his support, patience and encouragement. Lastly, I wish to thank my Mum, Dad and my sister Emma, for their unwavering support. You give me the strength, drive and encouragement to succeed, and I would not have made it this far without you. xi 1. BACKGROUND 1.1. Alzheimer's Disease 1.1.1. Alzheimer's Disease — epidemiology and impact Disorders affecting the elderly are increasingly prevalent in today's shifting population demographic. Alzheimer's Disease (AD) is a neurodegenerative disease characterized by progressive worsening of cognitive function, and resulting in premature death'. AD is the most common cause of senile dementia, affecting more than 4.5 million North Americans in 2003, triple the number diagnosed in 1980 2. Increasing longevity ensures that the prevalence of AD will continue to rise, with a projected 15 million people affected by the disease by the year 2050 2. As such, AD has become a major public health concern with significant economic and social impacts. Currently, no disease-modifying therapies exist for individuals diagnosed with AD, meaning that the existing therapies only provide symptomatic relief and do not alter the course of the disease. Acetylcholinesterase inhibitors such as donezepil, rivastigmine or galantamine, and the NMDA receptor agonist memantine are currently the only licensed therapies for AD, and each produces only moderate and in most cases, short-term, symptomatic benefits to mildly affected patients 3' 4. As such, there is a large field of ongoing research devoted to developing strategies to prevent or treat AD based on an improved understanding of its pathophysiology. 1.1.2. Diagnosis and clinical profile Although a confirmed diagnosis of AD is made post-mortem, living patients can be diagnosed with probable AD based on criteria formulated by the National Institute of Neurological and Communicative Diseases and Stroke - Alzheimer's Disease and Related Disorders Association (NINDS-ADRDA). Behavioural tests are used to 1 differentiate cognitive signs of AD from other types of dementia 5. The average age at diagnosis with probable AD is 73 years of age 6, and the average lifespan following diagnosis is 4-6 years 1 . Clinical manifestations of AD involve mild memory impairments in the early stages, progressing to global cognitive deficits and severe behavioural impairments in the later stages of the disease '. As the disease progresses, individuals lose the ability to care for themselves, and experience dramatic changes in personality, often including aggression and psychotic symptoms 8. Apathy, depression and anxiety are often comorbid with the progression of AD, contributing to the decline in quality of life for patients 9 . 19 . 1.1.3. AD and cognition Though all cognitive functions decline during disease progression in AD individuals, some brain functions are more adversely affected than others. Episodic memory, or memory for the events and experiences of an individual's life, is the first and most severely affected in AD 11 . The medial temporal lobe (MTL), which includes the hippocampus, appears to be substantially more affected in AD individuals than other brain areas 12. Consequently, the MTL and hippocampal areas are associated with memory encoding and consolidation, particularly that of episodic memories. The hallmark clinical symptom of AD is memory loss, specifically profound anterograde amnesia, or inability to encode and consolidate new information 13. Focal lesions in the MTL are known to cause amnesia, and interestingly, the MTL is the site of most severe neurofibrillary tangle deposition in AD brains 11 . Though episodic memories are compromised in the initial stages of AD, deficits in semantic memory also develop as the disease progresses. Semantic memory refers to memory for meanings and other concept-based knowledge that is unrelated to specific experiences in an individual's life. Together, episodic and semantic memory comprise declarative memory, or memory of 2 facts. Therefore, individuals with AD typically exhibit great confusion, inability to recognize people from day to day, problems with spatial processing and topographical orientation, and paranoia often with recurring themes of theft as a consequence of memory deficits 13 . 1.1.4. AD neuropathology The formal diagnosis of AD is made upon post-mortem analysis of the brain, and requires histological verification of the two classic neuropathological hallmarks of AD, which are extracellular amyloid plaques found in the parenchyma and cerebral vessels, and intracellular neurofibrillary tangles (NFT) of hyperphosphorylated tau protein 5 . Cerebral atrophy due to neuronal death is another characteristic of AD neuropathology, and is particularly prominent in MTL brain regions 14, 15 . Neurofibrillary tangles (NFTs) Intracellular NFTs are composed of paired-helical filaments of tau, a microtubule- associated protein normally found in axons of neurons, but redistributed to cell bodies in tauopathies 16 ' 17. The normally soluble tau forms filamentous protein aggregates in tauopathies such as AD, and tau found in NFTs is also abnormally phosphorylated 16 . NFTs are also observed in other neurodegenerative diseases, such as Pick's disease and other forms of dementia 18. Mutations in tau are associated with an autosomal- dominant type of frontotemporal dementia, suggesting tau and NFTs play an important role in the cause of dementia 16. Density of NFTs in the hippocampus and cortex of AD patients correlates well with dementia scores obtained a short time before death, indicating that NFTs play an important role in the pathogenesis of AD 19 . 3 Amyloid plaques Amyloid plaques are the other diagnostic neuropathological finding in the AD brain. The major component of the amyloid fibrils that comprise both the diffuse and mature amyloid plaques is beta-amyloid (A13) peptides. These amyloid fibrils have been shown to be directly toxic to neurons in vitro 20. The core of an amyloid plaque consists of fibrillar A13 peptides primarily composed of 40 or 42 amino acids 21 . The A1340 and A1342 fragments are cleavage products of a much larger protein, aptly named amyloid precursor protein (APP). Several types of morphologically different plaques exist in the AD brain. Diffuse plaques represent early lesions in AD brains, containing loosely associated Ap deposits. Mature plaques are more complex, containing AO but also a variety of other aggregated proteins including apolipoproteins E and J, anti-chymotrypsin and complement proteins 22' 23. In addition, mature plaques also contain cells and cellular elements such as dystrophic neurites, astrocytic processes and activated microglia 22. Along with the toxic effect of amyloid on neurons, subsequent microglial activation may be another important factor in the pathogenesis of AD. 1.1.5. Neuroinflammation The neuroinflammation hypothesis of AD is based on histological verification of the presence of microglia in amyloid plaques 24 ' 25. Microglia are the main form of active immune defense in the brain, acting much like macrophages in the periphery. Microglia have been shown to be activated by aggregated A13, and are unsurprisingly found associated with amyloid plaques in the AD brain 26. Microglial activation has several consequences; the first being that activated microglia secrete large amounts of oxygen free radicals, in a process known as a respiratory burst. This release of cytotoxic substances is intended to kill infectious agents, but can also directly damage neurons 4 and lead to cell death 27. Secondly, microglial activation also promotes activation of the complement system, another key player in the immune response. Full activation of complement results in assembly of the membrane attack complex (MAC), designed to kill invasive cells. However, if host cells are not adequately protected, they may be damaged as well in a process known as 'bystander lysis' 26. The presence of both microglia and amyloid appear to promote the formation of the MAC, which may contribute to accidental host cell targeting and consequent neuronal damage 28. In fact, chronic microglial activation may be responsible for much of the neuronal loss in AD. Histological analysis of dystrophic neurites from AD brains reveals that many apoptotic neurons are associated with MACs, indicating that the inflammatory response is likely central to the pathogenesis of AD 26 . 1.1.6. Amyloid cascade hypothesis The most widely accepted theory regarding AD is the amyloid cascade hypothesis, which holds that the presence of A13 and amyloid in the brain is the key pathogenic factor in AD, and that other AD neuropathologies occur as a result of Ap or amyloid-induced changes 28, 30. The strongest evidence supporting this hypothesis comes from research in families affected with an early-onset form of AD, which was found to be genetically linked. The most common form of AD has no genetic basis, and thus is known as sporadic AD. Sporadic AD accounts for —95% of reported AD cases 30. In contrast, familial AD (FAD) is associated with autosomal dominant mutations that cause early- onset AD in individuals between 35 and 60 years of age 30. FAD is commonly associated with mutations in the amyloid precursor protein (APP), the cleavage products of which include the A13 peptide. Two well-known mutations in APP are the Indiana (V717F) mutation, and the Swedish (KM670/671NL), a double mutation; both of which 5 flank the A13 region of APP 31, 32. FAD-related mutations are also found in presenilin-1 and presenilin-2, proteases involved in the cleavage of APP at the C-terminal boundary of the A13 region 33 ' M . These studies revealed that FAD mutations alter both APP itself, as well as the proteases involved in its cleavage, shifting to favour the production of AR 35. Though the cause of A13 accumulation is well documented in FAD cases, in sporadic AD the cause is still unclear. 1.1.7. APP processing The Ap peptide is a product of proteolytic processing of APP. APP is a typical type-I transmembrane glycoprotein expressed in many types of cells, including neurons 36. In neurons, full-length APP is rapidly transported down axons by the anterograde transport system, and is thought to be involved in axon elongation and synaptogenesis 28' 37 . Newly synthesized APP has also been shown to be transported anterogradely from the cell bodies of neurons in the entorhinal cortex, accumulating at presynaptic terminals in the hippocampus 37 . Proteolytic processing of full-length APP is accomplished by the activity of a, 13 and y secretases, and can essentially be divided into two pathways (Figure 1) 38. In the non- amyloidogenic pathway, APP is first cleaved by a-secretase which cuts within the Ap domain, thus precluding the formation of AR peptide. Following a-secretase activity, the y-secretase cleaves the C-terminal APP fragment, C83, generating a harmless peptide. The y-secretase is a complex consisting of presenilins 1 and 2, nicastrin, PEN-2, and APH-1 38 ' 4° . Alternatively, in the amyloidogenic pathway, APP is initially cleaved by the 13-secretase-activity-enzyme 1 (BACE-1) 41 .^This cleavage produces soluble extracellular domain sAPP13, and a membrane-bound C-terminal fragment, C99, which 6 undergoes a second cleavage by y-secretase. Cleavage of C99 by y-secretase produces Ar3 peptides. The fact that mice deficient in BACE1 or the y-secretase complex do not generate Ap peptides highlights these two amyloidogenic proteases as relevant drug targets for AD 39 ' 42 . 1.1.8. Mouse models of AD Due to the complex nature of AD, it has been very difficult to create an animal model for the disease that encompasses all the characteristics of the human disease. Therefore, most of the mouse models generated to date recapitulate some, but not all, of the pathologies associated with AD. The existing mouse models can be loosely grouped in two categories, those with tau-related pathologies, and APP-related models that develop Ap and amyloid. Recently, double and triple transgenic models have been developed in an attempt to overlay the pathologies and produce a more complete model of AD. The use of mice as a model for AD has presented a challenge to researchers. Whereas individuals expressing APP at 1.5 times the normal level develop autosomal dominant AD that is 100% penetrant by midlife, mice on the other hand, require more than 8-fold overexpression in order to develop amyloid pathology 43. It has also proved difficult to create a murine model of AD that shows cell loss and cerebral atrophy characteristic of the human disease, though it is likely that neurodegeneration does occur, as many of the models show cognitive impairments 43. No mouse model to date fully replicates all aspects of human AD, but APP-related models of AD currently provide the best in vivo system available for therapeutic research based on the amyloid hypothesis. 7 PDAPP mouse model of AD The PDAPP mouse model was reported by Games and colleagues in 1995 44 , and is now a well-established model of amyloid formation. PDAPP mice overexpress human APP (hAPP) with portions of APP introns 6-8, which permit alternative splicing to produce all three major human isoforms of APP 44' 45. The PDAPP mouse also expresses the FAD-associated "Indiana" mutation V717F in hAPP, where a valine at residue 717 is substituted by phenylalanine 45. This mutation has been shown to increase the generation of AI3, in particular favouring the more amyloidogenic form, Ap42 44. Expression of APP is driven by the platelet-derived growth factor p (PDGF(3), which stimulates high expression in hippocampal neurons and cortex 46. These mice express very high levels of APP, 10-fold higher than endogenous APP, and develop robust AD- like neuropathology, including amyloid plaques, dystrophic neurites and loss of synapse density 45. Notably, plaque formation begins early in the hippocampus, developing around 6-8 months, and progressing to other cortical areas with regional specificity resembling AD 44. Plaque density in these brain regions was also observed to increase with age. Several groups have demonstrated that PDAPP animals show cognitive deficits that manifest as impaired performance on the water maze both before and after amyloid deposition 47. However, recent studies confirmed that PDAPP mice display an age- dependent deterioration in spatial reference memory, with aged 13+ months PDAPP animals displaying a lower learning capacity index than young 6-9 month-old PDAPP animals 48. Interestingly, these behavioural findings correlate with the increasing amyloid plaque load in aging PDAPP mice. Chen et al. suggest that increasing amyloid deposition interferes with and thus reduces efficiency of synaptic transmission in the 8 hippocampus, resulting in an 'episodic-like' memory deficit whereby PDAPP mice are gradually less able to encode new memories and retrieve previously stored information 48 . These studies emphasize the PDAPP mouse model as a useful tool for researchers of the amyloid hypothesis, and also highlight the importance of using suitable behavioural assays to assess cognitive deficits associated with AD-like pathologies. 1.2. Cholesterol 1.2.1. Cholesterol - function, regulation and transport in cells Cholesterol is a 27 carbon molecule composed of 4 hydrocarbon rings, a hydrocarbon tail and a hydroxyl group. Cholesterol is integral for the formation and maintenance of cell membrane permeability and fluidity, and plays a role in cell signaling. Cholesterol also functions as the substrate for synthesis of bile acids in the liver, as well as the precursor for steroid hormones in the endocrine glands. With the exception of mature red blood cells, all eukaryotic cells can synthesize cholesterol from acetate. Cholesterol biosynthesis is a complex pathway, and the rate-limiting step is the formation of mevalonate from acetate, catalyzed by the enzyme 3-hydroxy-3-methylglutaryl- Coenzyme A reductase (HMGR) 49' 50 (Figure 2). Cellular cholesterol levels are precisely controlled by strict regulation of cholesterol synthesis, storage, export and uptake. Maintaining homeostatic levels of cholesterol within the cell is extremely important, as excess membrane cholesterol can be cytotoxic 51 . Under conditions of excess cholesterol, plasma membrane cholesterol is either effluxed to extracellular acceptors or esterified for storage within the cell 52 . Most mammalian cells are unable to catabolize cholesterol, and therefore must export it 53. Cellular efflux transfers excess cholesterol to extracellular lipoprotein carriers, which enable hydrophobic cholesterol to be 9 transported throughout the body. In the periphery, the process by which cholesterol is returned to the liver for catabolism and ultimately excreted through bile is known as reverse cholesterol transport. 