Detecting Early Cognitive and Synaptic Changes in an Amyloid Beta Mouse Model of Alzheimer's Disease

MISSING IMAGE

Material Information

Title:
Detecting Early Cognitive and Synaptic Changes in an Amyloid Beta Mouse Model of Alzheimer's Disease
Physical Description:
1 online resource (157 p.)
Language:
english
Creator:
Montgomery, Karienn Souza
Publisher:
University of Florida
Place of Publication:
Gainesville, Fla.
Publication Date:

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Medical Sciences, Neuroscience (IDP)
Committee Chair:
Bizon, Jennifer L.
Committee Members:
Notterpek, Lucia
Setlow, Barry
Foster, Tom
Candelario-Jalil, Eduardo
Leeuwenburgh, Christiaan

Subjects

Subjects / Keywords:
alzheimer's -- amyloid -- dementia -- disease -- learning -- memory -- model -- mouse -- transfer
Neuroscience (IDP) -- Dissertations, Academic -- UF
Genre:
Medical Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract:
Early identification of Alzheimer’s disease (AD) iscritical to designing treatments to prevent the progression of the disease and toultimately yield improved quality of life and significant cost savings tofamilies and society. While current cognitive assessments are highly accuratein predicting AD, these assays are only sensitive to late stages of the diseaseand most diagnostic assessments of cognition in humans are verbal and are,thus, not translatable across species. Such translation is important because animalmodels, particularly mouse models that are genetically modified to exhibitdifferent features of the disease, are used in the vast majority of studiesfocusing on disease etiology and mechanisms as well as in preclinicalassessments of potential therapies for AD. Whereas these tools have beeninvaluable for helping to uncover disease processes, recapitulating thecognitive/behavioral sequelae of AD remains a challenge. An ideal cognitiveassessment would be 1) sensitive to early stages of the disease; 2) adaptablefor cross-species comparisons and 3) suitable for repeated assessment andwithin-subjects experimental designs. Tasks which are based in associativelearning paradigms can be easily performed by both humans and rodents. One typeof learning, “transfer learning,” which refers to the ability to applypreviously learned information to novel contexts, appears to be impaired earlyin dementia and even predicts individuals who progress into AD1. These data suggest thattransfer learning may be useful for earlydetection of impairment and for cross-species studies. The overarching theme of this dissertation was to develop a mousetransfer learning task that is sensitive to hippocampal function (Chapter 2)and, using well-established mouse models of AD, to determine if transfer learning is sensitive to  neural alterations  that represent the earliest pathologicalfeatures of this disease (i.e., beta amyloid pathology and synapticdysfunction; Chapters 3 and 4). Together, the data support that transfer learningcan be used to assess hippocampal function in rodents, and in particular, transferlearning is sensitive to decline in mouse models that recapitulate some of theearliest pathological features of AD (i.e. amyloid ß deposition and synaptic dysfunction).
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility:
by Karienn Souza Montgomery.
Thesis:
Thesis (Ph.D.)--University of Florida, 2012.
Local:
Adviser: Bizon, Jennifer L.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2013-12-31

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Classification:
lcc - LD1780 2012
System ID:
UFE0044514:00001


This item is only available as the following downloads:


Full Text

PAGE 1

1 DETECTING EARLY COGNIT I VE AND SYNAPTIC CHANGES IN AN AMYLOID BETA By KARIENN ANTONIUK SOUZA MONTGOMERY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PAR TIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Karienn Antoniuk Souza Montgomery

PAGE 3

3 To Ga (meu pai)

PAGE 4

4 ACKNOWLEDGMENTS First I would like to acknowledge my mentor and friend Dr. J ennifer Bizon, who through the years has always encouraged me, has believed in my potential and invested in me : n o words can express the love and respect I have for you Dr. Barry Setlow, helped me to be become a better professional and scientist I would also like to thank my committee members, who have challenged me to aim for excellence in my work, and who always helped me realize how much there is still to learn about the brain. Dr. Lucia Notterpek: thank you for guiding me to think about various aspect s of the project; and also, for personally believing in my potential. Dr. T om Foster: thank you for allowing me to learn from your lab, and for always being available to discuss and teach me about all things electrophysiology, hippocampus and aging. Dr. Ed uardo Candelario: thank you for taking the time to teach me how to perform Western blots it was a great opportunity to learn from you Dr. Christiaan Leeuwenburgh: thank you for accepting to be a member of my committee at the last minute: your input has b een invaluable, and I wish y ou had joined earlier. Dr Brandi Ormerod (former me mber ): thank you for all your help and input on the project, for always being responsive and helpful when I asked questions, and for your encouragement and wisdo m when advising me on my scientific career I would also like to acknowledge all the individuals who played an integral part on execution of the experiments. By chronological order, the students and staff who helped me w ith the behavioral experiments: Rebecca Simmons, D avid Mathai Katherine Burton, Kayla Florio, Kelly Demars Vicky Kelly, Kim Brown and Kenneth Vera: thank you for your s trong work ethic and responsible behavior Several graduate students have helped in many ways th rough the extent of the project: Candi L asarge, Cristina

PAGE 5

5 Banuelos, George Edwards III, Ian Mendez, Marci Mitchel, Michael Guidi, Nick Simon, Nicolle Little Rya n Gilbert and Sofia Beas Further, I would like to thank collaborators, who have introduced new ideas and perspective s to my project a nd some who have generously offered support through sharing of lab materials : Dr s Yona Levites, Todd Golde, David Borchelt, Paramita Chakrabarty, Christopher Janus, Tom Foster, Ashok Kumar, Eduardo Candelario, Lucia Notterpek, Sooy eo n Lee, Ron Mandell Mi chelle Nicolle (Wake Forest University) Mark Gluck (Rutgers University) and Catherine Myers (Rutgers Univeristy) Last but not least I would like to acknowledge the people who have cheered me on and have encouraged me to continue pursuing a Ph.D. First m y husband, Stephen Montgomery who has walked beside me through these years ; and has helped by watching the kids during the nights and weekends when experiments had to be done, by supporting me when experiments did not work, and by always believing that I c ould accomplish anything I set my mind to. My parents ( Algacir and Silete) for doing everything in their power to offer me opportunities they did not have for believing in my potent ial and alw ays encouraging me to do my best : I love you very much and I am proud to be your daughter My grandma Nelly who is an inspiration to me, showing me that o bstacles can be overcome when you maintain your principles and Aunt Cleide and Uncle John (in memory) who have alwa ys believed in my potential inv ested in m e and always treated me like their own d aughter : I could never thank you enough for the opportunity you have given me Uncle Dan Leedy who ever since coming into our family has been encouraging and uplifting. Aunt Lilia ne and Uncle Larry, who ha ve treated me like a daughter as well : your home has always been a refuge, and

PAGE 6

6 your words have always encouraged me. M y in laws, Billy and Sandy who have also been an encouragement, and who ha ve provided immense physical, spiritual and emotional support t o Stephen and I : thank you for adopting me into your family. My brother Derryck, cousin Shawn and Aunt Elisete: you have encouraged me in so many ways through the years. And I would like to thank my two little angels, Connor and Leilani, who are my joy and inspiration and who kept me grounded when I wanted to give up: it was hard finishing what I started but never hard being your mommy and I would not do it any other way.

PAGE 7

7 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ .......... 11 LIST OF FIGURES ................................ ................................ ................................ ........ 12 LIST OF ABBREVIATIONS ................................ ................................ ........................... 14 ABSTRACT ................................ ................................ ................................ ................... 16 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 18 isease: Cognitive Dysfunction and Pathology ..... 18 The Hippocampus, Cognition and AD ................................ ................................ ..... 22 The Hippocampus and Transfer Learning ................................ ........................ 23 Development of a Human Transfer Learning Task ................................ ........... 25 Developing a Mouse Model of Transfer Learning: the Need for Cross Species Assessme nts ................................ ................................ ................... 28 Outline and Experimental Goals ................................ ................................ ............. 29 2 DEVELOPMENT OF A MOUSE TRANSFER TASK ................................ ............... 36 Methods ................................ ................................ ................................ .................. 37 Experiment 1 ................................ ................................ ................................ ........... 38 Can Transfer Learning be Assessed in Mice? ................................ .................. 38 Subjects ................................ ................................ ................................ ..... 3 8 Apparatus and task parameters ................................ ................................ 38 Shaping ................................ ................................ ................................ ...... 39 Testing ................................ ................................ ................................ ....... 39 Are Mice Sensitive to the Changes in Stimulus Features Associated with the Transfer Phase? ................................ ................................ ...................... 41 Sta tistical Analyses ................................ ................................ .......................... 41 Experiment 2 ................................ ................................ ................................ ........... 42 Is Transfer Learning Hippocampal Dependent? ................................ ............... 42 Subjects ................................ ................................ ................................ ..... 42 Surgical procedures (bilateral hippocampal lesions) ................................ .. 42 Transfer learning task ................................ ................................ ................ 43 Odor Detection Threshold Testing ................................ ................................ .... 43 Water Maze Assessment ................................ ................................ .................. 44 Histology ................................ ................................ ................................ ........... 46 Statistical Analyses ................................ ................................ .......................... 46 Results ................................ ................................ ................................ .................... 47