1.2.2. Cholesterol in the brain Despite accounting for only 2% of total body mass, the brain is the most cholesterol-rich organ in the body, containing approximately 25% of total body cholesterol 51 . Cholesterol is a major component of the myelin sheaths surrounding neuronal axons, which serve to maintain axonal conductivity. Approximately 70-80% of the cholesterol in the brain is found in myelin, with the remainder is found in cell membranes of neurons and glia, or in lipoprotein-like particles in the cerebrospinal fluid (CSF) 54. While cells in the periphery obtain cholesterol from endogenous synthesis as well as from dietary sources, the brain does not appear to import cholesterol from the peripheral circulation. Quantitative analyses show that cholesterol cannot traverse the blood-brain barrier (BBB). Therefore, the periphery and CNS do not share cholesterol pools, and all brain cholesterol must be synthesized in situ 53. In support of this, a study that monitored uptake of isotopically labeled cholesterol administered intravenously to baboons revealed that while all other organs contained large quantities of labeled cholesterol, the brain had acquired little or no labeled cholesterol 55. More recently, experiments with mice fed a high-cholesterol diet revealed that while serum lipids were increased 7-8 fold above controls, there was no evidence of cellular cholesterol accumulation in the brain, in contrast to peripheral organs 56 . While most cells in the adult brain are capable of endogenously synthesizing cholesterol, mature neurons often do not synthesize sufficient cholesterol for resource-consuming processes such as axonal growth and synaptogenesis. As a result, neurons rely largely 10 on a supply of cholesterol from glial cells in the brain. Though brain cholesterol is synthesized in situ, cells in the CNS are unable to degrade the sterol ring and as such, excess cholesterol must be delivered to the peripheral circulation for eventual excretion via the liver 57. Because cholesterol is unable to readily cross the BBB into the periphery, it is first converted to the oxysterol, 24S-hydroxycholesterol. 24S-hydroxycholesterol passively diffuses across the BBB, down its concentration gradient into the peripheral circulation. The conversion of cholesterol to its oxysterol form is catalyzed by the enzyme 24-hydroxylase (CYP46A1), which is expressed exclusively in subsets of neurons 58' 59. The mechanism by which glial cells clear cholesterol is less well understood, as glia do not express CYP46A1 51 . Intriguingly, knockout mice lacking 24- hydroxylase show an approximate 50% reduction in excretion of brain cholesterol, suggesting that cholesterol catabolism in these neurons is of crucial importance to brain cholesterol clearance 58. Additionally, a continuous turnover of cholesterol is required for normal learning and memory, as CYP46-deficient mice show no long-term potentiation (LTP) and exhibit profound impairments on spatial learning tasks 58 . Since the majority of the cholesterol in the adult brain is found in myelin and plasma membranes, brain cholesterol is largely metabolically inert 54. However, an estimated 0.02% of the total brain cholesterol pool is turned over every day 58. Extracellular transport of cholesterol in the CNS is accomplished by lipoproteins resembling HDL particles that circulate in the cerebrospinal fluid 52. The major apolipoprotein in the brain is apolipoprotein E (apoE), which is synthesized by astrocytes and microglia. Cholesterol delivered by apoE-containing lipoproteins has been shown to dramatically increase elongation of axons in times of growth or repair, and is necessary for efficient formation and maintenance of synapses 88, 81 . These findings highlight the crucial role of apoE cholesterol carriers in neuronal processes key for proper neuronal function. 11 1.2.3. Cholesterol - connection to AD Increasing evidence from recent research points to an important role for brain cholesterol metabolism in the pathogenic processes of AD. Notably, studies involving apoE genotype, plasma cholesterol levels, and the therapeutic use of statins all implicate cholesterol metabolism in the pathogenesis of AD. Apolipoprotein E ApoE, the main cholesterol-carrying apolipoprotein in the brain, is the only validated genetic risk factor for late onset, sporadic AD 62. In humans, apoE exists as three major isoforms encoded by distinct alleles (E2, E3, E4), which differ only by one or two amino acid substitutions. Of these allelic variations, apoE3 is the most common, comprising 70-78% of the population. ApoE4 is present in 15-20% of the population, and the rarest allele, apoE2, is only present in 7-8% of individuals 62-64. It is well established that inheritance of the apoE4 allele decreases the age of AD onset relative to the apoE3 allele, and that gene dose of apoE4 is inversely related to age at AD onset 62, 65. However, a single allele of apoE2 confers a protective effect, delaying the age of disease onset. For example, apoE3/E2 carriers develop AD at an average age of 84 years compared to an average age of 68 years in apoE4/E4 carriers 62 . Several mechanisms have been proposed to explain the differential effects of the three apoE isoforms on development of AD. In culture, apoE4-containing lipoproteins were shown to deliver less cholesterol to neurons than apoE3, with apoE2 delivering the most 60 . Neuronal growth and repair processes, particularly in the adult brain, are heavily dependent on apoE-derived cholesterol, thus apoE4 may be less efficient at repairing damaged neurons in AD 60. Interestingly, carriers of apoE4 also show worsened neurological impairments in head injury, stroke and multiple sclerosis, reinforcing the 12 concept that apoE4 is less effective at neurite maintenance and repair 66. An alternate mechanism suggests isoform-specific interactions of apoE with Ai3 may play a role in accelerating amyloid deposition 66. Mice deficient in apoE do not form fibrillar amyloid, suggesting that apoE is involved in conversion of soluble AR into insoluble amyloid fibrils 67 . Studies with transgenic mice expressing human apoE3 or apoE4 revealed that the apoE4 isoform specifically acts to enhance A13 aggregation and deposition, while the other isoforms promote disaggregation 65. It is possible that apoE4 decreases the age of AD onset by both promoting and accelerating amyloid deposition, and failing to effectively repair damaged neurons. Given that apoE is the only well-established genetic risk factor for sporadic AD, which accounts for more than 95% of AD cases, understanding how apoE contributes to the pathogenesis of AD is a relevant and worthwhile pursuit. Plasma Cholesterol Levels A role for plasma cholesterol in AD pathogenesis was first described by Sparks et al., who demonstrated that rabbits fed a high-cholesterol diet accumulated amyloid in the brain 68. Similar observations have since been made in other animal models, including transgenic mouse models of AD. The PSAPP mouse model of AD, when fed a high- cholesterol diet for several months, was found to have a significantly increased amyloid load, showing increases in both number and size of deposits 69. Refolo et al. also found that levels of A13 were positively correlated with high dietary cholesterol intake in these animals, suggesting that hypercholesterolemic conditions may alter processing of APP or the metabolism of Ar3 69. Results from long-term population studies in humans seem to concur with results from animal studies, suggesting that elevated plasma cholesterol is a risk factor for AD. A large human study in which 4 serum cholesterol measures 13 were taken over an average of 21 years found that cholesterol levels greater than 6.5mM in midlife were associated with a 2.6 greater chance of AD diagnosis later in life 70. Several other studies have also shown a relationship between elevated serum cholesterol levels in midlife and development of AD later in life 71-73. In addition, it has been suggested that lifestyle factors such as obesity, with its associated range of vascular factors often including dyslipidemia and hypercholesterolemia, can increase risk of developing dementia and AD in an aged individual 70. There are conflicting reports on whether plasma cholesterol levels are elevated in AD patients. Two experiments found serum cholesterol levels to be decreased after individuals developed AD, and other studies reported increased serum cholesterol and in particular, higher levels of LDL cholesterol, in AD patients as compared to controls 73. Results from these studies, though contradictory, collectively suggest that cholesterol levels are an important factor in AD, and that plasma cholesterol levels in midlife may be a better indicator of that risk than cholesterol levels upon diagnosis with AD. However, since the brain and periphery do not share cholesterol sources, the underlying mechanism driving the relationship between high serum cholesterol and development of AD has yet to be elucidated. Statins Retrospective studies have recently suggested that the use of cholesterol-lowering drugs reduces the prevalence of AD. Statins are widely prescribed to reduce LDL cholesterol in patients at risk for cardiovascular diseases, and have their effect by inhibiting the activity of HMG Co-A reductase, the catalyst for the rate-limiting step in cholesterol biosynthesis. Assessing the efficacy of statins in reducing AD prevalence or incidence has proved to be a challenge in both animal and human studies, often producing mixed or even contradictory results. In initial retrospective studies, Wolozin et al. observed a 14 60-73% decrease in the prevalence of diagnosed probable AD in patients taking two different statins, lovastatin and pravastatin, as compared to either non-treated individuals or those taking other medications for cardiovascular disease 74. However, results of recent prospective trials have been mixed. Studies designed to assess efficacy of simvastatin and pravastatin in hypercholesterolemic aged individuals did not detect a difference in the incidence of AD between treated and non-treated individuals over a 5 year follow up period 75 ' 76. In contrast, a prospective trial revealed that atorvastatin administered to individuals with mild to moderate AD significantly improved their cognitive performance 77 . Similarly mixed results were obtained in animal studies. Inhibition of cholesterol biosynthesis with simvastatin was shown to reduce levels of both A1340 and A1342 in guinea pigs 78. In contrast, another study in Tg2576 mice, an established AD mouse model, demonstrated that simvastatin did not affect A13 levels in the brain as compared to untreated mice, nor did it affect brain cholesterol levels or apoE. Interestingly, Li et al. found that treatment with simvastatin not only reversed learning and memory deficits in Tg2576 mice, but enhanced learning and memory when administered to non-transgenic mice 79. Since simvastatin administration caused no change in both AO load and total cholesterol levels in the brain, it must mediate its therapeutic effect through other pathways involved in learning and memory. 1.3. Down Syndrome and AD Down Syndrome (DS) is a genetic disease caused by inheritance of an extra copy of chromosome 21, affecting 1 in 700 newborns 88. Clinical features of DS include dysmorphic features of the head and limbs, short stature, hypotonia, heart defects and 15 mental retardation, ranging from mild cognitive impairments to severe and profound mental retardation. Due to these abnormalities, DS is also a risk factor for a number of other diseases, such as diabetes, childhood onset leukemia, and cardiac disease 80 . Clinical phenotypes vary with each DS individual, but mental retardation is the invariable hallmark of the disease. A prominent clinical phenotype of DS is inevitable development of AD neuropathology, including amyloid plaques and NFT by the mid to late 30s 81, 82. This onset of AD neuropathology manifests very early in DS individuals, decades earlier than non-DS individuals who typically exhibit AD-like neuropathology in their late 70s. Though severe AD-like neuropathology is observed in almost all DS individuals by the age of 40 82 , specific age at onset of AD cognitive symptoms and dementia is difficult to assess because of the highly variable degree of mental retardation previously existing in DS individuals. The majority of DS individuals are trisomic for the entire chromosome 21, which contains 225 genes whose specific functions are largely unknown 83. Notably, APP maps to chromosome 21, and the accelerated onset of AD-like neuropathology in DS individuals has been attributed to APP overexpression. This idea was reinforced by the case of a 78-year old DS individual who exhibited no amyloid deposition upon autopsy. The individual was found to have partial trisomy 21, in which the chromosomal material triplicated did not include the APP gene ". Interestingly, triplication of APP alone through small chromosomal rearrangements in non-DS individuals is also associated with early onset AD. Recent studies of five separate families with duplicated APP found that inheritance of three copies of APP was sufficient to cause early-onset dementia and neuropathology characteristic of AD in 21 out of 21 individuals, without other features of DS 85. These findings support the hypothesis that overexpression of APP is critical for the early onset of AD neuropathology in DS individuals, though they do not rule out the 16 possibility that additional genes on chromosome 21 may also participate in the accelerated onset of AD. In fact, it is possible that inheritance of extra genes on chromosome 21 may synergize with excess APP to accelerate the onset of AD neuropathology. 1.4. ATP-Binding-Cassette G1 1.4.1. ATP-Binding-Cassette superfamily The ATP-Binding-Cassette (ABC) transporter superfamily is one of the largest groups of protein families, and is divided into seven subgroups referred to as ABCA through ABCG. 49 transporters from this superfamily have been identified in humans to date, and they participate in the transport of a wide variety of substrates including amino acids, lipids, inorganic ions, peptides and sugars across biological membranes 66. ABC transporters are classified either as full-transporters, containing two ATP-binding domains, or half- transporters containing only one ATP-binding domain 87. Half-transporters must dimerize to form a functional protein unit. ABC transporters utilize energy derived from ATP binding or hydrolysis to carry out their function of unidirectional transport of substrates across the membrane. Typically, ABC transporters transport their substrate from the cytoplasm to the extracellular space, or when the substrate is a lipid, from the inner leaflet of the plasma membrane to the outer leaflet or to an extracellular lipid acceptor 86. Mutations in genes encoding ABC transporters have been linked to disorders such as Tangier disease or HDL deficiency, which results from mutations in ABCA1. ABC transporters ABCB1 and the breast cancer resistance protein, ABCG2, have also been implicated in multidrug resistance of cancers 66 . 17 1.4.2. The ABCG1 Transporter ABCG1 is the founding member of the ABCG subclass of the ABC transporter superfamily. ABCG transporters are generally composed of an N-terminal nucleotide- binding domain, followed by the C-terminal membrane domain containing six transmembrane a-helices and the Walker A and Walker B ATP-binding domains (Figure 3). Accordingly, ABCG1 is a half-size transporter and must dimerize in order to provide functional transporter activity. Human Abcgl is a homologue of the Drosophila white protein, which must form a heterodimer with one of two other ABCG-related proteins to transport eye pigment precursors in Drosophila 86. Both cholesterol efflux and distribution activities are present when ABCG1 is selectively expressed in cells, demonstrating that ABCG1 can function as a homodimer 89 ' 99. A recent study suggested that mammalian ABCG1 may also heterodimerize with a highly homologous transporter known as ABCG4, whose co-expression was essential for functional ATPase activity of ABCG1 in an insect cell system 91 . The assembly of functional ABCG1 transporters is thus an important consideration for overexpression studies. ABCG1 is widely expressed in tissues, though highest levels of expression are observed in spleen, lung, thymus and brain 89. ABCG1 has been most extensively characterized in macrophages, and numerous gene transcripts have been described arising from alternative splicing events or the use of different transcription initiation sites 86. Although ABCG1 is highly expressed in brain, the functions of ABCG1 in the CNS are not well characterized. 1.4.3. ABCG1 function and regulation Like several other members of the ABCG family, ABCG1 functions in the regulation of sterol homeostasis and in HDL metabolism. The established role of ABCG1 is to 18 facilitate the efflux of cellular cholesterol to lipidated apolipoproteins such as HDL, although the mechanism by which ABCG1 facilitates this transport remains elusive. As a result of the importance of sterol efflux in atherosclerosis, the majority of studies have focused on ABCG1 in the periphery. In macrophages, ABCG1 is thought to redistribute cholesterol to a cell-surface pool where it is accessible for removal by HDL particles 89 . Export of cholesterol from lipid-laden macrophages is thought to be the rate-limiting step in preventing atherosclerotic lesions in the vessel wall. Overexpression of Abcgl in cultured cells has been shown to promote cholesterol efflux to HDL, but not to lipid-poor apolipoproteins 90 . ABCG1 deficiency impairs efflux of cholesterol to HDL, and causes accumulation of cholesterol and neutral lipids in a variety of tissues in Abcgl knockout animals fed a high-fat diet 92. In light of these studies, ABCG1 seems to play a critical role in maintaining overall lipid homeostasis in tissues. ABCG1 mRNA is highly induced by the conversion of macrophages to lipid-laden foam cells, as well as by treatment with agonists of the Liver X Receptor (LXR), transcription factors that regulate a network of genes involved in lipid homeostasis 86, 93. LXR transcription factors regulate lipid homeostasis by binding oxysterol ligands (for example, 24S-hydroxycholesterol) in the cytosol and then translocating to the nucleus where they bind LXR response elements (LXREs) and promote transcription 94. Importantly, synthetic LXR agonists have been observed to reduce A13 production, amyloid burden and cerebral inflammation in AD mice 95 ' 96. Since Abcgl and a number of other genes are highly induced by LXR agonists, an understanding of the role of these genes in mediating that effect is important for AD therapeutics. Recently, ABCG1 has also been shown to regulate processing of the sterol regulatory element binding protein (SREBP-2), a transcription factor involved in regulation of 19 cholesterol synthesis. Under low sterol conditions, SREBP-2 is processed and then translocates to the nucleus, where it activates target genes and increases cholesterol synthesis 88. Tarr et al. demonstrated a link between overexpression of ABCG1 and increased SREBP-2 processing and target gene activation, suggesting that functional ABCG1 decreases flux through the cholesterol biosynthetic pathway 97 . 1.4.4. Human disease and ABCG1 Unlike other ABC transporters, a disease associated with ABCG1 has not yet been firmly established. Due to its regulatory role in sterol homeostasis, ABCG1 function has been implicated in both obesity and atherosclerosis. Interestingly, recent studies in Abcgl knockout mice showed that gene deletion markedly reduced size of fat cells in these animals and also protected them against diet-induced obesity 98. Massive accumulation of neutral lipids was observed in macrophages of mice with disrupted ABCG1, revealing an important role for functional ABCG1 in preventing excessive uptake of cholesterol, which can lead to foam cell formation M . The human Abcgl gene maps to chromosome 21, and is expected to be overexpressed by 1.4-2 fold in DS individuals with a third copy of chromosome 21 99' 188. Studies of individuals with DS have revealed some evidence of dysregulated lipid metabolism, though many of these studies have used very small cohorts and questionable controls. Fetuses with trisomy 21 were shown to have abnormalities in lipid metabolism resulting in hypercholesterolemia 101 . Another study showed a significantly higher prevalence of cholelithiasis in children with DS, 4.7% as compared with 0.2% for unaffected children 182 . Cholelethiasis is relatively uncommon in pediatric patients, and refers to the development of biliary gallstones associated with high plasma cholesterol levels. Though not conclusive, these findings suggest that DS individuals may display abnormal 20 cholesterol metabolism, and that genes on chromosome 21 such as ABCG1 may be associated with this phenotype. 1.5. Research Rationale and Hypothesis 1.5.1 Rationale Due to its location on chromosome 21 and its function in the maintenance of cholesterol homeostasis in the cell, ABCG1 is a candidate gene that may participate in the early development of AD neuropathology in DS individuals. Tansley et al. observed that both ABCG1 mRNA and protein levels are significantly increased in DS cortex, confirming that our gene of interest is, in fact, overexpressed in DS individuals. Experiments done by Tansley et al. also revealed that expression of ABCG1 enhances Af3 production in transfected HEK cells, and this increase was only observed with expression of a functional ABCG1 cholesterol transporter 1°0. Conversely, Kim et al. found reduced Af3 levels in CHO cells expressing ABCG1 183. These disparate in vitro findings demand an in vivo resolution where appropriate physiology is maintained. Though the effects of ABCG1 expression on A13 production are unclear, these findings provide a compelling rationale for investigating ABCG1 as a potential contributor to AD pathophysiology in DS individuals. Further, an important study by Kotti et al. demonstrated that cholesterol turnover in the brain is required for learning, as mice deficient in 24-hydroxylase do not show long-term potentiation (LTP) and are profoundly impaired in spatial learning tasks 58. These results suggest sterol homeostasis is of central importance to normal learning and memory, and indicate a potentially important role for the cholesterol-related functions of ABCG1 in cognitive processes. 21 1.5.2. Experimental hypothesis and objectives The hypothesis of this thesis is that ABCG1 contributes to the early onset of AD neuropathology and cognitive deficits observed in some DS individuals. Specifically, we hypothesize that overexpression of ABCG1 will accelerate the onset or progression of AD in vivo by causing more severe cognitive impairments in murine models. Specific objectives 1. To characterize baseline phenotype and cognitive profile of ABCG1 BAC Tg mice. 2. To determine whether cognitive deficits in the PDAPP mouse model of AD are altered by ABCG1 overexpression. 3. To determine the impact of ABCG1 deficiency on behaviour and cognition in mice. 4. To assess any differences in underlying neuronal and dendritic morphology in ABCG1 overexpressing mice. 22 2. MATERIALS AND METHODS 2.1. Transgenic Animals 2.1.1. 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 PCR. The validated BAC was then microinjected into F1 C57BI/6/CBA murine oocytes, and founder animals were identified by human specific PCR. 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. Braydon Burgess established the initial cohorts of ABCG1 BAC transgenic mice, and I subsequently generated experimental cohorts for behavioral analyses. Behavioural analysis cohorts included a total of 22 female mice at 12 months of age, with 11 ABCG1 BAC Tg mice and 11 wild-type littermate controls. 2.1.2. PDAPP Mice PDAPP mice were generously donated by Dr. Ronald DeMattos, Lilly Research Laboratories, Indianapolis, Indiana USA. PDAPP mice express a human APP mini-gene under the control of the PDGF-I promoter that promotes expression of the transgene in hippocampal and cortical neurons 44. The APP mini-gene encodes the FAD associated Indiana mutation (V717F) in addition to APP introns 6-8 to produce APP isoforms 695, 751 and 770 through alternative splicing. PDAPP mice are maintained by brother-sister matings as transgenic homozygotes on a mixed genetic background derived from C57/BI6, DBA and Swiss-Webster origins. 23 To establish experimental cohorts, heterozygous ABCG1 BAC transgenic males were crossed to PDAPP APP homozygous females to produce (ABCG1 BAC+ / APP+) and (ABCG1 wt, APP-) progeny. Female progeny were sacrificed at weaning to generate male cohorts for aging experiments. Cohorts of PDAPP/ABCG1 BAC Tg mice were set up by Braydon Burgess, and I subsequently managed cohorts selected for behavioral analyses. Final cohorts selected for behavioural analyses included 48 male mice at 12 months of age, with 25 PDAPP/ABCG1 mice and 23 PDAPP littermate controls. 2.1.3. ABCG1-deficient mice Abcg/-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 Internal Ribosome Entry Site -LacZ-Neo-pA cassette that places the 13- galactosidase gene under the control of endogenous Abcgl regulatory elements. Homologous recombination results in the deletion of 7 amino acids (GPSGAGK) 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. Experimental cohorts were generated by Braydon Burgess, and I subsequently selected and managed cohorts for behavioral analyses. 2.1.4. Diet Mice were maintained on a standard chow diet (PMI LabDiet 5010, containing 24% protein, 5.1% fat, and 0.03% cholesterol). 24 2.1.5. Acclimatization, enrichment and light cycle Two weeks prior to the start of testing, animals were transferred from their original housing facility to the behavioural testing facility for acclimatization. Mice were individually housed upon re-location to the testing facility to control for any home cage- specific influences, and all mice were provided identical enrichment materials (tubing and paper towel) in their individual cages. Mice were maintained on a 12h dark/12h reverse light cycle, and all testing was performed during the dark cycle, when the animals are naturally active. 2.2. Behavioural Testing I was solely responsible for the care and maintenance of the experimental colonies, set up of apparatus and administration of all behaviour tests. 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 2. 2.2.1. Testing conditions Changes in the pre-testing environment can have drastic effects on animal performance on behavioural tasks. To avoid extra stressors that may affect performance, only the experimenter, both prior to, and during the testing window, handled the mice. Cages were cleaned once a week following a testing session, to minimize effects of stress on performance in daily behavioural tasks. Perfumes and other strongly scented products were avoided when handling mice. 25 2.2.2. SHIRPA Phenotypic assessment of ABCG1 BAC Tg mice was carried out using the SHIRPA protocol 104 to screen for any differences in baseline phenotype, as these could markedly affect performance on cognitive tests. SHIRPA is a standard assessment tool, comprised of three stages. The first two provide a detailed general phenotype, including a behavioural observation profile, as well as assessment of motor behaviours and food and water intake. The third stage provides a more specialized screen for neurological deficits. 2.2.3. Open Field Test Locomotor activity, exploratory behaviour and general anxiety were analyzed in mice using the open field test. A circular arena with a diameter of 90 cm was set up directly underneath an overhead digital camera (Logitech QuickCam Pro 5000) in a room in an undisturbed, quiet location. The arena was brightly lit from above, and arena zones were defined using ANY-mazeTM computer software (Stoelting Co). A room divider shielded the experimenter and computer from the animal's view for the duration of the task. Animals were brought into the testing room, and given 5-10 minutes to adjust to the room's environment in their home cages. Each mouse was placed in the arena individually, and allowed to freely explore for 5 minutes, while its activities were tracked and recorded by the camera and ANY-mazer"". Upon completing the task, the mouse was removed from the arena by the experimenter and returned to the home cage. The number of defecations in the arena was recorded for each mouse, and the arena was cleaned with a mild soap following each test to avoid transfer of olfactory cues between animals. 26 Distance traveled, average speed and path length serve as measures of locomotor activity and exploratory behaviour. Anxiety is demonstrated by the time spent around the edges of the arena relative to the time spent in the centre, and also by freezing and defecation. 2.2.4. Elevated Plus Maze Emotionality was specifically measured using the elevated plus maze, a standard test of anxiety and fear. The apparatus consisted of an elevated maze with 2 open arms and 2 enclosed arms, set up in a room in an undisturbed, quiet location (Figure 4A). The maze was set up directly underneath an overhead camera (Logitech QuickCam Pro 5000), and zones were defined using ANY-mazeTM computer software. The experimenter and computer were behind a room divider, hidden from the animal's view for the duration of the task. Animals were brought into the testing room, and given 5-10 minutes to adjust to the room's environment in their home cages. Each mouse was initially placed in the centre of the maze, and allowed 5 minutes of free exploration. The animal's activities were tracked and recorded by the camera and ANY-maze -rm. Upon completion of the task, the mouse was returned to its home cage and number of defecations was recorded before the maze was cleaned with a mild soap prior to the next trial. Anxiety-related behaviour relates to the avoidance of the open arms of the maze, freezing/rearing and defecation. 2.2.5. Morris Water Maze Learning and memory was assessed using several variations of the Morris Water Maze task (Figure 4B). The water maze consisted of a circular white pool (100 cm diameter, 54 cm height) filled with water to a height of 36 cm. The water was rendered opaque 27 using white, non-toxic tempera paint. A circular platform (14 cm diameter, 35 cm height) was submerged in the pool (about 1 cm below the surface), providing a ledge upon which the mice could step to escape from the water. For each trial, a mouse was released into the pool at one of three release positions and allowed to search for the platform. Each trial lasted a maximum of 60 seconds, after which the mouse was manually guided to the platform and allowed to remain on the platform for another 10 seconds. All trials were tracked using an overhead digital camera (Logitech QuickCam Pro 5000) and ANY-maze r"" computer software. Visible platform task The visible platform task is used to screen for any visual deficits, as well as any motor problems (such as inability to swim) that may hinder an animal's ability to complete water maze training trials. The platform was located about 2 cm above the surface of the water, and a visually conspicuous marker, such as a small coloured flag, was attached to the top of the platform (Figure 4C). The platform location was changed for each of the 4 trials. Mice were released into the pool and allowed 60 seconds to locate the platform, while being tracked and recorded by ANYmazeTM. The experimenter noted any apparent deficits in vision or motor behaviours. Spatial reference memory task The reference memory task was used to assess learning and long-term memory for a fixed platform location (Figure 4D). Distinct geometric shapes were attached to the walls of the pool on three sides, and the experimenter and computer system were hidden behind a dividing wall throughout the experiment to avoid influence of extraneous cues. Mice were trained for 4 trials per day, for a total of 5 days. 24 hours after the last training session, a probe trial was administered, in which the platform was removed from 28 the pool and the mouse was allowed to swim freely for the full 60 seconds (Figure 4E). The amount of time spent in the quadrant of the pool that had previously contained the platform was measured, as an indication of memory retention. The following day, training sessions resumed with the platform in a new location, opposite to the original location. Mice were trained on this new location for 3 more days (12 trials total for each mouse). 24 hours following the last training session a second a probe trial was administered. All groups of mice received a total of 36 training trials and 2 probe trials each during the reference memory task. Spatial working memory task This task was used to assess any deficits in working memory, or ability of the mice to use information relevant to the task while performing the task. The pool set-up was as described above, with the exception of platform location, which changed to a new location on a daily basis. Mice were given 2 trials per day, for 4 days. A decrease in latency to the platform on the second trial of the day was an indicator of working memory, as the mice would use the visual cues to re-locate the platform more quickly the second time. Each mouse received a total of 8 training trials during the spatial memory task. 2.3. Golgi-Cox Staining 2.3.1. Preparing Golgi-Cox solution Golgi-Cox solution was prepared with a 5:5:4 ratio of the following components respectively: 5% potassium dichromate solution, 5% mercuric chloride solution and 5% potassium chromate solution. The solution was mixed in a glass flask covered with aluminum foil to avoid exposing the solution to light. After five days, the solution was filtered to remove any red precipitate that had formed and stored at room temperature until use. 29 2.3.2. Sacrifice and brain collection Animals were deeply anesthetized by intraperitoneal injection of 600mg/kg Avertin (Sigma-Aldrich), prepared according to manufacturer's instructions. Anesthetized mice were perfused transcardially with a 0.9% saline solution for 10 minutes. Whole brains were immediately transferred to a foiled 50mL conical tube containing 15mL of Golgi- Cox solution. Tubes were stored at room temperature for 24 hours, after which the solution in each conical tube was replaced with fresh Golgi-Cox solution. Brains were stored in the dark at room temperature for 14 days. Following this, the Golgi-Cox solution was emptied from each conical and whole brains were blotted gently with a Kimwipe. Conical tubes were rinsed out with distilled water and refilled with a 30% sucrose solution, upon which the brains were returned to their tubes. Tubes were stored in the dark for several days until the brains were no longer floating and had sunk to the bottom of the tube, indicating they were ready for sectioning. 