PAGE 8

8 Experiment 1 ................................ ................................ ................................ ........... 47 Development of the Transfer Learning Task ................................ .................... 47 Are Mice Sensitive to the Change in Context in the Transfer Phase? .............. 48 Experiment 2 ................................ ................................ ................................ ........... 48 Does Transfer Learning Depend on the Hippocampus? ................................ .. 48 Hippocampal lesion verific ation ................................ ................................ .. 48 Transfer learning performance ................................ ................................ ... 49 Odor detection threshold ................................ ................................ ............ 49 Morris Water Maze Performance ................................ ................................ ...... 50 Swim speed ................................ ................................ ............................... 50 Cued (visible) platform training ................................ ................................ .. 50 Spatial (hidden) platform training ................................ ............................... 50 Discussion ................................ ................................ ................................ .............. 51 3 DISEASE ......... 63 Choosing a Mouse Model of AD ................................ ................................ ............. 63 Methods ................................ ................................ ................................ .................. 66 Subjects ................................ ................................ ................................ ............ 66 Experiment 1: Is Transfer Learning Sensitive to Robust Plaque Pathology in Aged (12 months old) APPswePS1 Mice? ................................ .................... 66 Experiment 2: Is Transfer Learning Sensitive to A Elevations Prior to the Onset of Plaque Deposition in APPswePS1 Mice? ................................ ....... 66 Transfer learning task: apparatus and task parameters ............................. 66 Shaping ................................ ................................ ................................ ...... 67 Testing ................................ ................................ ................................ ....... 67 Odor detection threshold testing ................................ ................................ 68 Ex periment 3: Is Water Maze Performance Impaired in 12 Month Old APPswePS1 Mice? ................................ ................................ ....................... 68 Apparatus ................................ ................................ ................................ ... 68 Water maze analysis ................................ ................................ .................. 69 Statistical analyses ................................ ................................ .................... 69 Results ................................ ................................ ................................ .................... 71 Experiment 1 ................................ ................................ ................................ .... 71 Experiment 2 ................................ ................................ ................................ .... 71 Experiment 3 ................................ ................................ ................................ .... 73 Discussion ................................ ................................ ................................ .............. 73 4 EARLY DETECTION AND POSSIBLE MECHANISMS ................................ .......... 86 Methods ................................ ................................ ................................ .................. 89 Subjects ................................ ................................ ................................ ............ 89 Experiment 1: To What Extent are Transfer Learning Deficits Associated with A APPswePS1 and Tg SwDI Mice? ................................ 90 Transfer learning task ................................ ................................ ................ 90 ................................ ......... 91 Enzyme linked immunosorbent assay (ELISA) ................................ .......... 92

PAGE 9

9 Thioflavin(S) staining ................................ ................................ ................. 93 Statistical Analyses ................................ ................................ .......................... 93 Transfer learning assessment ................................ ................................ .... 93 A pathology ................................ ................................ .............................. 94 Experiment 2: Do Alterations in Markers of Synaptic Integrity Accompany Transfer Learning Deficits APPswePS1 and Tg SwDI Mice? ........................ 94 Western blotting ................................ ................................ ......................... 95 Statistical analyses ................................ ................................ .................... 96 Experiment 3: Is There Other Evidence for Synaptic Dysfunction in 6 Month Old APPswePS1 Mice? ................................ ................................ ................. 96 Statistical Analyses ................................ ................................ .......................... 98 Results ................................ ................................ ................................ .................... 98 Experiment 1: To What Extent are Transfer Learning Deficits Associated SwDI Mice? ................................ 98 Transfer learning performance in APPswePS1 mice ................................ 98 Transfer learning performance in Tg SwDI mice ................................ ........ 99 Comparison of transfer learning across APPswePS1 and Tg SwDI models ................................ ................................ ................................ .... 99 ................................ ................................ ................ 100 A 40 levels (ELISA) ................................ ................................ ................ 101 Relationship between A pathology and transfer learning performance in APPsweP S1 mice. ................................ ................................ ............ 102 Experiment 2: Do Alterations in Markers of Synaptic Integrity Accompany Transfer Learning Deficits APPswePS1 and Tg SwDI Mice? ...................... 102 Effects of chronological age on synaptophysin and PSD95 expression ... 102 Synaptic protein expression in APPswePS1 mice and Tg SwDI .............. 103 APPswePS1 model ................................ ................................ .................. 103 Tg SwDI model ................................ ................................ ........................ 104 Experiment 3: Is There Other Evidence for Synaptic Dysfunction in 6 M onth Old APPswePS1 Mice? ................................ ................................ ............... 104 Discussion ................................ ................................ ................................ ............ 105 Transfer Learning Deficits in APPswePS1 ................................ ..................... 106 Comparison of Transfer Learning Deficits in APPswePSI and Tg SwDI Mice 107 Synaptic Integrity and Transfer Learning Performance ................................ .. 109 Synaptic Function in APPswePS1 Mice ................................ ......................... 113 5 C ONCLUSION ................................ ................................ ................................ ...... 124 Implications of Chapter 2 and Future Studie s ................................ ....................... 124 Implications of Chapter 3 and Future Studies ................................ ....................... 125 Implications of Chapter 4 and Future Studies ................................ ....................... 126 APPswePS1: Implications and Future Studies ................................ ............... 127 Tg SwDI: Implications and Future Studies ................................ ..................... 129 Final Words ................................ ................................ ................................ ........... 132 LIST OF REFERENCES ................................ ................................ ............................. 134

PAGE 10

10 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 157

PAGE 11

11 LIST OF TABLES Table page 2 1 Odor and digging media pairs as paired in the transfer learning task ................. 61 2 2 Representative example of the order and sequence of complex stimuli presented in the mouse transfer task ................................ ................................ 62 4 1 Table summarizes APPswePS1 and Tg SwDI mouse models ..................... 123

PAGE 12

12 LIST OF FIGURES Figure page 1 1 Hypothetical model of dynamic biomarkers during AD progression .................... 32 1 2 Schematic of the medial temporal lobe system and associated structures in human. ................................ ................................ ................................ ............... 33 1 3 Hippocampal and non hippocampal model of stimuli encoding. ......................... 34 1 4 Human Transfer Lear ning paradigm and small scale study with non demented human subjects with and without hippocampal atrophy (HA) ............ 35 2 1 Transfer learning task in mice ................................ ................................ ............. 55 2 2 Validation of the transfer task for mice ................................ ............................... 56 2 3 H ippocampal lesions in C57BL/6J. ................................ ................................ ..... 57 2 4 Bar graphs show performance of sham lesioned (green bar) and ibotenic acid lesioned (red bar) mice on the transfer learning task ................................ .. 58 2 5 Odor detection threshold testing in sham (green bar) and lesioned (red bar) mice ................................ ................................ ................................ .................... 59 2 6 Water maze performance in sham (green) and Ib otenic acid lesioned (red) mice ................................ ................................ ................................ .................... 60 3 1 Amyloid pre cursor pr otein (APP) processing ................................ ...................... 77 3 2 Amylo id aggregation at the synapses ................................ ................................ 78 3 3 Photomicrographs show Thioflavin S stained coronal sect ions of the hippocampus in 3, 6 a nd 12 month old APPSwePS1 mice ................................ 79 3 4 Performance of aged (12 months) APPswePS1 (blue bars) and age matched NTg mice (red bars ) on the transfer learning task ................................ .............. 80 3 5 Performance of young (3 months) APPswePS1 (blue bars) and age matched NTg mice (red bars) on the transfer learning task ................................ .............. 81 3 6 Performance of the same APPswePS1 (blue bars) and NTg mice (red bars) shown in Figure 3 5, re tested for transfer learning at 12 mo ............................ 82 3 7 Previous exposure to the task did not alter the ability to detect a significant deficit in transfer learning ................................ ................................ ................... 83

PAGE 13

13 3 8 Odor detection threshold testing in young and aged APPswePS1 (blue bars) and NTg (red bars) mice ................................ ................................ ..................... 84 3 9 Water maze performance in aged APPswePS1 (blue bars) and NTg (red bars) mice ................................ ................................ ................................ ........... 85 4 1 Recording of field excitatory postsynaptic potentials (fEPSP) a t hippocampal CA3 CA1 synapses in brain slices. ................................ ................................ ... 115 4 2 Transfer learning performance in 3 and 6 month old APPswePS1 and Tg SwDI mice ................................ ................................ ................................ ........ 11 6 4 3 SwD I than in APPswePS1 hippocampus ............... 117 4 4 Linear regression plots showing relationship between A 42 levels and transfer learning performance in APPs wePS1 ................................ .................. 118 4 5 Synaptic protein expression in APPswePS1 mice ................................ ........... 119 4 6 GFAP expression in Tg SwDI mice ................................ ................................ .. 120 4 7 Baseline transmissio n and paired pulse facilitation ................................ .......... 121 4 8 Long term potentiation in APPswePS1 (blue circles) and age matched NTgs (red circles) ................................ ................................ ................................ ....... 122

PAGE 14

14 LIST OF ABBREVIATION S ACSF A rtificial cerebrospinal fluid AD APP Amyloid precursor protein Amyloid Beta Protein BSA Bovine serum albumin E East EDTA ethylenediamine tetra acetic acid ELISA Enzy me Linked Immunosorbent Assa y EPSP Excitatory post synaptic potential FA Formic acid H Hour HZ Hertz LTP Long term potentiation N North PBS Phosphate buffered saline PFA paraformaldehyde PFA Paraformaldehyde PPF Paired pulse facilitations PS1 Presinilin mu tation 1 PS2 Presenilin mutation 2 RIPA Radioimmunoprecipitation assay S South TBS Tris buffered saline TBS Tris buffered saline