2.3.3. Brain sectioning Whole brains were removed from sucrose solution and allowed to drain for 5 minutes on paper towel. Cerebellar and prefrontal regions were cut off evenly, using a non-feather blade. Brains were affixed to the vibratome platform, using the cut cerebellum side as a base, and then lowered into a chilled 6% sucrose solution in the stage compartment. Vibratome settings were adjusted to cut 200 p.m brain sections. Initial slices were discarded until the third ventricle was prominent and the hippocampus began to appear, after which slices were carefully transferred to a large petri dish containing 6% sucrose solution. Care was taken to minimize the amount of light exposure during this process, as Golgi solution is very light sensitive. 30 2.3.4. Slide preparation and staining Brain slices were transferred to 2% gelatinated slides, ensuring that the concave side of the slice was facing down on the slide. Slides were labeled with a histapen, and covered with a piece of weigh paper. Gentle pressure was applied evenly to all brains on the slide, and the weigh paper was carefully removed. Slides were then placed in the water incubator at 25-30°C for 30 minutes, and the incubator lid was covered with a black cloth to minimize light exposure. Immediately following incubation, slides were transferred to slide boats for the staining procedure. Fresh staining solutions were prepared in Wheaton jars, and slides were placed into each solution in the following order, for the corresponding amount of time. Solution Contents Time Notes 1 2 distilled water 100% ammonium hydroxide 1 minute 30 minutes wrap Wheaton jar in foil to avoid light exposure 3 4 distilled water KodaFix solution 1 minute 30 minutes wrap Wheaton jar in foil to avoid light exposure 5 distilled water 1 minute 6 50% ethanol 1 minute 7 70% ethanol 1 minute 8 95% ethanol 1 minute 9 100% ethanol 5 minutes 10 100% ethanol 5 minutes 1/3 ethanol 11 1/3 HemoDe/CitriSolv 5 minutes 1/3 chloroform 12 100%HemoDe/CitriSolv 15 minutes clearing solvent 13 100%HemoDe/CitriSolv 15 minutes clearing solvent Table 1. Protocol for preparation of Golgi-stained slides. All solutions were made fresh in Wheaton jars immediately prior to beginning staining protocol. Solutions were discarded and replaced with fresh ones following staining of 2-3 full slide boats. 31 Slides were covered with Permount and cover-slipped immediately after removal from the last solution. Sections were allowed to dry completely on open trays in a dark cupboard before being stored in a closed slide box. 2.3.5. Dendritic analysis Slides were coded and analysis was performed with the experimenter blinded to the genotypes of the original mice during analysis. All slides were viewed and hand drawn under the 40x objective lens on the microscope (Nikon Eclipse E600). Cells were only included in the analysis if the cell body and dendrites were clearly stained, and easily distinguishable from the processes of other surrounding cells (see Figure 5 for sample drawings). The aim was to draw 20 cells in total from the brain of each animal, with 10 cell representations from each of the inner and outer granule cell layers. However, if cells did not meet the criteria of being clearly visible and distinguishable from other cells, they were not drawn and included in the analysis, thus resulting in slightly different total cells drawn for each animal. The number of primary dendrites, or dendritic processes extending directly from the cell body, was counted for Golgi-impregnated cells in the dentate gyrus. A series of concentric rings, spaced 15 pm apart, was placed over each cell drawing, centered on the cell body (Figure 5C). Sholl analysis was conducted by recording the number of bifurcations per ring, as a measure of branching complexity. 2.4. Statistical analysis Data were analyzed by repeated-measures analysis of variance (ANOVA) for between- group differences, and followed up with Student's t tests, performed by GraphPad Prism (version 4.0) software. All in vivo data were conducted with the rater blinded to genotype. 32 3. RESULTS 3.1. ABCG1 BAC transgenic mice 3.1.1. ABCG1 is overexpressed in brain ABCG1 is shown to be overexpressed in the brain of ABCG1 BAC transgenic mice, with protein levels in transgenic animals between 3-8 fold higher than that observed in wild- type littermate control animals. The level of overexpression varied with brain region, with greater overexpression in the cortex and hippocampus as compared to the cerebellum (data not shown). This verified that our transgene does in fact significantly increase ABCG1 protein levels in the brain over endogenous levels. Braydon Burgess performed the western blots to verify that ABCG1 overexpression did in fact occur in the animal line to be used for these studies. 3.1.2. ABCG1 BAC Tg mice show normal baseline behaviour A cohort of 22 mice, (n=11 ABCG1 BAC Tg, n=11 wild-type) were assessed using the categories described by the SHIRPA protocol 104. We used this protocol to test whether overexpression of ABCG1 resulted in any gross phenotypic abnormalities. Mice were all female, and were approximately 10 months of age (±0.45) at the time of observation. Because ABCG1 BAC Tg mice had not displayed gross changes in activity and behaviour based upon general observations in their home cage environments, we expected most SHIRPA measurements to be normal, indicating that excess ABCG1 does not generate a gross metabolic disturbance that will dramatically impact behavioural results. Observations made using the SHIRPA protocol showed that animals expressing the ABCG1 BAC transgene had a similar baseline phenotype to wild-type littermates (Figure 6). ABCG1 BAC Tg mice did not display any abnormalities in physiological measures such as heart rate, body tone and grip strength or in behavioural measures such as irritability, aggression and exploratory activity. Since 33 ABCG1 BAC Tg mice have a baseline phenotype comparable to wild-type animals, these factors should have no effect on performance and results of further behavioural tests should be an accurate reflection of cognitive function. 3.1.3. ABCG1 overexpression has no effect on anxiety or general locomotor activity ABCG1 BAC Tg mice were next evaluated for anxiety level and general locomotor activity, as both factors can interfere with learning tasks, and therefore can drastically affect performance during cognitive assessment. Mice were approximately 12 months old (±0.45) at the time of testing, and cohorts included 11 ABCG1 BAC Tg mice and 11 wild-type littermate controls. The single-trial Open Field Test revealed no difference between groups in time spent in different locations in the arena. Both ABCG1 BAC Tg and wild-type mice spent the majority of their time exploring the edges of the open arena, and generally avoided the brightly lit centre of the arena. No differences were observed when mice were also analyzed for average speed in the arena, distance travelled and time spent immobile or frozen in the arena (Figure 7A), indicating that ABCG1 BAC Tg mice have normal exploratory and locomotor behaviour and do not show increased anxiety levels compared to wild-type animals. The Elevated Plus Maze, a behavioural task specifically designed to evaluate anxiety, was administered to verify that anxiety levels are not altered in ABCG1 BAC Tg mice. Wild-type and ABCG1 BAC Tg mice spent comparable amounts of time in the different locations within the maze, with both groups displaying preference for the closed arms of the maze (Figure 7B). There were no differences between groups in average speed or distance traveled in the maze. Time spent immobile, an indicator of anxiety, was also similar for both groups. 34 3.1.4. Overexpression of ABCG1 alone does not affect learning and memory Mice expressing the ABCG1 BAC transgene were next evaluated against wild-type littermates on the Morris Water Maze spatial reference memory task. The task consisted of a 5-day acquisition phase in which mice were given 4 trials per day and a maximum time of 60 seconds per trial to locate a platform fixed in the North quadrant. Twenty-four hours later, a probe trial was administered to test recall of the platform location. ABCG1 BAC Tg mice displayed nearly identical average escape latency, or time to locate the platform, over each day of acquisition as compared to wild-type animals, and the learning curve was very similar (Figure 8A), indicating that ABCG1 overexpression does not significantly affect learning during the acquisition phase of this water maze task. Repeated-measures ANOVA revealed a significant effect of training day, (p < 0.0001), but not group (p = 0.67). Performance on the single-trial probe test was also comparable between groups, with both wild-type and ABCG1 BAC Tg mice spending the greatest amount of time searching in the probe quadrant (North), which previously contained the platform during training days (Figure 8B). Both groups of mice displayed strong preference for the North quadrant, indicating that both wild-type and ABCG1 BAC Tg mice were capable of learning the task thoroughly, and were able to remember visual cues associated with the platform location during training trials. Additionally, the comparable performance of ABCG1 BAC Tg and wild-type mice on both learning during the acquisition phase or on memory recall in the probe trial, also suggests that overexpression of ABCG1 alone does not negatively affect hippocampal neuron function involved in the learning and memory required for this task. 3.2. ABCG1 and Alzheimer's Disease We next evaluated ABCG1 overexpression in a mouse model of Alzheimer's Disease. The PDAPP mouse is a well-established model of amyloid formation, developing 35 amyloid plaques similar to human AD pathology, as well as deficits in learning and memory, around 12 months of age 45. The reason for evaluating ABCG1 overexpression in an Alzheimer's model is supported by several factors. First, the ABCG1 gene is located on chromosome 21, which is triplicated, and verified to be overexpresed in DS 100. As almost 100% of DS individuals show early onset of neuropathological hallmarks of AD by the age of 40, we hypothesized that ABCG1 overexpression may have a role in accelerating the onset of AD in these individuals. In addition, in vitro experiments from our laboratory demonstrated that overexpression of ABCG1 increases levels of Ar3, the neurotoxic peptide species found in AD brains 1°° . In the following experiments, I aimed to clarify the in vivo effect of ABCG1 overexpression on behaviour, as well as to determine if it contributed to, or perhaps even accelerated, cognitive decline in an AD mouse model. A large cohort of PDAPP and PDAPP/ABCG1 mice was generated by cross-breeding with ABCG1 BAC Tg mice. Groups for behaviour were chosen at approximately one year of age, consistent with the literature showing that both plaque deposition and cognitive deficits are well-established in PDAPP mice at this age 105 . Cohorts of 23 PDAPP and 25 PDAPP/ABCG1 were selected for behavioural analysis. 3.2.1. ABCG1 overexpression does not alter anxiety or general locomotor activity in PDAPP mice Short behavioural tests were administered to 12-month old PDAPP and PDAPP/ABCG1 BAC Tg male mice, to assess anxiety level and general locomotor activity, to rule out effects of these factors on learning and memory, and also to familiarize the mice with handling. Mice were approximately 12 months old (±0.35) at the time of testing, with cohorts of 25 PDAPP/ABCG1 mice and 23 PDAPP littermate controls. The Open Field 36 Test revealed no difference between groups on time spent in different locations in the arena, and minimal values for standard error (Figure 9A). Both groups of mice spent more time exploring the edges of the open arena, and generally avoided the brightly lit centre of the arena. I observed no significant differences between groups when I analyzed average speed in the arena, distance travelled and time spent immobile or frozen in the arena (Figure 9A), indicating that PDAPP mice overexpressing ABCG1 have normal exploratory and locomotor behaviour, and do not show increased anxiety levels as compared to PDAPP mice. To confirm that PDAPP/ABCG1 mice had comparable anxiety levels to PDAPP mice, a single-trial Elevated Plus Maze was administered. Both groups of mice displayed a strong preference for the closed arms of the maze, but there was no significant difference in time spent in different locations of the maze due to expression of the ABCG1 BAC transgene (Figure 9B). There were also no significant differences between groups in average speed or distance traveled in the maze, as well as in time spent immobile during the test. Interestingly, these tasks did reveal some differences between PDAPP animals (Figure 9) and the wild-type animals assessed in the first group of experiments (Figure 7). PDAPP mice explored both the open field and elevated plus maze at a slower speed and thus also covered less distance than wild-type animals. PDAPP mice also showed increased time spent immobile on both tasks, indicating that the AD animals were more anxious than the wild-type. PDAPP animals also spent less time exploring the open arms of the elevated plus maze, choosing to spend the majority of their time in the enclosed arms. The AD phenotype appears to increase anxiety levels in mice, similar to anxiety symptoms observed in AD patients. In humans, increased anxiety may manifest in up to 70% of AD patients during the course of the disease, and is considered a noncognitive behavourial alteration of AD 9. Interestingly, although increased anxiety is one of the 37 most prevalent symptoms of human AD and negatively affects quality of life for AD patients, it has received far less attention than cognitive impairments. Increased anxiety in AD patients is associated with impairments in daily living activities, just as increased anxiety can interfere with task learning in animals. 3.2.2. Determination of a suitable cohort for water maze task Before beginning the acquisition phase of the Morris Water Maze, PDAPP and PDAPP/ABCG1 mice were first administered a visual test in latency to locate a visually conspicuous platform marked with a flag. The visual water maze task revealed no apparent deficits and no significant differences in performance between the two groups (Figure 10). Repeated-measures ANOVA revealed no significant effect, either by group (p = 0.8) or by quadrant (p = 0.58), indicating that the animals do not have a pre-existing bias to one of the four quadrants which could affect water maze results. During the visual test, mice were also screened for inappropriate swimming behaviours in the water maze such as spinning, thigmotaxis (or the tendency to remain close to walls) around the pool perimeter, and floating, and any mice consistently displaying these behaviours were removed from the study. Approximately 5% of the mice from each group (n=2 PDAPP, n=1 PDAPP/ABCG1) were removed from further study and were not included in the analysis. 3.2.3. Overexpression of ABCG1 did not affect performance on the spatial reference memory water maze task Deficits in spatial learning have previously been characterized in the PDAPP mouse model, and are apparent even in PDAPP mice younger than 6 months of age, prior to full onset of A(3 neuropathology 48. Several groups have also shown that PDAPP mice exhibit an age-dependent worsening in spatial learning on water maze tasks 48 ' 105. The water maze is an extremely effective tool in detecting deficits in spatial learning, and has 38 been used and modified by many groups to evaluate cognitive decline in AD model mice. Here I used a variation of the water maze to assess spatial reference learning and memory in both PDAPP mice and PDAPP mice overexpressing ABCG1. On the spatial reference memory task (fixed location/hidden platform), both groups of mice showed faster latency to locate the platform after several days' training, with the learning curve leveling off after the fourth day (Figure 11A). Both PDAPP and PDAPP/ABCG1 mice also displayed a decrease in swim path length after 5 days of training (Figure 11 B), indicating that they had learned the location of the platform. However, performance on the task did not differ significantly between the two groups. Repeated-measures ANOVA revealed a significant effect of training day (p < 0.0001), but not group (p = 0.14). Another important parameter to assess is swim speed in the water maze, as speed can have a great effect on escape latency values. As shown in Figure 11 C, there was no difference in swim speed between the two groups, although a decrease in mean swim speed over the duration of the training trials was observed. In the probe trial (Figure 11D), both groups of mice displayed a strong preference for the training quadrant (quadrant that contained the platform during training trials), showing evidence of spatial memory for the platform location. Though both PDAPP and PDAPP/ABCG1 mice preferred the training quadrant, no significant difference existed between the groups on the amount of time spent searching each of the quadrants during the probe trial. Although no differences were observed between PDAPP groups with normal and elevated ABCG1 levels, it is interesting to note that the PDAPP mice in general showed poorer performance on the task than the wild-type mice that were previously tested (Figure 8, Figure 11A and 11D). This indicates that the PDAPP mice do show cognitive 39 impairments on the water maze at 12 months of age, which is consistent with reports in the literature on these animals 45' 48, 105. But more importantly, this observation validates that the variation of the water maze task that was used in the present study is sensitive enough to detect the expected cognitive deficits in symptomatic PDAPP mice. 3.2.4. Performance on the spatial working memory water maze task was not affected by overexpression of ABCG1 Mice were administered a 4-day spatial working memory task, consisting of two trials per day in which the animals had to locate a submerged platform. Animals were required to learn a new platform location each day with the location remaining constant over the two trials for a given day. Performance on the second trial of the day was expected to improve, indicating that mice had learned the relevant information (the new platform location) during the first trial for that day. Figure 12 (A through D) shows latency to find the platform for each of the 4 days of testing. Analysis of the data using repeated- measures ANOVA did not reveal any significant effect by group (all p values > 0.1), but did show significant differences by trial (p values < 0.05). Both groups of mice appear to have learned the task by the second day, as they take less time to locate the platform on the second trial. However, there was no significant difference between the performance of PDAPP and PDAPP/ABCG1 animals, suggesting that overexpression of ABCG1 does not affect working memory in this specific water maze task. Figure 12E represents mean performance on Trials 1 and 2 over the 4-day duration of the task. There was no significant difference in performance between the two groups on either of the trials, although the PDAPP/ABCG1 group displayed a greater decrease in latency to the platform on the second trial. 40 3.2.5. ABCG1 overexpression does not influence performance of PDAPP mice on behavioural tests Overexpression of ABCG1 in PDAPP animals did not affect general exploratory behaviour and locomotor activity, nor did it change the levels of anxiety in these mice as indicated by the results of the Elevated Plus Maze (EPM). Performance of PDAPP/ABCG1 mice in the water maze was also comparable to PDAPP mice that did not express the ABCG1 BAC transgene. Although I did not directly compare wild-type and PDAPP mice in the same experiment, I did observe that PDAPP animals exhibited the expected impaired performance on this task compared to previously tested wild-type mice. The PDAPP AD animals performed worse than the wild-type mice at a similar age both during the learning phase (training trials) and on memory recall (probe trial). From this I can conclude that the water maze task used in the present study was valid, as it was sensitive enough to detect the expected AD-related cognitive deficits in the PDAPP animals. Although the data conclusively show no effect of ABCG1 overexpression on cognition in both wild-type and PDAPP mice, the possibility that excess ABCG1 may affect other neuronal functions that impact learning and memory cannot be ruled out. Though the water maze task proved to be valid for detecting deficits due to the PDAPP phenotype, it is possible that the changes caused by ABCG1 overexpression are too subtle to be detected by this task. Further behavioural tests are needed to confirm that ABCG1 overexpression has no effect on learning and memory. 3.3. ABCG1-deficient mice Although the previous experiments revealed no behavioural phenotype for ABCG1 BAC transgenic mice, others in the Wellington lab have shown that ABCG1 overexpression 41 does cause changes at the cellular level 188. Compared to wild-type animals, the levels of cholesterol precursors in the ABCG1 BAC transgenic brain are lower. Most notably, the cholesterol precursors lathosterol, lanosterol and desmosterol were reduced, and a similar trend was observed with 24S-hydroxycholesterol, a product of cholesterol catabolism 187. This suggests an important role for ABCG1 in maintaining homeostasis of cholesterol metabolism within the brain. Tall et al. also conclude that ABCG1 has a physiological role in mediating efflux of cholesterol and biosynthetic intermediates from cells 93 . Cholesterol metabolism appears to be an important component of the learning process at a cellular level, as suggested by the findings that brain cholesterol turnover is a required process for learning in mice 58. Russell et al. demonstrated that mice lacking 24-hydroxylase, an enzyme responsible for cholesterol catabolism, show severe deficiencies in spatial, associative and motor learning, as well as deficits in LTP 58. It seems that intact cholesterol metabolism pathways are essential for the learning process to occur, and given that ABCG1 is a key player in cholesterol metabolism in the brain, this brings to light the question of how deficiency of ABCG1 might affect learning. Mice deficient in ABCG1 have been reported to exhibit impaired cholesterol efflux to HDL, elevated plasma apoE levels, and accumulation of lipids in tissues (particularly in lung) when fed a high-fat, high-cholesterol diet 108, 109. In the brain, ABCG1 deficiency was recently reported to increase the levels of desmosterol, a sterol intermediate in cholesterol synthesis 93. However, to our knowledge no one to date has investigated the cognitive function of ABCG1-deficient mice. In this preliminary study, I used the water maze to assess learning and memory in ABCG1 deficient (-/-), hemizygous (+/-) and wild-type (+1+) animals. Gene dose does not always linearly correspond to phenotype 42 because of compensatory pathways that might operate, so it was important to determine the behavioural phenotypes of each of the three groups of animals. 3.3.1. ABCG1-deficient mice do not express ABCG1 protein ABCG1 deficiency in ABCG1 knockout mice was verified using Western blot analysis. ABCG1 -'- mice had a complete absence of ABCG1 protein expression, while hemizygous animals (+/-) showed some expression, but markedly less than wild-type (+/+) mice (data not shown). Western blots were performed by Braydon Burgess. 3.3.2. Loss of ABCG1 expression tends to worsen probe trial performance To determine if ABCG1 deficiency has an effect on learning and memory, three groups of mice were evaluated using a 5-day spatial reference memory water maze task. The groups had different levels of ABCG1 expression, consisting of wild-type (+/+) mice, mice hemizygous for ABCG1 (+/-) and ABCG1-deficient (-/-) mice. Animals were all female with an average age of 4.3 months ((±0.73), and cohorts consisted of 10 wild- type, 10 hemizygotes and 11 ABCG1-deficient animals. Groups were evaluated for escape latency (time to locate platform) as well as for swim speed and path length, or swimming distance in the maze. Repeated-measures ANOVA revealed a significant effect of training day (p = 0.006), but not group (p = 0.12). There was no significant difference in performance of the three groups on any of training days during the acquisition period (Figure 13A-C), indicating that ABCG1 expression levels do not influence day-to-day learning of the task. On the probe test for memory recall, however, ABCG1 deficiency appeared to worsen recall for the location of the platform learned during the acquisition phase (Figure 13D). ABCG1-deficient mice spent comparably less time searching in the training quadrant 43 that had previously contained the platform, and crossed over the exact location of the platform less often than did the other two groups. Mice hemizygous for ABCG1 appeared to perform at a level intermediate to the wild-type and ABCG1-deficient animals, crossing over the exact platform location less often than wild-type animals but more often than the ABCG1-deficient mice (Figure 13E). Though significance was not achieved when comparing ABCG1-deficient mice to the wild-type mice in this experiment, the p value was approaching significance at 0.08. If this preliminary experiment were repeated with a larger cohort of animals, the data suggest it is likely that significance would be achieved, at least between the performance of the wild-type and ABCG1 -'- animals. It is important to note that the animals tested in this experiment were all female, so it would be interesting to assess male mice for any sex-related differences in performance. Also, these results were obtained in young 4-5 month old mice, therefore it would be very interesting to observe the performance of aged ABCG1-deficient mice on this same task to see if any significant differences exist. Lastly, future experiments should include a spatial working memory water maze task, as well as perhaps tests for other types of learning such as associative and motor learning. 3.4. Morphological analysis of Golgi-stained neurons Although the behavioural experiments revealed no significant difference in learning and memory due to overexpression of ABCG1, it still cannot be concluded that ABCG1 has no effect on cognition. Though dramatic changes at the neuronal level are usually reflected in behaviour, it is possible that ABCG1 overexpression may cause only subtle changes and that the behavioural tests used in these studies are not sensitive enough to detect these changes. To this end, I investigated the effects of ABCG1 overexpression at a neuronal level using Golgi staining methods to visualize whole cells and their 44 dendritic trees. Morphological changes in dendrite branching are thought to underlie learning and memory processes, and increased dendritic complexity is associated with synaptic plasticity in the hippocampus. Cohorts included 5 mice for each of the following groups: wildtype, ABCG1 BAC Tg, PDAPP and PDAPP/ABCG1 mice. Mice were all male with an average age of 12 months, and approximately 8-10 whole cells were selected for drawing from the brain of each animal. Brains selected for Golgi staining were from mice with no prior testing history and were not selected from cohorts that had been trained on behavioural tasks, as these would not be representative of baseline measures. I examined Golgi-stained neurons from two regions of the dentate gyrus of the hippocampus, the inner and outer granule layers (Figure 14). The dentate gyrus is a major site of neurogenesis in the adult brain 110, 111, and thus neurons in this region play an important role in new learning and the formation of new memories. I assessed the neurons for number of dendrites as well as branching complexity and length. Complexity of dendritic arborization is thought to correspond to learning and memory capacity, as shown by studies demonstrating the powerful positive influence of exercise and environmental enrichment on neuronal morphology 112, 113, and additionally, studies show that chronic stress negatively affects dendritic branching 114. Dendritogenesis, or the process by which dendrites form and grow, is directed by chemical signals within the environment of the neuron 115 . The physical act of growing occurs when vesicles travel down microtubules and exocytose at the end, fueling the growth of the branch with plasma membrane. Cholesterol is an important component of the plasma membrane, and ABCG1 expression could potentially affect branching because of its role in cholesterol metabolism and homeostasis within the cell. 45 3.4.1. ABCG1 overexpression does not affect number of primary dendrites, branch order analysis or length to branch point To assess effect of ABCG1 overexpression on underlying neuronal morphology, brain slices from wild-type, ABCG1 BAC Tg, PDAPP and PDAPP/ABCG1 mice were prepared using the Golgi-Cox method. Slides were analyzed for the number of primary dendrites, length to branch point and branch order analysis. Sholl analysis was conducted to assess branching complexity for both inner and outer granule layer cells of the dentate gyrus in the hippocampus. Those dendrites extending directly from the cell body were counted and recorded as primary dendrites. No significant differences in the number of primary dendrites were found between groups (Figure 15A and B), indicating that expression of ABCG1 does not influence the initial formation of the dendritic tree. The length of each primary dendrite was measured, from where it originates at the cell body to the first branch point. This measurement was taken to represent the average length of the primary dendrites for each cell, as after the first branch point it was impossible to determine which branch was the extension of the original primary dendrite. As shown in Figure 15D, there was a significant difference (p=0.053) in dendrite length in the IGZ region, with PDAPP/ABCG1 dendrites appearing significantly shorter than PDAPP dendrites. However, due to limited mouse availability and a resulting low number of PDAPP/ABCG1 primary dendrites measured (4) as compared to PDAPP dendrites (19), these data are not conclusive. In general, no significant differences were observed between the mean lengths (Figure 15C and D) of the primary dendrites from both the inner and outer granule zones, though there does appear to be a trend toward decreased length in the ABCG1-overexpressing animals. 46 Branch order analysis was also used to assess dendritic arborization. Dendrites originating from the cell body were labeled as primary dendrites, those originating from the first branch point were labeled as secondary, and so on. There were no significant differences between the wild-type and ABCG1 BAC Tg animals (Figure 16A and C) and also between the PDAPP and PDAPP/ABCG1 animals (Figure 16B and D). The general trend in these results points to smaller values for the higher branch orders in the animals overexpressing ABCG1, regardless of AD phenotype. This suggests that perhaps ABCG1 overexpression, while having little effect on the initial dendrite formation, may become a factor in the later stages of dendritogenesis during the formation of higher order branches. It is important to note the large standard deviation values of some of these measurements, particularly for cells in the inner granule zone area, indicating that while these trends may cause interesting speculation, further analysis and a larger number of cells are required to draw any solid conclusions. 3.4.2. Overexpression of ABCG1 results in less complex dendritic branching patterns Sholl analysis was conducted as a means to quantify dendritic complexity. A series of concentric rings, spaced 15 Jim apart, was placed over the cell tracing, centered over the cell body (Figure 5C). The number of branches visible as a function of distance (by each concentric ring) was recorded. Figure 17 depicts the results of Sholl analysis of Golgi- stained neurons in the inner and outer granule layers of the dentate gyrus. Analysis of dendritic complexity revealed that animals overexpressing ABCG1 show significantly less dendrite-ring intersections, particularly from the 4 th through 10 th rings, as compared to wild-type and PDAPP animals that do not overexpress ABCG1. This observation also lends support to the general trend observed in the branch order analysis (Figure 16), suggesting that ABCG1 overexpression results in less branched dendrites and thus, in decreased dendritic complexity. Interestingly, an effect of AD phenotype is also 47 observed in this analysis, with the dendrites of PDAPP animals extending over less distance than those of wild-type or ABCG1 BAC Tg mice. Figures 17A and 17B both represent dendritic branching as a function of distance, however, differences due to the AD phenotype of PDAPP animals are evident when comparing the two figures. Figure 17B shows both a smaller number of dendrites, as well as generally shorter dendritic trees indicated by less total intersections with the concentric rings, as compared to the dendrites from mice without the AD phenotype represented in Figure 17A. This suggests that PDAPP animals have less complex dendritic branching patterns than wild- type mice, a result that might be expected based on the PDAPP mouse model's previously established cognitive deficits. Again, the fairly large standard deviation values in these measurements must be noted, indicating that these experiments should be extended to verify the results. 