PAGE 15

15 TBST Tris buffered saline and Tween W West

PAGE 16

16 Abstract of Dissertation Presented to the Graduate School of the University of Flor ida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy DETECTING EARLY COGNITIVE AND SYNAPTIC CHANGES IN AN AMYLOID BETA By Karienn Antoniuk Souza Montgomery December 2012 Chair: Jennifer Lynn Bizon Major: Medical Sciences Neuroscience is critical to designing treatments to prevent the progression of the disease and to ultimately yield improved quali ty of life and significa nt cost savings to families and society. While current cognitive assessments are highly accurate in predicting AD, these assays are only sensitive to late stages of the disease and most diagnostic assessments of cognition in humans are verbal and are, thus not translatable across species. Such translation is important because a nimal models, particularly mouse models that are genetically modified to exhibit different features of the disease, are used in the vast majority of studies focusing on disease etiol ogy and mechanisms as well as in preclinical assessments of potential t herapies for AD. Whereas these t ools have been invaluable for helping to uncover disease processes, recapitulating the cognitive/behavioral sequ e l ae of AD remains a challenge. An ideal cognitive assessment would be 1) sensitive to early stages of the disease; 2) adaptable for cross species comparisons and 3) suitable for repeated assessment and within subjects experimental designs. Tasks which are based in associative learning paradigms can be easily performed by both humans and rodents. ransfer learning

PAGE 17

17 learned information to novel contexts, appears to be impaired early in dementia and even predicts individuals who progress into AD 1 These data suggest that transfer learning may be useful for early detection of impairment and for cross species studies. The overarching theme of this dissertation was to develop a mouse transfer learning task that is sensitive to hippocampal function (Chapter 2) and using well established mouse models of AD to determine if transfer learning is sensitive to neural alterations that represent the earliest pathological features of this disease (i.e., beta amyl oid pathology and synaptic dysfunction; Chapters 3 and 4). Together the data support that transfer learning can be used to assess hippocampal function in rodents, and in particular, tran sfer learning is sensitive to decline in mouse models that recapitula te some of the earliest pathological features of AD (i.e. and synaptic dysfunction).

PAGE 18

18 CHAPTER 1 INTRODUCTION nervous system, characteri zed by loss of memory and higher cognitive functions. An estimated 24 million people have AD (dementia) today. This number is believed to double every 20 years, and could rise to 42 million by 2020, assuming no effective prevention strategies or curative t reatments are discovered 2 The increase in t he prevalence of AD presents a harrowing social and economic challenge that will impact not only the affected individuals but also the burden on families and other caregivers. Tests developed by cognitive aging researchers have provided evidence that norma l aging is accompanied by declines in speed of information processing, memory, executive function and reasoning 3 The nature of cognitive deficits in dementia is not fundamentally different, with patients presenting abnormalities of memory, problem solving, language, calculation, visuo spatial pe rception, and judgment 4 6 As AD progresses, unlike normal aging, some patients may develop psychotic symptoms such as hallucinations and delusions. In the final stages of the disease, general bodily functions are impaire d, and individuals are mute, incontinent, and bedridden 2,7 8 Currently, there are no tests that can definitively establish a living diagnosis of A D type d ementia. Thus, diagnosis is based on medical history, anecdotal information from patients and relatives 9 laboratory assessments of biomarkers (discussed in more detail later in the Chapter) and physical and neurological examinations, along with neuropsychological testing. Cognitive testing is important to this diagnostic battery in that cognitive decline, particularly associated with hippocampal dysfunction, significantly

PAGE 19

19 improves accuracy of an AD diagnosis 1,10 15 As such much effort has been directed toward the development of cognitiv e screens, and a plethora of cognitive testing paradigms are presently available to clinicians. The mini mental state exam (MMSE) is one of the most widely used assessments 16 for AD This test assesses proficiency in a number of cognitive domai ns (e.g., spatial orientation, memory, language, and visual construction) 16 and is designed to provide an overall measure of cognitive abilities. Assessmen ts tailored to more specific aspects of cognitive decline that show some promise for dete ction of AD 17 include the Tower of London (TOL), which is used for evaluating planning skills (executive function) 18 and the Digit Span (DS) te st, which measures auditory attention, immediate span of learning, and working memory 19 However, memory is one of the earliest and most profoundly affected aspects of cognition that is affected in AD. Most memory assessments that have been used for assessment in AD have focused on declarative memory which can be simply defined as conscious recollection of previously learned information about people, places and events 10 These include the Logical Memory (LM) and t he California Verbal Learning (CVLT) tests 20 22 Importantly, however, while declarative memory assessments are quite accurate in diagnosing AD, they are not usefu l as pre clinical assessments, as they are not sensitive to incremental changes in disease associated pathology 23 It is well accepted that the cognitive symptoms of AD are the result of typical pathology observed in the brains of diseased individuals 24 28 The main pathological hallmarks of AD are extracellular cytoplasmatic amyloid beta (A ) plaques, intracellular neurofi brillary tangles (NFTs; phosphorylated tau) and neuronal and synaptic loss 4,29 34 Although considerable research efforts have been focused on uncovering the molecular

PAGE 20

20 pathways of AD, exactly how these pathologies interact to produce cognitive symptoms remains poorly understood. Nevertheless, levels of A have each been linked to cognitive impairment in AD, and more recently, due to advances in vivo detection, it has been determined that the development of these pathologies may precede the symptomatic phase of the disease 35 Despite these findings, current diag nostics (for time line, see yellow ure 1 1) are only sensitive to deficits well after significant pathological burden has accumulated and p roduced irreversible detriment to the brain (i.e., significant neuronal loss). As such, a major challenge in the field is to define diagnostic criteria that are sensitive to pre symptomatic levels of pathology and that have strong predictive value for dise ase progression. The ability to identify individuals early would afford an optimal treatment window in which to administer disease modifying therapies that could offer benefit beyond current treatments that largely focus on symptom management (see pink 1). Thus far, much of the research centered on early detection in AD has been focused on identifying molecular biomarkers of the disease. B iomarkers are parameters (physiological, biochemical, anatomic al ) that can be measured in vivo and that reflect specific features of disease related pathophysiological processes 25 Biomarkers occupy an essential place in diagnostic criteria for AD, as they allow for identification of the pathophysiolo gical processes underlying cognitive impairment At present, the earliest detectable pathological change clinical manife stations of AD 36 37 Based on the data supporting this idea, Jack and

PAGE 21

21 colleagues 37 have developed a biomarker model which is illustrated in Figure 1 1. This model identifies before the appearance of clinical symptoms (red The severi ty and change over time in synaptic markers (orange curve evidenced by positron emission tomography ) tend to correlate with clinical symptoms du ring symptomatic stages of AD 32,34,38 43 whereas brain structure abnormalities (blue curve, evidenced by structural MRI studies) are not evident until later when marked neuronal loss is present. Brain structure ret ains a close relationship with cognitive performance (purple curve) through the clinical phases of AD 44 Note in Figure 1 1 that none of the biomarkers are static; biomarker s change over time and follow a non linear time course 37 Furthermore, it is not known how early detrimental A begins prior to onset of clinical symptoms but current theories suggest that the lag between A pathology and synaptic dysfunction may be more than a decade (s ee Chapter 5 for discussion) 31,45 51 A lthough molecular biomarkers offer promising improvements to early diagnosis of AD, in many instances, the presence of a biomarker does not result in manifestation of like dementia 2,13,32,52 53 Further more, in general, the initial reason why a patient seeks professional medical help is because cognitive difficulty has began affecting day to day functions. Thus, testing for biomarkers is warranted only after the clinician has already established that cog nitive function is impaired 17,21,26 Thus, accurate and sensitive cognitive assessments are an essential complementary tool for early detection of AD 2,13,32,52 53 .One strat egy for improving the sensitivity of cognitive assessments including preclinical AD is to identify aspects of cognitive func tion that are sensitive to the early pathologies associated with the disease (i.e., A pathology

PAGE 22

22 and synaptic dysfunction). Such assays could aid in improved, early diagnosis of AD during a window in which disease progression could be halted or even reversed (see ideal diagnosis window pink box in Figure 1 1 ). The Hippocampus Cognition and AD As mentioned above, loss of hippocampal dependent memory is a common feature of AD and molecular and functional changes in the hippocampus are among the earliest to manifest in the disease 54 55 Thus, a comm on feature across many cognitive assessments used diagnostically in AD is their dependence on the hippocampal formation. The hippocampus is a major component of the medial temporal lobe system (Figure 1 2A modified from Bizon and Nicolle, 2006 ), a brain s ystem that is highly conserved across species. One aspect of hippocampal function that is often assessed in AD is declarative/explicit memory 10,13,56 57 58 The hippocampus undergoes significant alterations over the course of the disease and these changes can result both in hippocampal atrophy and in impairments on tasks that assess declarative memory, including delayed recall assessment of episodic lo ng term memory and verbal memory impairment 10,56 Even though tasks like delayed paragraph recall are among the most reliable and accurate tools for a living diagnosis of AD, by the time individuals are diagnosed the vast majority already have life interfering mnemonic dysfunction and pronounced, multifaceted forms of hippocampal pathology 56,59 In addition to declarative memory, however, the hippocampus is also involved with other aspects of cognitive functioning and, as such, there may be alternative forms of hippocam pal supported cognition that are sensitive to pathol ogies associated with early AD. Figure 1 2B shows a diagram of information flow involving the hippocampus and s urrounding circuitry. Note that the hippocampal region is inter connected with all