48 4. DISCUSSION AND FUTURE DIRECTIONS 4.1. Discussion 4.1.1. ABCG1 overexpression and cognition Using ABCG1 BAC Tg mice, I investigated the effects of ABCG1 overexpression on both baseline behavioural phenotype and cognition. ABCG1 BAC Tg mice were indistinguishable from wild-type animals when assessed using the SHIRPA protocol, indicating that overexpression of ABCG1 does not result in any gross metabolic disturbances that could affect observable behavioural or physiological measures. Cognitive and noncognitive measures were also unaffected in these mice. The performance of ABCG1 BAC Tg mice on tasks measuring anxiety, general locomotor activity and spatial learning and memory were entirely comparable to wild-type animals. Notably, the results obtained from these behavioural and cognitive tests on ABCG1 BAC Tg animals are highly reproducible, with minimal standard deviation between individual values (Figures 7-8). These data strongly suggest that overexpression of ABCG1 alone is not sufficient to cause changes in cognition. However, ABCG1 overexpression does cause changes in the brain at the cellular level, affecting the cholesterol biosynthetic pathway. Though total brain cholesterol levels were unaffected, overexpression of ABCG1 resulted in decreased levels of sterol precursors and 24S-hydroxycholesterol in the brain, suggesting that ABCG1 suppresses flux through the cholesterol biosynthetic pathway 107. Kotti et al. established that changes in cholesterol flux have an effect at the level of cognition, as CYP46-deficient mice have greatly reduced flux in brain cholesterol biosynthesis, and are profoundly impaired in spatial learning and memory 58 . However, the present study revealed no effect of ABCG1 overexpression at the behavioural level. A possible explanation for these findings is the existence of a compensatory pathway that acts to balance out the 49 effects of ABCG1 overexpression and maintain normal cognitive functions in the brain. The ABCG4 transporter is highly homologous to ABCG1. ABCG4 is thought to interact with ABCG1 in heterodimers, and is co-expressed with ABCG1 in neurons and astrocytes 97 . Moreover, ABCG4 is proposed to be important in cholesterol homeostasis, although the role of ABCG4 is unclear. Tarr et al. recently reported that Abcg/-deficient mice and Abcg4-deficient mice both show repression of SREBP-2 target genes involved in cholesterol synthesis, suggesting that ABCG4 does function in regulation of cholesterol levels. Unlike ABCG1 however, expression of ABCG4 is not regulated by the LXR family of transcription factors, and may constitute an alternate constitutive pathway by which cholesterol homeostasis is maintained 97. It is a possibility that the ABCG4 pathway compensates for imbalances caused by overexpression of ABCG1, preventing major changes at the level of cognition. However, further studies are required to determine the exact functions of ABCG4, and whether it is indeed part of a pathway that could compensate for alterations in ABCG1 function. It is also an important consideration that the changes caused by overexpression of ABCG1 may be too subtle to have globally obvious effects on cognition. Changes at the neuronal level due to ABCG1 overexpression may be below the detection threshold of the behavioural and cognitive tests used in our experiments. Additionally, it is possible that different brain pathways, perhaps those associated with noncognitive measures, are affected by overexpression of ABCG1, in which case our behavioural tests may not be suitable to detect these changes. 4.1.2. Overexpression of ABCG1 in PDAPP mice Spatial learning deficits have previously been characterized in PDAPP mice as young as 4-5 months of age, though an age-related deterioration occurs following substantial A13 50 deposition and produces more severe deficits in mice at 10-12 months of age 48 ' 188. The experiments carried out in the present study did reveal the expected cognitive deficit in PDAPP mice at 12 months of age, when compared to the performance of age-matched wild-type mice on the same water maze task in the first set of experiments (Figure 8 and Figure 11). This finding suggests that the water maze paradigm used was effective because it was sensitive enough to detect an expected deficit. However, it is important to consider that cognitive deficits in PDAPP animals have been shown to exist from a very young age and may be quite pronounced by the age of 12 months, thus these deficits might be detected very easily by the water maze tasks. Although these water maze tasks were able to detect an obvious deficit, they may not be sensitive enough to detect very subtle changes in cognitive function. PDAPP mice overexpressing ABCG1 performed similarly to PDAPP controls with normal ABCG1 expression, on both water maze tasks administered. Spatial reference memory was first assessed using the classic version of the Morris Water Maze, where mice are trained on a fixed location, hidden platform. This test paradigm provides an assessment of long-term memory, in which mice must learn to associate visual cues with the platform location in training trials over many days. This task is analogous in humans to an individual learning the location of the parkade where they park their car each day. The parkade will obviously be in a fixed spot, and though the individual may take different routes when returning to the parkade on any given day, they are still able to use visual cues to locate the parkade. The water maze task was also adapted to create a spatial working memory task, which involved learning a new platform location on a daily basis. This protocol was similar to the one used by Chen et al., though abbreviated, and required mice to selectively retrieve memory encoded for the new platform location each day 48. This task aims to test the 'episodic-like' component of memory, as the mice must 51 demonstrate day-specific learning; finding the platform on the first trial and remembering that location for the second trial. In humans, this would be analogous to an individual remembering the specific location of their car in a parkade, as it is likely to be parked in a different location each day. Interestingly, Chen et al. found that impairments in working or 'episodic-like' memory in PDAPP mice were age-dependent, whereas deficits in spatial reference memory were already present and did not correlate with age 48. The spatial working memory water maze task did not reveal any significant differences between PDAPP and PDAPP/ABCG1 mice, though the day-to-day performance of both groups appears quite variable, a factor that could possibly be eliminated by the administration of more trials per day. 4.1.3. Benefits and limitations of behavioural testing Although behavioural tests like the Morris Water Maze are widely used and successfully adapted for testing learning and memory in various capacities, these tasks do have limitations in their application and relevance to human disease. Episodic memory is the first, and most severely affected type of memory in humans with AD 11 , and thus, specifically testing this type of memory is an important consideration for behavioural AD studies in mice. Episodic memory in humans is a type of declarative memory, or memory that can be consciously discussed or declared. Whether mice displaying recall of their learning on a spatial working memory task is indicative of 'episodic-like' memory is a difficult question to answer. In human AD, the medial temporal lobe (MTL) containing the hippocampus (Figure 14) is the most severely affected brain region. Spatial navigation in mice is known to be hippocampal-dependent 118, and thus the Morris Water Maze task seems to be an appropriate test of hippocampal function. Moreover, humans with temporal lobe damage display severe impairments in learning and memory, including the recall of spatial locations and in solving spatial maze tasks 11 , 52 indicating that spatial learning and memory in humans is also localized to the hippocampus and MTL. The object recognition task (ORT) is another task that is considered a measure of hippocampal function, and is becoming a popular task in assessing cognitive function of AD mouse models. The ORT is based on spontaneous exploration of novel and familiar objects by rodents, and measures whether they can recall objects previously presented to them in a training trial. Normal animals will spend more time exploring novel objects, rather than objects they have previously seen and explored. Recognition memory, as measured by this task, is thought to consist of two components, a recollective component (episodic) and a familiarity component 117. Thus, the ORT does provide a measure for 'episodic-like' memory in animals, though it is difficult for this task to distinguish between the components of recognition memory. However, rodents with large hippocampal lesions have shown clear deficits on the ORT 117  suggesting this task may become an important addition to behavioural testing in our AD mouse models. Just as no animal model is able to fully replicate all aspects of a human disease, behavioural and cognitive tests in animal models have their limitations in application to human diseases. However, these tests can provide us with vital insights into human diseases, making them invaluable tools for research. 4.1.4. ABCG1 and dendritogenesis Although no abnormal behavioural phenotype was observed in ABCG1 BAC Tg mice, Sholl analysis of dendritic branching revealed less complex branching patterns of neurons from ABCG1-overexpressing brains. The number of dendrites per ring intersection was significantly decreased in the outer rings (5 to 11), particularly in 53 ABCG1 BAC Tg animals, although a similar trend was observed for PDAPP animals overexpressing ABCG1 (Figure 17). What accounts for the less complex dendritic branching patterns in ABCG1-overexpressing animals? Recent studies have shown that both overexpression and deficiency of ABCG1 do not change total brain cholesterol levels, although cholesterol metabolism is affected in both cases, including levels of cholesterol precursors and catabolites 93' 187. Although ABCG1 has an established role in cholesterol efflux, ABCG1 may also act as a sterol sensor that detects low levels of cholesterol within the cell and activates transcription factors such as SREBP-2 to induce cholesterol biosynthesis 97 . The majority of cholesterol in the brain is sequestered in myelin, therefore small changes in membrane cholesterol levels may have large effects on neuronal structure and function. Adult neurons synthesize cholesterol at a low basal rate, and as such, often rely largely on glial cells to meet their lipid requirements for axonal growth processes and synaptogenesis 118. Dendritic branching is a complex process, regulated by chemical signals and extracellular factors 115. Elongation of dendrites and formation of new branches are membrane-extending processes, and so require fusion of vesicles containing membrane components 115. Given that cholesterol is an important plasma membrane component, it is possible that neurons from ABCG1-overexpressing brains are unable to develop extensive dendritic branches because insufficient cholesterol is available for membrane extension. Lipid rafts, which contain sphingolipids and cholesterol, are thought to play an important role in axonal growth, guidance and synaptic transmission 119. Lipid rafts may also play an important role in dendrite growth, although axons have been the predominant focus of most studies to date. It is interesting to note that cholesterol depletion disturbs lipid rafts 119,120, thus it is possible 54 that under ABCG1-overexpressing conditions, lipid rafts may not be enriched in cholesterol and therefore may not be able to participate in dendritogenesis. Changes in levels of cholesterol precursors and metabolites in the brain could account for the altered dendritic morphology of ABCG1-overexpressing mice, though it is curious that this is not accompanied by detectable corresponding phenotypic changes or cognitive deficits in these animals. Since other genes, such as ABCG4, may also participate in regulating cholesterol levels, it is possible that these other pathways compensate for changes in ABCG1 expression and prevent manifestation of cognitive deficits, as previously suggested. These alternate pathways warrant investigation, since little is known about the function of ABCG4. Neurons from the dentate gyrus (DG) of the hippocampus were used for morphological analysis using Golgi methods, as the DG is thought to be important in learning and memory processes 121 . The DG is one of the few places where neurogenesis occurs in the adult brain, and altered neurogenesis is associated with altered hippocampal plasticity 122. Synaptic plasticity, or changes in the efficiency of synapses, is thought to underlie the learning process and to be required for memory storage. Accordingly, cells in the DG show efficient synaptic strengthening, known as long-term potentiation (LTP). Our colleagues in the Christie lab have assessed ABCG1 BAC Tg animals for any changes in LTP. Stimulation of Schaeffer axon collaterals that project from CA3 pyramidal cells to the CA1 region of the hippocampus elicited similar field excitatory postsynaptic field potentials (fEPSP) in both ABCG1 BAC Tg and wild-type animals (Appendix I), indicating that ABCG1 overexpression does not affect LTP. In addition, no significant difference in paired-pulse facilitation was observed between ABCG1 BAC Tg mice and wild-type littermates both prior to, and after stimulation (Appendix I), 55 suggesting that the presynaptic properties of ABCG1-overexpressing mice are intact. Overall, these results demonstrate that ABCG1 BAC Tg mice have comparable LTP to wild-type mice, further reinforcing the behavioural results that determined ABCG1 overexpression does not affect learning and memory processes. 4.1.5. ABCG1 deficiency and cognition Investigation of the effects of ABCG1 overexpression revealed no differences at the behavioural level, but changes in the cholesterol biosynthetic pathway were observed on the cellular level. To ascertain the effects of altering ABCG1 expression and to further clarify the role and functions of ABCG1, we investigated the effect of ABCG1 deficiency at both behavioural and cellular levels. It is important to note that hemizygotes (+/-) were also assessed, along with ABCG1-deficient (-/-) animals and wild-type controls. This is an important consideration, as gene dose does not always correspond to phenotype. The pathways by which genes are transcribed into mRNA and subsequently translated into proteins are complex and tightly regulated on multiple levels, thus it is important to determine the effect of gene dose as well. A deficiency of ABCG1 had the opposite effect to ABCG1 overexpression on cholesterol precursors in the brain, reportedly increasing levels of desmosterol 93. Correspondingly, Burgess et al. found that ABCG1-deficient mice displayed increases in the cholesterol precursors lathosterol, lanosterol and desmosterol, which interestingly, were increased in a gene dose- dependent manner 107 . Although total cholesterol levels in the brain were unchanged by both overexpression and deficiency of ABCG1, levels of the cholesterol catabolite 24S- hydroxycholesterol were increased by 20% in ABCG1-deficient mice 107 . These findings demonstrate that ABCG1 deficiency alters cholesterol metabolism and biosynthesis in the brain, reinforcing the important role of ABCG1 in maintaining cholesterol 56 homeostasis and turnover, a process known to be vital to normal learning and memory processes ". Behavioural testing of ABCG1-deficient mice on the water maze reference memory task yielded an interesting result. Loss of ABCG1 expression appeared to worsen recall for the location of the platform on the probe trial (Figure 13E) in a gene dose-dependent manner. ABCG1-deficient mice crossed over the exact location of the platform less frequently than both wild-type and hemizygous (+/-) animals, and the data almost achieved significance (p=0.08) when comparing ABCG1 -/- to wild-type. However, it is interesting that no deficits were apparent on the training trials, as both ABCG1 -/- and +/- mice displayed a learning curve similar to wild-type mice (Figure 13A). This observation suggests that ABCG1-deficient mice were capable of learning normally during training trials, but had difficulty recalling the task after a delay, since the probe trial was a single trial administered 24 hours following the last set of training trials. It is important to note that this study was conducted with young, 4-5 month old ABCG1-deficient mice, thus these data cannot be directly compared from those obtained using ABCG1 BAC Tg mice tested at 12 months of age. Learning and memory deficits may be more pronounced in older animals, as aging often contributes to poorer performances on behavioural tasks such as the water maze 123 . Although not conclusive, this interesting finding suggests that ABCG1 is necessary for normal memory processes, and a deficiency of ABCG1 may disrupt cholesterol metabolism in a way that interferes with memory recall. In any case, the effects of ABCG1 deficiency on cognition merit further investigation. 