PAGE 23

23 primary neo cortical association areas which process information about visual, auditory parahippocampal perirhinal and entorhinal cortices (ErC) Connections carrying information be tween hippocampus and associational cortices via entorhinal cortex are reciprocal 58,60 and after being funneled back t o the ErC, the information is thought to be redistributed to associational neocortical areas 58,60 61 This pathway implicates the hippocampus as important for forming connections and relationships between information acqui red by the association cortices allowing associations between two or more stimuli to be formed. The ability to form associations between stimuli experienced together (or to integrate stimuli into an existing context) aids in subsequent recall (i.e., memory). In this view, the hippocampal system aids in the encoding of related information into long term storage so that it can remembered as associated and recalled by more than one sensory cue later 58,62 The Hippocampus and Transfer L earning Building on its role in associative learning, Drs. Mark Gluck and Catherine Myers have developed unique c omputational models of hippocampal function. These models predict that, given its role in encoding associations and relationships between stimuli, the hippocampus is particularly important in the ability to apply learned information in one context to a novel problem or situation (hereon referred to as transfer learning ) 12,62 66 During transfer learning, it is hypothesized that the hippocampal region integrates/ manipulates information, and allows transfer of previously learned information to new contexts through what has been previously defined as generalization 65 67 Evidence that the hippocampus is involved in transfer learning finds empirical support in animal studies showing that rats with hippocampal region damage may

PAGE 24

24 uli (or stimulus features) during learning conferring impairments when aspects of the learned stimuli are altered 68 69 Unitization refers to inflexible learning of com pou nd stimuli as if they were one. For example but ar learning that rewa ot true). I n a study by Eichenbaum and colleagues, rats were trained to associate an o dor w ith a reward (e.g., A+), and tra ined further to distinguish between pairs of odors (e.g., A+B X + Y etc. with + signifying the correct choice for finding a reward). Rats were then able to respond appropriately when these learned associations about stimulus reward relationships were re configured into novel combinations of the f amiliar odors (e.g., A+Y X +B ) 70 However, rats with hippocampal region dysfunction resulting from fornix lesions performed at chance on the problems involving novel combinations of previously learned odors 71 This effect can be interpreted as indicating that hippocampal lesioned animals unitized th e odor stimuli, perceiving the AB compound as a unit rather than as its component odors A and B P resentation of A Y a novel compound rather than a novel c ombination of familiar would require new asso ciations to learn the relation ship of this stimulus compound to reward. Similarly, in other studies in which animals with hippocampal damage show impairments when novel configurations are presented 72 73 hippocam pal lesioned animals can learn the original associations as well as controls under conditions that favor stimulus unitization 74 76 These data would suggest that transfer learning, or generalization requires that information is effectively encoded via the hippocampal region during initial learning.

PAGE 25

25 Indeed, this region is crucial for efficient flexible manipulation of information about stimuli previously processed by the association cortices when that information is experienced in new contexts 62 64,70,77 Figure 1 3A and B illustrate how associative learning of stimuli might be encoded both with and without hippocampal mediation. In A, stimulus input is represented by grey ci rcles. This input is comprised of information such as colors, shapes, and odors. The information is learned (light blue circles) and encoded as separate but related features, and the hippocampus allows these stimuli to be used flexibly, so that information about stimuli paired together can be easily re combined or transferred in novel contexts (multiple dark circles represent potential new combinations of stimuli; each circle is a different context). In the absence of a functional hippocampal, as illustrate d in Figure 1 3B, encoding of associations (learning) is still possible (light blue circles), but the flexibility is lost, and learning is restricted to simpler stimulus response learning that does not support transfer to new contexts/combinations if neces sary 78 (unitization of stimuli, one dark blue circle). Thus, without the hippocampus, co presented stimuli resul t in fused or unitized representations of these stimuli, which are combi ned into one new, non flexible, representation 79 Thus, when a new context or association between sti muli is needed, these associations must be re learned I t is important to note that under normal conditions, learning is believed to occur through both pathways (combined) acting simultaneously Development of a Human T ran sfer Learning T ask The use of tra nsfer learning as a detection method for hippocampal dysfunction has been developing in the last ten years 1,10,12,64,80 81 Based on animal studies described above 70,72 74, 76,78 79,82 Drs. Gluck and Myers created a novel computer based associative learning task to assess transfer learning in humans which has been

PAGE 26

26 empirically tested in aged individuals with mild hippocampal atrophy who are at risk for cognitive decline 1,81 As illustrated in Figure 1 4A and 1 4B this task involve s a series of two item visual discrim ination problems in which subjects learn, via trial and error, to choose the correct object from each pair presented (Figure 1 4A). Objects have two stimulus features (shape and color) but within each pair, objects differ with respect to either color or s hape but not both. In other words, only one of the features ( color in Fig ure 1 4A top ) is relevant to the correct choice within a particular pair, whereas the other feature ( shape in Fig ure 1 4A) is identical between the two items. Pairs of objects are pr esented one at a time, upon which the subject is required to select a right or left key (in a keyboard), representing each object on the screen respectively. Object pairs are randomly chosen by the computer program so that color and shape are relevant an e qual number of times. Once all discrimination pairs (presented pseudo randomly) are learned to criterion ( learning phase ; Figure 1 4A ), an unsignaled transfer phase occurs in which the irrelevant feature of each object pair is changed but the relevant feat ure remain s the same and is still predictive of the correct choice ( transfer phase ; Figure 1 4B ). A study with medically healthy, non demented individuals with mild atrophy in the hippocampus (assessed by structural MRI), has provided initial evidence tha t t he hippocampal region is necessary for transfer learning, and further, that the transfe r learning task may be useful for detecting mild hippocampal dysfunction. In this study, hippocampal atrophied (HA) subjects performed almost identically to non atro phied c ontrols (no HA) in the learning phase of the human transfer task 1 as evidenced by a comparable number of errors in the initial, associative learning phase of the task (Fig ure

PAGE 27

27 1 4C). However, HA individuals were impaired in the transfer phase, as evidenced by significantly more errors relative to no HA controls. This deficit was evident despite the fact that the stimulus feature predictive of the rewa rd (the relevant informa tion) was not changed 1 (Figure 1 4D) Notably, this deficit was observed in the absence of other cognitive deficits in the HA subjects (Figure 1 4E; delayed paragraph recall, no d ifference between groups). The fact that transfer learning deficits correlate with hippocampal atrophy in aged individuals and are detectable prior to deficits on other standard assessments (like delayed paragraph recall) suggests that transfer learning ma y be particularly sensitive to early dysfunction of the hippocampus 1,81 The results from the stu dy with non demented hippocampal atrophied patients served as a foundation for extending the transfer learning task as a possible colleagues found that including the tran sfer learning assessment into the standard cognitive assessment administered to non symptomatic aged individuals significantly enhanced the predict ive value of the cognitive battery to distinguish those elderly individuals who progress ed into clinical AD d uring a 2 year window 11 Indeed, the predictiv e accuracy of AD diagnosis was 91% with the transfer lear ning assessment versus 78% pre diction using standard testing alone 11 These findings, along with the results of the study with the HA subjects, are consistent with the idea that the hippocampus may have a unique role in transfer learning, which supports subsequent generalization or flexible use of learned information when t hat information is encountered in a new context 1,10 1 2,62 65,70,72 73,78,83 84 Importantly, these studies further suggest that transfer learning may be particularly sensitive to modest hippocampal

PAGE 28

28 dysfunction associated with pathological burden that manifests prior to the symptomatic, clinical AD stage (s e e Figure 1 1). Developing a Mouse Model of Transfer Learning: the Need for Cross Species A ssessments Although far from perfect, genetic m ouse models have contributed significantly to the understanding of molecular pathogenesis of AD and are used in the vas t majority of molecular neurobiological and preclinical studies 85 T he development of the first genetically based mouse models of AD in 1991 86 89 introduced new opportunities for preclinical testing, and studies using such models have provided insight into biological mechanisms of AD. More specifically, the advantages of using mouse models include, but are not limited to, the ability to: 1) dissect speci fic pathways of AD pathogenesis, 2) examine the significance of individual pathologies associated with AD on brain function, 3) examine the i nter re latio nship b etween differing pathways pathologies and cognition, and 4) test specific agents and disease modifying therapies. S ince the origination of genetic models, over 300 interventions have been tested and reported in AD mouse models 85,90 91 To date, however, there is still no cure for AD as promising therapies in preclinical models have failed to translate well to the clinic. One possible reason for this failure is the inability to robustly detect and therefore a ccurately evaluate the effects of disease modifying therapies on cognition at preclinical stages. Current standard tasks used in the assessment of hippocampal function in mice consist mainly of spatial memory tasks, such as the Morris water maze. In this hippocampal dependent test of spatial reference memory 75,84,92 94 rodents must use extra maze cues to find an escape platform submerged in a tank of water. Although this task has been used extensively and has bee n successful for detect ing age related

PAGE 29

29 hippocampal impairments in aged rats 57,93 102 in mouse models, spatial learning behavior in the water maze is much less consistent 103 106 In most cases, spatial deficits in transgenic mouse models of AD tend to be modest observed late in the lifespan 107 110 and fail to consistently correlate with reliable biomarkers of AD 111 113 Finally, spatial assessments are mostly used in between subject ex perimental designs to assess efficacy of putative treatments. A lthough it is possible to introduce novelty to spatial tasks by changing platform locations and cues, rodents still tend to improve their performance with training 114 117 and the observed savings of learning in the water maze can make it difficult to detect potential benefits of the therapy being evaluated. Thus, improved cognitive assessments for mice could offer substantial benefit to the field of AD and ultimately to the identification and accurate evaluation of potential therapies. Ideally, such an assessment would be 1) sensitive to early aspects of disease pathology, 2) adaptable for cross species comparisons, and 3) suitable for repeated assessment and within subjects experimental designs that are optimal for intervention studies. Outline and Experimental Goals A key value of the transfe r learning task is that it is nonverbal and thus, unlike many tests of hippocampal dependent declarative memory (e.g., delayed paragraph recall), should be adaptable across species. Rodents, like humans, can readily discriminate objects that contain more than one stimulus feature 98,118 119 As such, it seems possible that a task analogous to the human transfer learning task could be developed for mice. Such a task would enable individual aspects of AD pathology and potential disease modifying therapies to be effectively evalua ted in relation to cognitive function, offering greater comparative power between species Based on this idea, t he