57 4.2. Conclusions This research was based on the hypothesis that ABCG1 overexpression contributes to the early onset of AD neuropathology and cognitive symptoms observed in some DS individuals. I did not observe an effect of ABCG1 overexpression, alone or in conjunction with mutated hAPP in a mouse model, on behavioural parameters such as anxiety and general locomotor activity, or on learning and memory as measured by behavioural and cognitive tasks. These observations were reinforced by the findings of our colleagues in the Christie lab, who determined that LTP in ABCG1 BAC Tg brain slices is comparable to wild-type (Appendix I). However, a significant effect of ABCG1 overexpression was observed at the cellular level when dendritic branching patterns of DG neurons were evaluated, suggesting that ABCG1 does play an important role in cholesterol metabolism and dendrite growth. Additionally, ABCG1 deficiency resulted in difficulties in memory recall on the water maze probe trial, suggesting that alternate pathways that could make up for changes caused by excess ABCG1 may not be able to compensate for a total deficiency. Taken together, this research concludes that ABCG1 overexpression is unlikely to contribute to the accelerated onset of AD pathology in DS individuals, even though ABCG1 does play an important role in maintaining cholesterol homeostasis. 4.3. Future Directions The conclusion of this study is that ABCG1 overexpression alone is unlikely to play an important role in the accelerated onset of AD pathology observed in DS individuals, and does not promote AD-like cognitive deficits in a mouse model. However, other genes on chromosome 21 are overexpressed in DS individuals, including genes critical for neuronal growth and development, and synapse maintenance 124. It remains a possibility that any of these genes may synergize with excess APP to accelerate the 58 onset of AD pathology. Additionally, it is possible that the accelerated onset of AD in DS individuals is a result of a combination of several overexpressed genes, or their effects on downstream pathways. As such, it is necessary to investigate the effects of other genes with altered expression in DS, in a similar fashion to this ABCG1 study, to systematically determine and rule out the roles of these genes in the early development of AD pathology. With this study, we have shown the ABCG1 overexpression, both alone and with excess APP, does not cause AD-like cognitive deficits in a mouse model and thus is unlikely to contribute to early development of AD pathology. However, the results of this study raise other significant questions about the role of ABCG1 in cholesterol metabolism, and the importance of cholesterol turnover in cognitive processes. The roles of ABCG1 and also of other cholesterol-related genes in the brain definitely merit further study. 4.3.1. Behavioural characterization of ABCG1-deficient mice A comprehensive behavioural and cognitive profile for ABCG1-deficient mice was not determined in this thesis, but is an important next step in an ongoing study as suggested by preliminary water maze data (Figure 13). Large cohorts of young and aged ABCG1- deficient mice could be analyzed using the Morris Water Maze variations described here, along with other tasks such as the ORT. In addition, tests of fear, anxiety and other behaviours could be used to determine if ABCG1-deficient mice have any sort of alterations or damage in hippocampal, amygdalar or prefrontal cortex areas. This research would conclusively answer whether ABCG1 deficiency has a significant impact on learning and memory, and may provide new insight into how the cholesterol metabolism functions of ABCG1 are related to cognition. 59 4.3.2. Investigate the effects of ABCG4 and ABCG1 on cognition Little is known about the functions of ABCG4, though it is proposed to be similar to ABCG1 and also to form heterodimers with ABCG1. This thesis did not investigate the effect of altering ABCG4 expression on cognition, though the results obtained here and in other recent studies 97 suggest that ABCG4 may be part of an alternate, non-LXR dependent pathway involved in the regulation of cholesterol levels. The evaluation of learning and memory in ABCG4 -/-, as well as in ABCG1-/- / ABCG4 -/- double null mice using water maze tasks and the ORT could provide a clearer understanding of the roles of these two genes and their interactions. 4.3.3. Conduct morphological analysis of ABCG1- and ABCG4-deficient dendritic trees The dendritic patterns of ABCG1-deficient, and ABCG4-deficient mice were not analyzed in this study, but could provide useful data on the effects of these genes on dendrite development and the potential effects of cholesterol availability on aspects of this process. A large cohort of animals should be evaluated to ensure that standard deviation values are minimal. 60 Amyloid Precursor Protein Cysteine Acidic^11- a-^7-rich il ivi I^I^1^I^I N 1 Cytosolic C terminus a-sec etase^13-secr tase sAPPa sAPPO I y-secretase y-secr tase •  p3 fragment AR fragment NON-AMYLOIDOGENIC PATHWAY AMYLOIDOGENIC PATHWAY Figure 1. Processing of the amyloid precursor protein. Full length APP undergoes cleavage at the plasma membrane by either a -secretase or R- secretase, then is cleaved by y-secretase to produce non-pathogenic p3 fragments or AR fragments, respectively. ** Location of the Indiana V717F mutation. 61 1 I 110 CYP46A 1 KO CYP46 -› learning and memory deficits i Geranylgeranyl-PP i Summary of the Cholesterol Biosynthetic Pathway HMGR 2 Acetyl CoA ---■ Mevalonate -■ Isopentenyl-PP i  T Statins Geranyl-PP 1 Cholesterol ^ Squalene ,___ Farnesyl-PP 24S-hyd roxycholesterol Geranylgeranylated Proteins (ie. Rho) Figure 2. Summary of the cholesterol biosynthetic pathway. Two Acetyl CoA molecules are converted to mevalonate by HMGR, in the first and rate limiting step of cholesterol biosynthesis. The mevalonate pathway produces cholesterol and nonsterol isoprenoids such as farnesyl diphosphate and geranylgeranyl diphosphate. In the brain, cholesterol is converted by Cyp46A1 to 24S-hydroxycholesterol, a compound which then is able to diffuse across the BBB for elimination through bile. This cholesterol turnover is necessary for learning, as CYP46-deficient mice show profound learning and memory deficits 58 . 62 Walker A&B ATP Binding Domains • Varia. ble Region Figure 3. Predicted topology of the ABCG1 monomer. Each ABCG1 monomer is predicted to encode 6 trans- membrane alpha-helices, and the Walker A and Walker B ATP binding domains. Monomers must dimerize to form functional transporters. 63 A Almr \IPA 111J Visual test with ^ Training trials with ^ Probe trial without marked platform hidden platform platform Figure 4. Behavioural testing apparatus and methodology. (A) Elevated plus maze apparatus, with 2 open arms and 2 enclosed arms. Mouse is placed in the centre section and allowed to freely explore for 5 minutes. (B) Morris Water Maze pool apparatus and set up. (C) Visible platform task to assess visual proficiency. Mouse swims to find a visually conspicuous platform marked with a flag. (D) Training trials. Mouse searches for hidden platform using visual cues located around the pool, then climbs onto it. Location of platform remains constant, start position of mouse varies each trial. (E) Probe trial for memory of platform location. Platform is removed and mouse swims freely for 60 seconds. 64 ••••\ ti• N 6 Figure 5. Cell drawings of Golgi-stained neurons and dendrites, viewed under 40x objective lens. Representative cell drawings of (A) neuron from outer granule zone (OGZ) layer, and (B) neuron from inner granule zone layer (IGZ), both visualized under 40x objective and 10x zoom. From brain of ABCG1 BAC Tg mouse. (C) Concentric rings used for Scholl analysis, spaced 15 jim apart and centered over the soma. Number of bifurcation/ring intersections was measured for cells in the IGZ and OGZ. A 65 SHIRPA Primary Screen  INN Wild-type = ABC G1 O E .0 c.) Figure 6. ABCG1 mice show normal baseline phenotype. SHIRPA primary screen on ABCG1 and wild-type mice. Primary screen involves physiological profiling of each mouse, using a number of test categories and assigning a rating (defined by the SHIRPA protocol found at mmary.html) for each mouse in each category. Cohorts included n=11 mice per genotype. The mice observed were all female, and had an average age of 10 months. All p values are non-significant. 66  • • •..• • •• • • WT •••••• • •: a on OD 0-----On 0 ABCG1 VVT • • ABCG1 Do • • • VVT ABCG1 A ^ Open Field Test ^ B ^ Elevated Plus Maze Edges of arena Centre of arena Open Arms Closed Arms Centre W 0.07 . -E 0.0^• • -0 0.0 • o.... o.• 0. c .O CO a o. 0. WT •• .• •• WT • •.• ••••• • ♦ •• •• •• co 0.1 E 0.1 -o 0.1 CD 0.13 fa 0.12 U) 0.11 C • 0.1 0. 0.00 a) C.)C a) LES 0 E .E a) E 0 00 O ABCG1 0 on 0 Doo no 0 ABCG1 0 0000 OCI 00 013 0 ABCG1 Figure 7. ABCG1 overexpression does not alter anxiety or general locomotor activity. ABCG1 BAC Tg (n=11) and wild-type mice (n=11) were assessed for any differences in anxiety or locomotor activity using the Open Field Test (A) and the Elevated Plus Maze (B). Data from each test were analyzed to include measures of time spent in maze locations, mean speed and distance and freezing, or time spent immobile in the maze. All p values are non-significant. 67 B .,... .12 -0 .- ._ w Eta a) .E 1 E E 3 Cl) Probe quadrant SENW SENW Day MN VVild-type CI ABCG1 Figure 8. Performance of 12-month old ABCG1 BAC Tg mice on the spatial reference memory water maze task. 11 animals were assessed per genotype (n=11). (A) Latency to platform. Each point represents an average of 4 daily trials during the training period. No significant difference was observed between the groups in the time taken to find the hidden platform. (B) Probe trial results from a single-trial test. Both control and ABCG1 over-expressing groups showed similar preference for the North (N) quadrant, which contained the platform on training trials. All p values are non-significant. 68 Edge ^ Centre 300 250 - 200 • 200 U) a) 150 • E 100 • 100 50 • nir-i.ji^10 0 e 1=3 PDAPP Imo PDAPP/ABCG1 Open^Closed^Centre Arms^Arms I I 0.100 0.075 0.050 0.025 0.000 ABCG1- 301 20 10 0^ ABCG1- 0.035 0.030 0.025I 0.020 0.015 0.010 0.005 0.000 r 10.0 7.5 5.0 2.5 0.0 1501 100 50 0^ ABCG1+ ABCG1+ AB6G1- ABCG1+ ABCG1+ ABCG1- A^Open Field Test^ Elevated Plus Maze Figure 9. PDAPP mice overexpressing ABCG1 show no differences in anxiety or locomotor activity. PDAPP (n=23) and PDAPP/ABCG1 mice (n=25) were assessed for any differences in anxiety or locomotor activity using the Open Field Test (A) and the Elevated Plus Maze (B). Data from each test were analyzed to include measures of time spent in maze locations, mean speed and distance and freezing, or time spent immobile in the maze. All p values are non-significant. 69 11111PDAPP N 30-, `1C4 20- w 0 >, 10- C -J o Trial 1 [ J 1 I Trial 2 Trial 3 Trial 4 PDAPP/ABCG1 Figure 10. No visual deficits revealed by water maze visual task. The single-day task consisted of 4 trials, in which PDAPP and PDAPP/ABCG1 mice had to locate a visually conspicuous platform that was moved to a new location on each trial. Maximum trial time was 60 seconds. All p values are non-significant. 70 MN PDAPP I=1 PDAPP/ABCG1+ Day ^ Day ^ DaY D Figure 11. Overexpression of ABCG1 in PDAPP mice did not affect performance on the spatial reference memory water maze task. (A) Latency to platform. Each point represents an average of 4 daily trials during the training period. No significant difference was observed between the groups in the time taken to find the hidden platform. However, ABCG1 BAC+ mice appeared to learn the task faster for the first 3 days, with the trend leveling out by the fourth day of training. (B) Swim distance and (C) swim speed were analyzed for any differences which might affect performance. (D) Probe trial results from a single-trial test. Both PDAPP and PDAPP/ABCG1 groups showed similar preference for the probe quadrant, which contained the platform on training trials. All p values are non-significant. 71 O 0 t _Ta.^71; a. ri o o ›...^ ›.c..) 0e ca) 1 m 1 1; ta'_1^ _i A Day 1 ^ B Day 2 ^ C Day 3 Trial ^ Trial ^ Trial D Day 4 E -16- PDAPP --0- PDAPP/ABCG1 OM PDAPP ED PD4PP/ABCG1 Trial  Figure 12. Performance of PDAPP mice on spatial working memory water maze task was not affected by overexpression of ABCG1. Graphs (A) through (D) represent daily latency to locate the platform. The platform is moved to a novel location at the beginning of each day and the location is held constant over the duration of two trials for each given day. (E) shows an average of trials 1 and 2, over the 4 days of testing for both PDAPP and PDAPP/ABCG1 groups. There was no significant difference in performance between the two groups, although the PDAPP/ABCG1 group displayed a greater difference in latency to locate the platform between the two trials. All p values are non-significant. 72 tsi^E S^W Quadrant p = 0.08 p = 0.21 +/- 0^ E 7 (6) 6 C 5 0 4 0 E 3 0 2cB 0- 1 0 D 1 A^ B^ C  12. E 10. c 7. to -0 5. E 0. E co O (.)a 2 ^ 2 ^ 2 Day Day Day Figure 13. Loss of ABCG1 expression tends to worsen water maze probe trial performance. A cohort including 10 wild-type, 10 ABCG1 heterozygotes and 11 ABCG1 knockouts was assessed using the Morris Water Maze. (A) Escape latency to locate platform. (B) Mean speed and (C) swim distance in the maze. (D) Probe trial to test recall of platform location. P values * = 0.12, ** = 0.09 (E) Number of crossings over the exact previous location of the platform during the probe trial. 73 C^Dentate gyrus A  Medial temporal lobe Hippocampus IGZ Figure 14. Hippocampal anatomy. (A) Medial temporal lobe. Structures include the hippocampus. (B) Hippocampus (C) Dentate gyrus of the hippocampus. GCL = granule cell layer. Outer (OGZ) and inner granule (IGZ) zones comprise the GCL. 74 OGZ IGZ OGZ IGZ NMI PLAPP n=6 G1 BAC n=5 IGZ PD6FP n=6 o PDAPPG1 130C n=5 OGZ ** A Figure 15. ABCG1 overexpression does not affect the number of primary dendrites or length to branch point. The dendrites of Golgi-stained neurons were analyzed from both wild-type and ABCG1 BAC Tg mice, as well as from PDAPP and PDAPP/ABCG1 mice. (A) Number of primary dendrites in inner and outer granule layers of the dentate gyrus of wild-type and ABCG1 mice. Numbers on bars represent number of cells assessed. (B) Primary dendrite count for brains of PDAPP and PDAPP/ABCG1 mice. Numbers on bars represent number of cells assessed. (C) Average length of primary dendrites measured to first branch point from brains of wild-type and ABCG1 mice. Numbers on bars represents number of primary dendrites measured. (D) Average length to branch point for dendrites of PDAPP and PDAPP/ABCG1 mice. Numbers on bars represents number of primary dendrites measured. P values * = 0.058, ** = 0.053 75 2^3 Branch Order PDAPP PDAPP/G1 BAC PDAPP -El- PDAPP/G1 BAC 2^3^4 Branch OrderC IGZ 6 a 45 4 t 3 112 1 0 -1111-G1 BAC- -e- G1 BAC+ -11-G1 BAC- -0- G1 BAC+ 2^3^4 Branch Order Figure 16. Branch order analysis is not significantly affected by overexpression of ABCG1. Dendritic arborization was assessed by labeling primary dendrites originating directly from the cell body, secondary dendrites emerging from the first branch point, and so on for both wild-type and ABCG1 mice, as well as PDAPP and PDAPP/ABCG1 mice. ABCG1 mice do not show signifcant overall differences in dendritic branch order as compared to wild-type mice in both the outer granule layer cells (A) and inner granule layer neurons (C). Branch order analysis of PDAPP and PDAPP/ABCG1 mice in the OGZ (B) and IGZ (D) revealed no significant differences between groups. * represents a p value of <0.05 76 —e— Wild-type —e—ABCG1 Tg * * * 1 2 3 4 5 6 7 8 9 10 11 12 13 14 * * * 8 9 10 11 12 13 14 B OGZ C 2 7oa) Sc cmc --*— PDAPP * -a-- PDAPP/ABCG1 * 3 4 5 6 7 8 9 10 11 12 13 14 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 D IGZ c 2 7 mmm m c ca c A OGZ Figure 17. Overexpression of ABCG1 results in less complex dendritic branching patterns. Dendritic arborization of Golgi-stained neurons was analyzed using Sholl analysis. The number of dendrites intersecting with each concentric ring was recorded and mean values were graphed for both outer and inner granule zone layers. (A) and (C) Number of dendrites/ring intersection for wild-type and ABCG1 BAC Tg animals, outer (OGZ) and inner (IGZ) zones respectively. (B) and (D) Dendrites per ring intersection for outer (OGZ) and inner (IGZ) granule layers of PDAPP and PDAPP/ABCG1 mice. * represents a p value of <0.05. 77 REFERENCES 1.Larson EB, Shadlen MF, Wang L, et al. Survival after initial diagnosis of alzheimer disease. Ann Intern Med. 2004;140:501-509. 