PAGE 30

30 goals o f this dissertation are: 1) to develop a mouse version of the human task and to determine if as in the human task, the hippocampus is critical for transfer learning in the mouse version of the task, 2) to test the utility of the mouse transfer task as an ; and 3) to investigate the role of early AD mediating transfer learning deficits in two mouse models of AD. Data from Chapter 2 show that mice were able to learn a series of concurrent discriminations that contained two stimulus features (odor and digging media) and could readily transfer this learned information to new problems in which the irrelevant feature in each discrimination pair was altered. Additional experiments demonstrated that the hippocampus is essential for transfer learning as lesions of this structure did not affect learning about initial stimulus reward associations but rel iably impaired the transfer of this learned information when the mice were presented with novel configurations of the stimuli. Data from Chapter 3 showed robust transfer learning deficits in separate cohorts of 12 month APPswePS1 mice (a transgenic mouse m od sectional ly and longitudinally. Transfer learning deficits were observed earlier than impairments in another hippocampal dependent task (spatial water maze), and suggest that amyloid deposition in the hippocampus could be mediating t ransfer learning performance. These data from Chapter 3 were further extended in experiments in Chapter 4, which were designed to determine the degree by which transfer learning deficits are mediated by early AD unction). courses of plaque deposition (APPswePS1 and Tg SwDI), transfer learning deficits were evident at the

PAGE 31

31 age at which plaque deposition begins to emerge in each model (6 months in APPsw ePS1 and 3 months in Tg SwDI). Synaptic changes (amount of synaptic proteins and synaptic transmission) are observed concurrently to the appearance of transfer learning impairments in the APPswePS1 model. In the Tg SwDI, no significant changes in synaptic protein levels were observed. Together, the data from these experiments indicate that hippocampal dependent transfer learning abilities are strongly associated with A multiple secondary mechanisms by which A pathology might mediate loss of transfer learning deficits (e.g., synaptic dysfunction). Furthermore, these findings support that the mouse transfer learning task should be an impor tant tool for both mechanistic and preclinical studies related to AD

PAGE 32

32 Figure 1 1 Hypothetical model of dynamic biomarkers during AD progression. A curve), as identified by cerebrospinal fluid A assay or PET amyloid imaging, begins accumulating earlier in the disease, prior to cognitive dysfunction (purple curve). Synaptic dysfunction (orange curve), evidenced by positron emission tomography (FDG PET) or functional magnetic resonance imaging (fMRI), would occur next. Dashed line indicates that s ynaptic dysfunction occur before detectable A deposition, for example in carriers of the APOE gene. Brain structure (blue curve), evidenced by structural magnetic resonance imaging, is detected last, and correlates very well with cognitive abnormalities (purple curve). Biomarkers change from norma l to maximally abnormal ( y axis) as a function of disease stage ( x axis). The temporal trajectory of clinical disease from preclinical to dementia is also illustrated, as characterized by cognitive and behavioral measures. Pink box illustrates the ideal di agnostic window, while the yellow box represents the time window in for current diagnosis (Modified from Jack et al. 2011).

PAGE 33

33 Figure 1 2. Schematic of the medial temporal lobe system and associated structures in human (A). Information pathway circuitry (B). In (A), m id sagittal view of the human brain with medial temporal lobe is shaded in light blue and the hippocampal formation shaded in dark blue (adapted from Bizon and Nicolle, 2006). In (B), the hi ppocampal formation is connected with all neocortic al association areas either directly or thr ough the parahippocampal cortex perirhinal cortex and entorhinal cortex (ErC) The pathway loops back to association cortex, through reciprocal connectivity.

PAGE 34

34 Figure 1 3. Hippocampal and non hippocampal model of stimuli encoding. In A) the hippocampal region network forms new internal representations that compress and differentiate information so that multiple novel combinations of stimuli can be made. In B) the association cortex network receives the same sti mulus input, but can only re construct the representations in a compressed or unitized form.

PAGE 35

35 Figure 1 4 Human Transfer Learning paradigm and small scale study with non demented human subjects with and without hippocampal atrophy (HA). (A) Learning Phase: I n each trial, the subject is presented with a pair of objects on the computer screen and is asked to choose the left or right object which has either the same shape or same color (irrelevant to choice) The chosen object is raised and, if the sub by chance a smiley face is revealed. Otherwise there is no smiley face. (B) Transfer Phase: The irrelevant feature (shape or color) in each object pair is altered, but the relevant feature remains identical (i.e., the relevant fe ature for the pair during the learning phase still predicts reward location). (C) In a study with patients with HA (purple bar) and without HA (noHA, green bar) hippocampal atrophy, p erformance in the concurrent discrimination learning phase of the task di d not differ in the hippocampal atrophy HA and no HA groups both groups learned the associations successfully (D) The HA group averaged significantly more errors on the transfer phase of the task when irrelevant features were changed (E) HA group perfo rmed comparably to the no HA group in the delayed paragraph recall test demonstrating that the transfer learning task may be more sensitive to hippocampal dysfunction than other tasks Figure adapted with permission fr om Myers, Gluck and colleagues e t al. (2002).

PAGE 36

36 CHAPTER 2 DEVELOPMENT OF A MOU SE TRANSFER TASK Despite considerable research efforts focused on uncovering the pathogenesis of effective treatments for this devastating condition remain elusive. One hindrance to developme nt of interventions is the absence of standardized, reliable cognitive assays that both identify individuals at early preclinical stages and that translate well to rodent models, from which the vast majority of neurobiological data (e.g., amyloid plaques, neurofibrillary tangles, synapse and neuronal loss) on the mechanisms of AD are derived As such, there is significant interest in cognitive assessments that are sensitive to mild decrements in vulnerable circuitry and that have poten tial for translational studies. The hippocampus has long been implicated in declarative learning (spatial learning and memory) and is affected early in AD 13,58,62,120 121 Indeed the majority of assessments used for diagnosis of AD are hippocampal dependent (e.g., delayed paragraph recall). However, as discussed in detail in Chapter 1, current hippocampal dependent diagnostic tests are accurate in determining individuals who already suffer from AD, but are not particularly helpful in de tecting impairment at early, pre symptomatic stages 11,56,59 Transfer learning, which is defined as the generalization of previously learned information from one context to another appears to be dependent on the h ippocampus, as individuals with mild hippocampal atrophy were impaired in the human transfer learning task (see Chapter 1 10,62,64 65,82 ) Aside from the benefit of the potential for early detection of hippocampal d ysfunction, the human transfer learning paradigm has strong potential for cross species translation due to the fact that transfer learning can be assessed nonverbally,

PAGE 37

37 and is based on a type of learning that rodents can perform readily 1,11 12 Naturalistic digging tasks used in rodent associative learning tasks 98,118,122 provide a base in which to develop a highly analogous transfer learning task for mice. Rodents are very adept at digging for food rewards, and th ere are a large number of unique digging media that can be paired with odors to serve as stimuli (to parallel colors and shapes in humans, see Figure 1 4). Hence, digging media and odors in the mouse task would parallel the colors and shapes used in the hu man task to achieve multiple contexts. The goal of Chapter 2 is to develop a task that is as analogous as possible to the human transfer learning assessment described in Chapter 1 (Fig ure 1 4). Ideally, the mouse transfer task should: 1) be confined to a s ingle session; 2) demonstrate a dissociation between associative learning and transfer learning (just as in the human task); 3) depend selectively on relevant circuitry (hippocampal region); and ultimately 4) be sensitive to early features of disease patho logy. Methods For Experiments 1 and 2, mice were individually housed in the AAALAC accredited vivarium in the Psychology Building of Texas A&M University (College Station, TX). Mice were maintained on a 12 h light/dark cycle (lights on at 0800) and climate controlled at 25 o C. All testing was conducted during the light cycle and mice in the study were screened daily for health problems. All animal procedures were conducted in accordance with approved institutional animal care procedures and NIH guidelines. A ll mice were given at least two weeks with ad lib access to food and water to habituate to the vivarium upon arrival. After the habituation period, mice were food restricted to 85% of their free feeding weight and handled for one week prior to testing.

PAGE 38

38 Exp eriment 1 Can Transfer Learning be Assessed in M ice? Subjects Female C57BL/6J (n=10) obtained from The Jackson Laboratory, Maine, USA were used to develop the transfer learning task in Experiment 1 Apparatus and task parameters The testing apparatus desi gned for the mouse transfer learning task is illustrated in Figure 2 1A. It consists of an open topped black Plexiglas box (12 inches width; 18 inches length; 8 inches deep) with two small terra cotta flower pots (1.7 inches diameter; 1.3 inches deep) secu rely attached to the floor. Each discrimination problem contained two stimulus features: an odor that was applied directly to the rim of the pots and the digging media that filled the pots and hid a food reward (one chocolate flavored food pellet, 20 mg, A IN 76A from TestDiet), placed on the bottom of the pot. Table 2 1 lists the complex discrimination pairs used in the study. To disguise the odor of the rewar d, crushed chocolate food pellets were sprinkled over the surface of each pot. The position (left o r right) of the rewarded pot varied pseudo randomly across trials. For the first four trials of each new discrimination problem, mice were allowed to dig in both pots until they obtained the reward (i.e. they were allowed to self correct if they dug in the incorrect pot). On these trials, only their first choice was scored (as correct or incorrect). On trials thereafter, mice were removed from the test chamber after only one dig (either correct or incorrect). A dig in a pot was scored if a mouse displaced t he digging media with either its paws or nose.