2. Hebert LE, Scherr PA, Bienias JL, Bennett DA, Evans DA. Alzheimer disease in the US population: Prevalence estimates using the 2000 census. Arch Neural. 2003;60:1119-1122. 3. Birks J. Cholinesterase inhibitors for alzheimer's disease. Cochrane Database Syst Rev. 2006;(1):CD005593. 4. Hull M, Berger M, Heneka M. Disease-modifying therapies in alzheimer's disease: How far have we come? Drugs. 2006;66:2075-2093. 5. McKhann G. Criteria for the clinical diagnosis of alzheimer's disease. excerpts from the NINCDS-ADRDA work group report. J Am Geriatr Soc. 1985;33:2-3. 6. Li YJ, Scott WK, Hedges DJ, et al. Age at onset in two common neurodegenerative diseases is genetically controlled. Am J Hum Genet. 2002;70:985-993. 7. Higgins GA, Jacobsen H. Transgenic mouse models of alzheimer's disease: Phenotype and application. Behav Phannacol. 2003;14:419-438. 8.Schneider LS, Dagerman KS. Psychosis of alzheimer's disease: Clinical characteristics and history. J Psychiatr Res. 2004;38:105-111. 9. Robertson J, Curley J, Kaye J, Quinn J, Pfankuch T, Raber J. apoE isoforms and measures of anxiety in probable AD patients and apoe-/- mice. Neurobiol Aging. 2005;26:637-643. 78 10. Lopez-Pousa S, Vilalta-Franch J, Garre-Olmo J, Pons S, Cucurella MG. Characterisation and prevalence of the psychological and behavioural symptoms in patients with dementia. Rev Neurol. 2007;45:683-688. 11. Nestor PJ, Fryer TD, Hodges JR. Declarative memory impairments in alzheimer's disease and semantic dementia. Neuroimage. 2006;30:1010-1020. 12. Mickes L, Wixted JT, Fennema-Notestine C, et al. Progressive impairment on neuropsychological tasks in a longitudinal study of preclinical alzheimer's disease. Neuropsychology. 2007;21:696-705. 13. Levy JA, Chelune GJ. Cognitive-behavioral profiles of neurodegenerative dementias: Beyond alzheimer's disease. J Geriatr Psychiatry Neurol. 2007;20:227-238. 14. Barkhof F, Polvikoski TM, van Straaten EC, et al. The significance of medial temporal lobe atrophy: A postmortem MRI study in the very old. Neurology. 2007;69:1521-1527. 15. Bigler ED. The clinical significance of cerebral atrophy in dementia. Arch C/in Neuropsychol. 1987;2:177-190. 16. Goedert M, Jakes R. Mutations causing neurodegenerative tauopathies. Biochim Biophys Acta. 2005;1739:240-250. 17. Grundke-lqbal I, lqbal K, Tung YC, Quinlan M, Wisniewski HM, Binder LI. Abnormal phosphorylation of the microtubule-associated protein tau (tau) in alzheimer cytoskeletal pathology. Proc Nat/ Acad Sci U S A. 1986;83:4913-4917. 18. Guillozet-Bongaarts AL, Glajch KE, Libson EG, et al. Phosphorylation and cleavage of tau in non-AD tauopathies. Acta Neuropathol. 2007;113:513-520. 79 19. 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. 20. Lorenzo A, Yankner BA. Beta-amyloid neurotoxicity requires fibril formation and is inhibited by congo red. Proc Natl Acad Sci USA.  1994;91:12243-12247. 21. Fagan AM, Mintun MA, Mach RH, et al. Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans. Ann Neural. 2006;59:512-519. 22. Schmidt ML, Robinson KA, Lee VM, Trojanowski JQ. Chemical and immunological heterogeneity of fibrillar amyloid in plaques of alzheimer's disease and down's syndrome brains revealed by confocal microscopy. Am J Pathol. 1995;147:503-515. 23. Gellermann GP, Appel TR, Tanned A, et al. Raft lipids as common components of human extracellular amyloid fibrils. Proc Natl Acad Sci USA.  2005;102:6297-6302. 24. 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. 25. 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. 26. McGeer EG, Klegeris A, McGeer PL. Inflammation, the complement system and the diseases of aging. Neurobiol Aging. 2005;26 Suppl 1:94-97. 27. Gehrmann J, Matsumoto Y, Kreutzberg GW. Microglia: Intrinsic immuneffector cell of the brain. Brain Res Brain Res Rev. 1995;20:269-287. 80 28. McGeer PL, McGeer EG. Inflammation, autotoxicity and alzheimer disease. Neurobiol Aging. 2001;22:799-809. 29. Price DL, Tanzi RE, Borchelt DR, Sisodia SS. Alzheimer's disease: Genetic studies and transgenic models. Annu Rev Genet. 1998;32:461-493. 30. Tanzi RE, Bertram L. Twenty years of the alzheimer's disease amyloid hypothesis: A genetic perspective. Cell. 2005;120:545-555. 31. Cai XD, Golde TE, Younkin SG. Release of excess amyloid beta protein from a mutant amyloid beta protein precursor. Science. 1993;259:514-516. 32. 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. 33. 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. 34. 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 APP mutations linked to familial alzheimer's disease. Nat Med. 1996;2:864-870. 35. 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. 36. Tanzi RE. Molecular genetics of alzheimer's disease and the amyloid beta peptide precursor gene. Ann Med. 1989;21:91-94. 81 37. Moya KL, Benowitz LI, Schneider GE, Allinquant B. The amyloid precursor protein is developmentally regulated and correlated with synaptogenesis. Dev Biol. 1994;161:597-603. 38. 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. 39. Lichtenthaler SF, Haass C. Amyloid at the cutting edge: Activation of alpha-secretase prevents amyloidogenesis in an alzheimer disease mouse model. J Clin Invest. 2004;113:1384-1387. 40. He G, Qing H, Tong Y, Cai F, Ishiura S, Song W. Degradation of nicastrin involves both proteasome and lysosome. J Neurochem. 2007;101:982-992. 41. 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. 42. 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. 43. 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. 44. 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. 45. Spires TL, Hyman BT. Transgenic models of alzheimer's disease: Learning from animals. NeuroRx. 2005;2:423-437. 82 46. 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 Neunasci. 1997;17:7053-7059. 47. Eriksen JL, Janus CG. Plaques, tangles, and memory loss in mouse models of neurodegeneration. Behav Genet. 2007;37:79-100. 48. Chen G, Chen KS, Knox J, et al. A learning deficit related to age and beta-amyloid plaques in a mouse model of alzheimer's disease. Nature. 2000;408:975-979. 49. Goldstein JL, Brown MS. Regulation of the mevalonate pathway. Nature. 1990;343:425- 430. 50. 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. 51. Dietschy JM, Turley SD. Cholesterol metabolism in the brain. Curr Opin Lipidol. 2001;12:105-112. 52. Beffert U, Stolt PC, Herz J. Functions of lipoprotein receptors in neurons. J Lipid Res. 2004;45:403-409. 53. Reiss AB, Siller KA, Rahman MM, Chan ES, Ghiso J, de Leon MJ. Cholesterol in neurologic disorders of the elderly: Stroke and alzheimer's disease. Neunabiol Aging. 2004;25:977-989. 54. Snipes GJ, Suter U. Cholesterol and myelin. Subcell Biochem. 1997;28:173-204. 55. Wilson JD. The measurement of the exchangeable pools of cholesterol in the baboon. J Clin In vest. 1970;49:655-665. 83 56. 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. 57. Bjorkhem I, Diczfalusy U, Lutjohann D. Removal of cholesterol from extrahepatic sources by oxidative mechanisms. CU? Opin Lipidol. 1999;10:161-165. 58. Kotti TJ, Ramirez DM, Pfeiffer BE, Huber KM, Russell DW. Brain cholesterol turnover required for geranylgeraniol production and learning in mice. Proc Natl Acad Sci U S A. 2006;103:3869-3874. 59. 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. 60. 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. 61. 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. 62. Strittmatter WJ, Roses AD. Apolipoprotein E and alzheimer disease. Proc Natl Acad Sci USA.  1995;92:4725-4727. 63. 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. 64. Corder EH, Saunders AM, Strittmatter WJ, 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. 84 65. Dolev I, Michaelson DM. A nontransgenic mouse model shows inducible amyloid-beta (abeta) peptide deposition and elucidates the role of apolipoprotein E in the amyloid cascade. Proc Natl Aced Sci U S A. 2004;101:13909-13914. 66. Baum L, Chen L, Ng HK, Pang CP. Apolipoprotein E isoforms in alzheimer's disease pathology and etiology. Microsc Res Tech. 2000;50:278-281. 67. 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 U S A. 1999;96:15233-15238. 68. 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. 69. 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. 70. 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. 71. 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. 72. 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. 73. Yanagisawa K. Cholesterol and pathological processes in alzheimer's disease. J Neurosci Res. 2002;70:361-366. 85 74. Wolozin B, Kellman W, Ruosseau P, Celesia GG, Siegel G. Decreased prevalence of alzheimer disease associated with 3-hydroxy-3-methyglutaryl coenzyme A reductase inhibitors. Arch Neurol. 2000;57:1439-1443. 75. 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. 76. Shepherd J, Blauw GJ, Murphy MB, et al. Pravastatin in elderly individuals at risk of vascular disease (PROSPER): A randomised controlled trial. Lancet. 2002;360:1623-1630. 77. Sparks DL, Sabbagh M, Connor D, et al. Statin therapy in alzheimer's disease. Acta Neurol Scand Suppl. 2006;185:78-86. 78. 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. 79. Li L, Cao D, Kim H, Lester R, Fukuchi K. Simvastatin enhances learning and memory independent of amyloid load in mice. Ann Neural. 2006;60:729-739. 80. Rachidi M, Lopes C. Mental retardation in down syndrome: From gene dosage imbalance to molecular and cellular mechanisms. Neurosci Res. 2007;59:349-369. 81. Tanzi RE. Neuropathology in the down's syndrome brain. Nat Med. 1996;2:31-32. 82. Lott IT, Head E. Alzheimer disease and down syndrome: Factors in pathogenesis. Neurobiol Aging. 2005;26:383-389. 83. Head E, Lott IT. Down syndrome and beta-amyloid deposition. Cuff Opin Neural. 2004;17:95-100. 86 84. Prasher VP, Farrer MJ, Kessling AM, et al. Molecular mapping of alzheimer-type dementia in down's syndrome. Ann Neural. 1998;43:380-383. 85.Cabrejo L, Guyant-Marechal L, Laquerriere A, et al. Phenotype associated with APP duplication in five families. Brain. 2006;129:2966-2976. 86.Velamakanni S, Wei SL, Janvilisri T, van Veen HW. ABCG transporters: Structure, substrate specificities and physiological roles : A brief overview. J Bioenerg Biomembr. 2007;39:465-471. 87.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. 88.Dean M, Hamon Y, Chimini G. The human ATP-binding cassette (ABC) transporter superfamily. J Lipid Res. 2001;42:1007-1017. 89.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. 90.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. 91.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. 87 92. 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. 93. Wang N, Yvan-Charvet L, Lutjohann D, et al. ATP-binding cassette transporters G1 and G4 mediate cholesterol and desmosterol efflux to HDL and regulate sterol accumulation in the brain. FASEB J. 2007. 94. Kalaany NY, Mangelsdorf DJ. LXRS and FXR: The yin and yang of cholesterol and fat metabolism. Annu Rev Physiol. 2006;68:159-191. 95. Riddell DR, Zhou H, Comery TA, et al. The LXR agonist TO901317 selectively lowers hippocampal Abeta42 and improves memory in the Tg2576 mouse model of alzheimer's disease. Mol Cell Neurosci. 2007;34:621-628. 96. Lefterov I, Bookout A, Wang Z, Staufenbiel M, Mangelsdorf D, Koldamova R. Expression profiling in APP23 mouse brain: Inhibition of abeta amyloidosis and inflammation in response to LXR agonist treatment. Mol Neurodegener. 2007;2:20. 97. Tarr PT, Edwards PA. ABCG1 and ABCG4 are coexpressed in neurons and astrocytes of the CNS and regulate cholesterol homeostasis through SREBP-2. J Lipid Res. 2008;49:169-182. 98. Buchmann J, Meyer C, Neschen S, et al. Ablation of the cholesterol transporter adenosine triphosphate-binding cassette transporter G1 reduces adipose cell size and protects against diet-induced obesity. Endocrinology. 2007;148:1561-1573. 99. 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. 88 100. 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;48:1022-1034. 101. 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. 102. Toscano E, Trivellini V, Andria G. Cholelithiasis in down's syndrome. Arch Dis Child. 2001;85:242-243. 103. 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. 104. Rogers DC, Fisher EM, Brown SD, Peters J, Hunter AJ, Martin JE. Behavioral and functional analysis of mouse phenotype: SHIRPA, a proposed protocol for comprehensive phenotype assessment. Mamm Genome. 1997;8:711-713. 105. Hartman RE, Izumi Y, Bales KR, Paul SM, Wozniak DF, Holtzman DM. Treatment with an amyloid-beta antibody ameliorates plaque load, learning deficits, and hippocampal long- term potentiation in a mouse model of alzheimer's disease. J Neurosci. 2005;25:6213-6220. 106. 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;48:1022-1034. 107. Burgess BL, Parkinson PF, Racke MM, et al. ABCG1 influences brain cholesterol synthesis but does not affect amyloid precursor protein or apolipoprotein E metabolism in vivo. J Lipid Res. 2008. 89 108. 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. 109. Out R, Hoekstra M, Habets K, et al. Combined deletion of macrophage ABCA1 and ABCG1 leads to massive lipid accumulation in tissue macrophages and distinct atherosclerosis at relatively low plasma cholesterol levels. Arterioscler Thromb Vasc Biol. 2008;28:258-264. 110. Eriksson PS, Perfilieva E, Bjork-Eriksson T, et al. Neurogenesis in the adult human hippocampus. Nat Med. 1998;4:1313-1317. 111. Kornack DR, Rakic P. Continuation of neurogenesis in the hippocampus of the adult macaque monkey. Roc Natl Acad Sci U S A. 1999;96:5768-5773. 112. Dierssen M, Benavides-Piccione R, Martinez-Cue C, et al. Alterations of neocortical pyramidal cell phenotype in the Ts65Dn mouse model of down syndrome: Effects of environmental enrichment. Cereb Cortex. 2003;13:758-764. 113. Eadie BD, Redila VA, Christie BR. Voluntary exercise alters the cytoarchitecture of the adult dentate gyrus by increasing cellular proliferation, dendritic complexity, and spine density. J Comp Neural. 2005;486:39-47. 114. Murmu MS, Salomon S, Biala Y, Weinstock M, Braun K, Bock J. Changes of spine density and dendritic complexity in the prefrontal cortex in offspring of mothers exposed to stress during pregnancy. Eur J Neurosci. 2006;24:1477-1487. 115. Scott EK, Luo L. How do dendrites take their shape? Nat Neurosci. 2001;4:359-365. 116. Wolfer DP, Stagljar-Bozicevic M, Errington ML, Lipp HP. Spatial memory and learning in transgenic mice: Fact or artifact? News Physiol Sci. 1998;13:118-123. 90 117. Squire LR, Stark CE, Clark RE. The medial temporal lobe. Annu Rev Neurosci. 2004;27:279-306. 118. Funfschilling U, Saher G, Xiao L, Mobius W, Nave KA. Survival of adult neurons lacking cholesterol synthesis in vivo. BMC Neurosci. 2007;8:1. 119. Tsui-Pierchala BA, Encinas M, Milbrandt J, Johnson EM,Jr. Lipid rafts in neuronal signaling and function. Trends Neurosci. 2002;25:412-417. 120. Hooper NM. Detergent-insoluble glycosphingolipid/cholesterol-rich membrane domains, lipid rafts and caveolae (review). Mol Membr Biol. 1999;16:145-156. 121. Gould E, Beylin A, Tanapat P, Reeves A, Shors TJ. Learning enhances adult neurogenesis in the hippocampal formation. Nat Neurosci. 1999;2:260-265. 122. Donovan MH, Yazdani U, Norris RD, Games D, German DC, Eisch AJ. Decreased adult hippocampal neurogenesis in the PDAPP mouse model of alzheimer's disease. J Comp Neural. 2006;495:70-83. 123. Gower AJ, Lamberty Y. The aged mouse as a model of cognitive decline with special emphasis on studies in NMRI mice. Behav Brain Res. 1993;57:163-173. 124. Head E, Lott IT, Patterson D, Doran E, Haier RJ. Possible compensatory events in adult down syndrome brain prior to the development of alzheimer disease neuropathology: Targets for nonpharmacological intervention. J Alzheimers Dis. 2007;11:61-76. 91 a 210 180 * 150 a. 0 a.4" 120 90 eaC ..c ▪ 60 *C. 30a. o wt (n = 9 slices, 5 mice) ■ abcgl+ (n=10 slices, 5 mice) -30 Appendix 1 — Supplementary Data -20 ^ 0 ^ 20 ^ 40 ^ 60 Time (min) Before Tetanus ^ Post Tetanus Figure Al. Intact long-term potentiation and paired pulse facilitation in CA1 of aged ABCG1+ mice. In vitro electrophysiology was performed on 400um thick hippocampal slices derived from aged mice overexpressing ABCG1 (abcgl+) and wild-type littermates. a) Normal long term potentiation (LTP). High frequency stimulation was applied to the Shaeffer collaterals to induce LTP in the CA1 region. No significant difference was observed between wild-type and transgenic mice. b) Normal paired-pulse facilitation (PPF). Two pulses were applied to the Shaeffer collaterals, inducing paired-pulse facilitation in the CA1 region. No significant difference was seen in wild-type and transgenic mice. This work was done by Timal Kannangara. 92 THE UNIVERSITY OF BRITISH COLUMBIA Pamela Parkinson has successfully completed the online training requirements of the Canadian Council on Animal Care (CCAC) / National Institutional Animal User Training (NIAUT) Program Chair, Animal Care Committee^ Veterinarian Certificate #: 0963 Date Issued: August 22, 2005


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items