PAGE 39

39 Shaping Shaping took place in the box described above. Initially mice were presented with two empty pots containing a chocolate pellet in the bottom of each. The re baiting of the pots was contingent on the mo use consuming both pellets. After 12 trials of successively retrieving both pellets, mice were then presented with two pots containing chocolate pellets in the bottom but filled with progressively more mixed digging media (12 trials per 33, 50 and 100% ful l). The day after successfully retrieving both pellets with full pots, mice received a simple discrimination to habituate them to the learning aspect of the task. Testing In developing the mouse transfer learning task and confining it to a single behaviora l session (to closely mimic the human version), satiation and motivation were initial concerns. A series of pilot experiments, in which food restricted C57BL/6J mice were allowed free access to the chocolate food pellets used in this task, revealed that mi ce would consume many more pellets than were ever needed for a full testing session. Moreover, all mice shaped, learned all problems to criterion, and completed all 30 trials in the transfer portion of the task irrespective of genotype or age, supporting t hat there were no motivational differences within sessions or between experimental groups. Just as in the human version of the transfer learning task 1,62 all testing was performed within a single session. Figure 2 1 illustrate s an example of the order and sequence of presentation of the two feature stimuli in the learning (Fig 2 1B) and transfer phases (Figure 2 1C) of the task. Initially, mice were trained on a series of three concurrent discrimination problems. As in the human version of the task, new

PAGE 40

40 discrimination pairs were progressively introduced and intermixed with alr eady learned pairs as criterion performance was reached on each pair. For each problem, mice learned to discriminate between pots with two stimulus features (odor and digging media ). One of the two pots was consistently paired with a food reward with eith er the odor or the digging media as the relevant feature (Fig ure 2 1B, Table 2 2 ). For example, if odor was the relevant feature, the pots differed in odor (with one odor always predictive of the food reward; e.g. rose+ ver sus citrus as shown in Figure 2 1B ) but contained the same digging media (e.g. sequins) that was thus irrelevant to the correct choice. For other discrimina tion problems, the digging medium was the relevant feature that differed between pots and predicted the reward (e.g., yarn+ v ersus beads ; as shown in Fig ure 2 1B ), and the pots were scented with the same odor (e.g. van illa), the irrelevant feature. The positive and negative features in each discrimination pair and the sequence of discrimination problems were randomized across mice, although eac h mouse received discrimination problems in which the relevant feature was alternated between odor and digging media as shown in Table 2 2 As in the human task, initially several problems (2 for the mice) were presented in pseudorandom order and after criterion was reached on these problems (6 consecutive correct choices including 3 from each problem), a third problem was introduced and the three pairs were presented in a pseudorandom fashion until mice reached criterion performance (6 co nsecu tive choices including 2 f rom each problem). This design ensured that each discrimination pair was learned prior to the transfer.

PAGE 41

41 After reaching criterion performance, mice were immediately assessed for transfer learning. The mice were presented with 30 t rials in which only the stimulus feature in each discrimination pair that was irrelevant to the reward was changed, and the 3 new combinations were presented pseudo randomly (including 10 each of the three discriminations; see Fig ure 2 1C). As in the human version of the task, this design affords an opportunity to assess the ability of the mice to transfer the predictive value of a previously learned relevant feature (e.g., a particular odor) to a food reward in a new context (e.g., pots containing a novel digging medium). Are Mice S ensitive to the Changes in Stimulus Features Associated with the Transfer P hase? To determine if mice were sensitive to the shift in context presented in the transfer phase (new stimulus features that are irrelevant to the correc t choice), an experiment was designed to compare performance of mice that continued to perform identical problems in the transfer phase (no shift in context) with mice that received new problems with the irrelevant features shifted (transfer learning). C57 BL6/J mice (n=10) were trained as described above on 3 concurrent discrimination problems. Immediately after reaching criterion performance, half of these mice (n=5) received 30 trials in which the irrelevant feature within each discrimination problem not predictive of food was altered ( transfer phase, Figure 2 1C, trials that were identical to those learned initially ( continual with pairs from learning phase: Statistical A nalyse s Total number of trials and errors to criterion (criterion: 6 correct trials on a row) on the last portion of the learning phase, in which all 3 concurrent pairs were presented,

PAGE 42

42 were compared between experimental groups. Due to lack of difference between the two measures, only errors to criterion were reported and shown. For the transfer phase, in which a fixed number of trials were presented (30 total trials), percent error was compared between groups. Unpaired t tests were used for analyzes of performance in bot h learning and transfer phases. Statistics were performed using StatView software (version 5.01) and p<0.05 was considered significant. Experiment 2 Is T ransfer Learning Hippocampal D ependent? Subjects Female C57BL/6J (n= 14) obtained from The Jackson La boratory, Maine, USA were used. Mice were housed as in Experiment 1 Surgeries were performed at 3 months of age. After being allowed to recover for at least two weeks, mice were subsequently tested in the transfer learning task also as described in Exper iment 1 assessed for olfactory detec tion abilities and then tested on spatial and cued versions of w ater maze. Surgical procedures (bilateral hippocampal lesions) Approximately 2 weeks after arrival, mice were anesthetized in a chamber with i sofluorine an d placed in a stereotaxic apparatus with the skull fixed (Kopf Instrume nts, Tujunga, California, USA). A midline incision was made and holes were drilled in the skull over the lesion sites. A 30 gauge needle attached to a 10 L micro syringe (Hamilton, Ren o, Nevada, USA) mounted on a timer controlled infusion pump (Sage Instruments, Boston, Massachuset ts, USA) was used to inject ibotenic acid (Sigma, St Louis, MO, USA). Ibotenic acid was dissolved in phosphate buffered saline (PBS; final pH 7.4) at a concen tration of 10 m g /mL. A volume of 50 nL was injected at each of four

PAGE 43

43 injection sites per hemisphere, according to the following posterior (P), lateral (L) and vertical (V) coordinates in relation to bregma and the skull surface 123 124 : P1.7, L 1.3 and V 1.9 mM; P 2.1, L 2.8 mM and V 3.0 mM; P 2.4, L 2.9 and V 4.0 mM; P 2.8, L 3.1 and V 3.7 mM. Sham operated mice underwen t the same procedure except that PBS alone was injected. The incision was closed with wound clips and mice received an injection of 2% lidocaine and saline subcutaneously and monitored until recovery from anesthesia. Mice received at least a two week recov ery period prior t o onset of behavioral training. Transfer learning task Mice were assessed on the transfer learning task in a manner identical to that described above in Experiment 1 (but note that all mice received the change in the irrelevant features d uring the transfer phase). Odor D etection Threshold T esting Odor learning impairment has been associated with hippocampal dysfunction 70,78 As olfactory abilities are integral to performance on the transfer learning task, odor detection abilities were assessed in order to determine whether the surgical procedure may ha ve resulted in a decreased ability to detect odors in the ibotenic acid lesioned mice. In a separate session following the completion of transfer learning testing, mice were trained to criterion on one additional olfactory discrimination problem using a fu ll strength odorant (Sandalwood), which was always paired with a food reward, versus a pot with mineral oil (no odor) applied to the rim. Both pots were filled with mixed digging media. Once achieving criterion performance, the mice were tested in a series of discrimination problems during which the full strength odorant (e.g., Sandalwood) was

PAGE 44

44 systematically diluted (1/10, 1/100, 1/1000, 1/10000). Mice were given 16 trials at each dilution and the percent error was used as the measure of performance. Water Maze Assessment The spatial reference memory version of the Morris water maze is a standard task used in rodents to assess hippocampal/medial temporal lobe function. To directly compare the sensitivity of the rodent assessment of transfer learning to spati al learning ability, the ibotenic acid lesioned mice and sham lesioned mice (at approximately 6 months of age) were tested in hidden (hippocampal dependent) and visible cued (non hippocampal dependent) versions of the Morris water maze task after all othe r testing Apparatus: The water maze consisted of a 4 foot diameter circular tank filled with water (24 to 27C) made opaque by the addition of nontoxic white tempera paint. The tank was surrounded by black curtains, to which large (15 X 15 inches) white g eometric cues (made with fabric) were affixed. The tank was divided into four imaginary quadrants, each with a platform position equidistant from the center to the wall. During cue training, the tank was filled to 1 cm below a black visible platform. Each swim was tracked and analyzed using a computer based video tracking system (Water 2020, HVS Image, UK). On each trial, mice were carefully placed into the water facing the wall of the tank at one of four start points (N, S, E, or W). Throughout the experiment, mice that failed to reach the platform within 60 s econds were guided to it by hand. During the spatial reference memory assessment (hidden platform training), a retractable escape platform (12 cm diameter, HVS Image, UK) was located in the sou thwest quadrant of the maze and submerged 1. Initially, mice were trained in the non hippocampal dependent cued version of the task in order to assess visual ac uity and sensorimotor abilities as described

PAGE 45

45 previously 103,109,125 126 During cue training, the platform position and start positions were varied on each trial such that the platform position and extra maze cues were made nsisted of 2 days (6 trials per day). On each trial mice were given 60 seconds to find the visible escape platform. After finding the platform or being guided there by an experimenter, mice remained on the platform for 30 seconds before being removed from the tank. At the conclusion of each trial, mice were returned to a heated holding cage (about 30C) for a 10 minute inter trial interval. Beginning the day immediately following completion of cue training, mice received 6 consecutive days of training to fi nd a hidden, stationary platform, to assess hippocampal dependent spatial reference memory (4 trials per day) The start position (N, S, E or W) was pseudo ra ndomly varied across trials so that mice needed to rely on the position of the platform relative t o the extra maze cues to effectively escape across trials. During each trial, mice were given 60 seconds to search for the hidden platform, followed by a 30 second post trial period in which they were allowed to remain on the platform. After each trial, mi ce were placed in the holding cage for a 10 minute inter trial interval. The fourth swim on days 2 and 6 of hidden platform training were probe trials, during which the platform was retracted to the bottom of the tank for the first 30 seconds of the 60 sec ond trial. Rodents with a good memory for the platform location concentrate their search in that area even when the platform is removed from the tank, and performance on probe trials is particularly sensitive to detecting hippocampal impairment in a variet y of aging models 95,97 98,103,115

PAGE 46

46 Histology Following the completion of behavioral testing (transfer learning, odor threshold testing and water maze), mice were injected i.p. with 0.1 mL of pentobarbital and perf used intracardially with 0.9 % saline followed by 4% paraformaldehyde. The brains were extracted and stored in 4% paraformaldehyde for 24h ours after which they were place in a 30% sucrose PBS solution for 24h ou r s Brains were then frozen and sliced on a f reezing microtome at 30 m sections through the hippocampus (every section was mounted). Sections were mounted on gelatin coated glass slides, hydrated in ethanol, and stained with thionin (0.25%). Next, sections were dehydrated in ascending concentrations of ethanol, delipidated in xylene and coversliped with Permount 127 For lesion verification, sections were visualized in a Zeiss AxioImager.M2 epiflourfescent microscope and traced with the assistance of Stereo Investigator software (Microbrightfield, Williston, VT). Tracings were compared to plates from the Mouse Brain Atlas corresponding to the rostrocaudal extent of hippocampus 123 Percent loss of hippocampal volume and sparing were collected by using the ruler tool from ImageScope software. Sparing was compared to the whole extension of the tracing per section. As all mice presented less than 5% sparing of the hippocampal region bilaterally, none were excluded. Statistical Analyses Analyzes of performance in the learning and transfer phase were perform ed identicall y to analyzes in Experiment 1. For the odor detection thres hold testing, an unpaired t test was used to compare total number of trials and errors during the simple discrimination of full strength odor versus no odor (mineral oil). Errors and tot al trials to criterion were analyzed, and due to

PAGE 47

47 no difference in results, only errors were reported and shown. A two factor repeated measures ANOVA (lesion condition x odor dilutions) was then used to compare lesion group performance (mean % correct at ea ch dilution) on the progressive dilutions of the full strength odorant. Statistics were performed with StatView software (version 5.01), and in all cases p<0.05 was considered significant. The water maze data were analyzed using a computer based video tra cking system, Water 2020, developed by HVS Image (Hampton, UK). Primary performance measures analyzed were path length (the total distance from the start position to the platform in centimeters) on hidden and visible training trials and percent time in the target quadrant on probe trials. Separate two factor repeated measures ANOVA (lesion condition x blocks/ day) were used to evaluate differences across visible platform training trials, hidden platform training trials, and probe trials. Finally, swim speed (centimeters per second) between groups was also analyzed using unpaired t tests on hidden and visible trials respectively. Statistics were performed as described above for each task using St atView software (v ersion 5.01). Results Experiment 1 Developmen t of the Transfer L earning T ask Mice learned the initial associations readily after shaping to dig, and were able to successfully perform on the task within an average of 2 hours. Food restriction to 85% of the original body weight did not appear to impair mouse behavior, as mice appeared no more active or anxiou s than they were during the non food restricted acclimation period. This was evidenced by observation of active grooming during the testing, no hyper active behaviors such as circling, and no increa se in fecal boli during testing 128

PAGE 48

48 Salience and difficulty of odor and digging medium problems appeared comparable as no significant differences in performance on odor and digging media discriminatio ns were observed in any group (data not shown) Additionally, analyses of individual pairings did not result in any differences, indicating comparable salience within individual pairs. Are M ice S ensitive to the Change in Context in the Transfer P hase? Assi gnments to control and test groups were made such that group means on initial discrimination le arning were equivalent (Figure 2 2A, t ( 8) = 0, ns). During the transfer phase, the test group performed significantly worse than the control group (Figure 2 2 B, t (8) = 2.96 p < 0.05). These data demonstrate that, similar to subjects in the human version of the task, mice are sensitive to alterations of the irrelevant stimulus feature. Experiment 2 Does Transfer Learning Depend on the H ippocampus? Hippocampal le sion verification In Figure 2 3A, histology of the hippocampus from a representative ibotenic acid lesioned and sham lesioned mouse are sho wn. As indicated, ibotenic acid lesioned mice showed complete removal of the pyramidal (CA1 CA3) cells with very litt le or no damage to the overlying cortex and the subiculum. A schematic reconstruction of the extent of hippocampal damage resulting from ibotenic acid lesions is shown in Fig ure 2 3A The largest lesion is represented in gray and sparing from the smaller l esions is represented in black. Numbers located nex t to the diagram in Figure 2 3B represent relative position of the section in mm from bregma. Histological evaluation of the

PAGE 49

49 damag e indicated that all 7 ibotenic acid lesioned mice had neuronal loss which extended through the full extent of dorsal and ventral hippocampus. Transfer learning performanc e In order to evaluate the role of the hippocampus in transfer learning, ibotenic acid and sham lesioned mice were tested in the transfer learning task. The pr ediction was that ibotenic acid lesioned mice would be able to learn the initial associations (learning phase) comparably to sham lesioned mice, but would be impaired relative to shams on the transfer phase of the task, when the irrelevant dimension was ch anged. Consistent with these predictions, both ibotenic acid lesioned and sham lesioned mice performed comparably during the learning phase of the task (Figure 2 4A; t test: (12) = 0.53 n.s.). In contrast, the i botenic acid lesioned mice performed signifi cantly worse than the sham group in the transfer phase, making consistently more errors when the irrelevant dimension was changed (Figure 2 4B; t test: (12) =2.22, p<0.05). Odor detection threshold In order to determine if the impairment observed in the ib otenic acid lesioned mice was due to an inability to detect odors as well as sham lesioned mice, both cohorts were tested in the odor detection threshold tes t. Both sham and Ibotenic acid lesioned mice learned the simple discrimination between full streng th sandal wood and mineral oil comparably (t (12) = 0.43, n.s.; Figure 2 5A). Moreover, a two factor repeated measures ANOV A (dilution X lesion condition) revealed that all mice decreased accuracy of choices as the odor was made more dilute (F(3,36)= 5.03 p< 0.05) but there was no difference between lesion conditions in their ability to perceive and respond to decreasing concentrations of t he odorant (t (12) = 0.31, n.s.; Figure 2 5B).

PAGE 50

50 Morris W ater M aze P erformance Following all other testing procedures, the sham and ibotenic acid lesioned mice were trained in visible (cued) and hidden platf orm versions of the water maze. Swim speed To assess sensorimotor differences between groups, swim speed was assessed on the first trials of visible cued and hidden p latform training. These trials were chosen for swim speed analysis as they are only minimally confounded by learning. No main effect of lesion condition was found on swim speed during the first trial of cue training (t (12) = 1.87, n.s.) or on first h idde n platform training trial (t (12 ) = 0.76 n.s., Figure 2 6B). Cu ed (visible) platform training Both groups of l esioned mice showed comparable pathlengths to reach the visible platform (pathlength averaged across both trials of cue training: t(12)= 1.69, n. s.). Together with the swim speed data, these data indicate that ibotenic acid lesions of hippocampus did not produce any gross sensorimotor or motivational effects that could confound interpretation of performance in the hidden platform version of the tas k. Spatial (hidden) platform training In contrast, as shown in Figure 2 6A, a two factor repeated measure ANOVA (lesion group X trial block) revealed that, while all mice improved over the course of training (F (3, 36) = 3.41, p<0.05), ibotenic acid lesio ned mice performed significantly worse rel ative to sham lesioned mice in the hidden (spatial) version of the task (F (1, 12) = 9.0, p<0. 05). No interaction was observed between lesion condition and trial block (F (3, 36) = 0.60, n.s.). Notably, there was n o difference in probe trial performance between groups, although a repeated factors ANOVA (lesion condition X probe trial) did

PAGE 51

51 indicate a trend such that ibotenic acid lesioned mice tended to spend less time in the target region than did sham lesioned mice (F(1,12)= 2.459, p=0.14). Discussion The overall goals of experiments in Chapter 2 were: 1) to develop a mouse version of the human transfer learning task; and 2) to confirm whether the hippocampus was necessary for transfer learning. Computational modeli ng in humans predicts that the hippocampus plays an important role in transfer learning, which is the ability to apply previously learned information to novel problems and situations 62,65,70 This prediction was directly tested in Chapter 2. A mouse transfer task analogous to the human transfer learning assessment was developed. The mouse task allowed for transfer learning to be assessed in a single session, and all mice success fully completed both phases of the task. Throughout testing, mice were motivated to dig for the food reward (averaging 2 hours to complete the task), and showed no signs of satiation. Data from Experiment 1 show that mice can readily learn multiple intermi xed stimulus reward associations (learning phase). Young C57BLJ/6 mice counterbalanced for comparable performance on the learning phase were further sensitive to the change in the irrelevant feature. This was evident by significantly decreased accuracy of performance in a test condition during which the irrelevant feature was altered compared to performance of mice that only had to recall unchanged associations from the initial learning problems. These data indicate that the transfer phase of the task engag es cognitive processes (and likely neural substrates) beyond those required for recall of simple discrimination information. Also notable from this initial experiment is that although mice performed significantly worse when the irrelevant stimulus feature was altered, they still demonstrated relatively proficient performance on the transfer

PAGE 52

52 phase, providing evidence that this task offers the parametric space necessary to detect a range of decrements mild to severe associated with disease. Data from the co mputer based transfer task performed in humans (described in detail in Chapter 1) show that individuals with hippocampal damage are able to learn a series of complex discriminations with two stimulus features (shape and color), but are impaired in their ab ility to transfer this information to newly configured problems in which one of the features is altered 10 This deficit occurred despite the fact that the feature predictive of the reward (the relevant information) was not changed 10 These data suggest that the transfer learning task is dependent on the hippocamp al region, and that subtle morphological changes (such as minor hippocampal atrophy) can result in impairment in transfer learning 1,10 Experiment 2 demonstrated that the hippocampal region is indeed necessary for transfer learning. Ibotenic acid produced complete lesions that extended the full rostro caudal extent of the hippocampus but that were largely confined to the hippocampal region. As predicted, and comparable to human studies, ibotenic acid lesioned mice were able to learn the associations on par with the sham lesioned mice during the l earning phase of the task. However, when the irrelevant dimension was changed on the transfer phase of the task, ibotenic acid lesioned mice performed significantly worse than sham lesioned mice. This finding is consistent with the idea that the hippocampu s is essential for flexibly utilizing acquired information about the associations between stimuli and rewards during the learning phase and applying (transferring) this information to the new problems p resented in the transfer phase.

PAGE 53

53 It is notable that rem oval of the hippocampus did not entirely impair performance during the transfer learning phase of the task. Instead, it resulted in approximately 30% percent reduction of performance. This finding is not surprising as, over the course of the 30 trials (10 of each problem), it would be expected that new associations would be acquired through new learning in a manner comparable to that used during the learning phase. However, t his finding does highlight one feature of the transfer learning task, which is that detectin g hippocampal related dysfunction. The hippocampus has also been implicated in spatial memory 58,120 and removal of thi s brain region has been shown to impair water maze performance in rats 94 Thus, spatial learning and memory were also assessed in ibotenic acid lesioned and sham lesioned mice by using the hidden version of the water maze task. Ibotenic acid lesions ing the training trials, as evidenced by longer path lengths to reach the platform across days. This result was not due to sensorimoto r deficiencies in ibotenic acid lesion ed mice, as swim speeds were not different during cue training. No differences were observed in percent time or pathlength in quadrant between groups during probe trials, which are known to be a more sensitive measure of memory in rats 96 97 This lack of sensitivity in mice has been observed by other researchers, who have proposed that water maze measures are not sens itive in mice, likely due to an increased use of non hippocampal strategies to find the platform during the hidden version of the water maze 116,129 130 One example of such strategy is the non spatial chaining stra tegy (described by Janus et al., 2004), which can be used as an alternative systematic search strategy. Adoption of any other

PAGE 54

54 systematic search strategy (e.g., chaining, circling) can result in successful location of the hidden escape platform as the mous e can rely on alternative strategies that are non hippocampal dependent 116 Thus, transfer learning may be a more specifi c and sensitive measure of hippocampal function, superior to water maze as an assessment o f hippocampal function in mice. It is worth noting that some initial limitations were identified with respect to the mouse transfer learning task. The dependency on o lfaction renders this task inappropriate for testing anosmic mice, as odor detection impairment could confound transfer learning. Thus, odor detection threshold testing is an important and essential control for interpretation of transfer learning results ( see odor detection results for Experiment 2). Moreover, mice are required to be food restricted to 85% of their free feeding weight, and differences in motivation and metabolism could affect performance in some strains 131 138 For example, in a study assessing the effects of caloric restriction in rats, Carter et al.( 2009) found that restriction resulted in an increase in physical activity, which could be a confounder when interpreting outcomes in behavioral tasks. Thus, it is important that weight restriction is closely monitored during the extent of the transfer task, so that differences in activity and motivation can be kept at a minimum 138 Exp eriments in this chapter describe the development and validation of a mou se transfer learning assessment that is analogous to the human task, providing a behavioral paradigm for assessing early hippocampal dysfunction in mouse models which may in bridge th e gap in translational research between rodents and humans. Thus, the utility of transfer learning as a detection tool for hippocampal dysfunction in a mouse model of AD was assessed in the following chapters (Chapter 3 and 4).

PAGE 55

55 Figure 2 1 Transfer lea rning task in mice. (A) T est apparatus for transfer learning in mice. Brown cylinders represent terracotta pots which were filled with a variety of digging media and scented with different odorants. In this schematic, the rose odor is positive (+) and is p redictive of a food reward buried in the pot. In (B), e xamples of discrimination pairs used in the mouse transfer learning task. In each pair, either the odor or digging medium in the pot is relevant to the correct choice (+) but not both. (C) After reach ing criterion performance on the learning phase (6 correct trials in a row), without signaling, mice are presented with the re configured stimuli shown under the the learning phase remains the same but the irrelevant feature is altered. During the learning and transfer phases of the task, each discrimination pair is presented sequentially and pseudo randomly and correct choices are rewarded with a food reward (chocolate pellet) buri ed in the pot (See Methods section for d etails).

PAGE 56

56 Figure 2 2 Validation of the transfer task for mice. Panel (A) show performance of young C57BL/6J mice (n=10) during learning of three concurrent discriminations with multiple stimulus features. The immediate recall of those (B). In (A), although the two groups of C57BL/6J mice performed identic ally received a change in the irrelevant stimulus feature made significantly more initial discrimin ations (percent error). See text for s tatistical analysis.

PAGE 57

57 Figure 2 3. Hippocampal lesions in C57BL/6J. (A ) Photomicrographs of Nissl stained coronal sections of the hippocampus in sham control (top) and ibotenic acid lesioned mice (bottom), rost ral t hrough ventral hippocampus (B) Histological reconstructions of the smallest sparing ( black ) and largest (grey) hippocampal lesion in serial coronal brain sections. Numbers indicate the relative position of the sections in mm from bregma.

PAGE 58

58 Figure 2 4 Bar graphs show performance of sham lesioned (green bar) and ibotenic acid lesioned (red bar) mice on the transfer learning task. Panel (A) shows errors to criterion (mean S.E.M.) in the learning phase of the task and panel (B) shows percent error (mean S.E.M.) in the transfer phase of the task. While there was no difference in number of errors on the initial discriminations, mice with ibotenic acid lesions were significantly impaired relative to sham lesioned mice on the transfer phase of the task. Se e text for statistical analysis.

PAGE 59

59 Figure 2 5 Odor dete ction threshold testing in sham (green bar) and lesioned (red bar) mice. Errors to criterion (mean S.E.) to learn a novel odor discrimination pair are shown in (A) and percent error of response s to decreasing d ilutions of the odorant in (B). detect the odors decreased with diminishing concentrations of the odorants, nearing chance performance at a 1:100,000 dilution of that used during training. No effect of c ondition was observed in the ability to detect odorants See text for statistical analysis.

PAGE 60

60 Figure 2 6. Water maze performance in sham (green) and Ibotenic acid lesioned (red) mice. Panel ( A ) shows that the path from the start position to the station ary hidden platform decreased across the 6 training days (one block consists of training trials for 2 days) for both groups, with a main effect of con dition, such that Ibotenic acid lesion ed mice performed worse than sham lesioned mice. There was no intera ction of block with condition. Panel (B ) shows that both groups were able to find a visible platform comparably, with similar average swim speeds on the first trial of cue training demonstrating a lack of sensorimotor or motivational differences between g roups. See text for statistical analysis

PAGE 61

61 Table 2 1. Odor and digging media pairs as paired in the transfer learning task Discrimination Pairs C orresponding Irrelevant Dimension Learning phase Transfer phase beads x rubber string jasmine vanilla patchouli x mulberry pompoms cork paper x cut balloon chocolate W atermelon jasmine x vanilla beads rubber string pompons x cork patchouli M ulberry chocolate x watermelon paper cut balloon cut straws x cheese cloth pumpkin cucumbe r cucumber X pumpkin cut straws cheese cloth cut paper x guinea pig bedding lemon rosemary bergamot x sage sequins R affia cut bench pad x wood bedding lavender cinnamon vetiver x geranium yarn alphadri bedding raffia x sponge bergamot S age cinnamon x lavender wood bedding cut bench pad alphadri bedding x yarn vetiver G eranium

PAGE 62

62 Table 2 2. Representative example of the order and sequence of complex stimuli presented in the mouse transfer task. Abbreviations: O=odor ; M= digging media. Stimulus features predictive of reward are marked with an asterisk (*). Each number indicates a novel stimulus. As in the human version of the task, new discrimination pairs were introduced progressively, intermixed with previously lear ned pairs as criterion performance was reached in each pair. After reaching criterion in pair 1, the pot with O1 remained baited and O2 unbaited but the irrelevant dimension was changed from M1 to M5. Note that all odors and digging medium were used only once throughout training (O1 O6; M1 M6) Dimensions Combinations Discriminations : Relevant Irrelevant Pot 1 (+) Pot 2 ( ) Learning phase (pair 1) Odor Media O1* with M1 O2 with M1 Learning phase (pair 2) Medi a Odor M2* with O3 M3 with O3 Learning phase (pair 3) Transfer phase (pair 1) Odor Odor Media Media O4* with M4 O1* with M5 O5 with M4 O2 with M5 Transfer phase (pair 2) Media Odor M2* with O6 M3 with O6 Transfer phase (pair 3) Odor Media O4* with M6 O5 with M6

PAGE 63

63 CHAPTER 3 TRANSFER LEARNING IN A MOUSE MODEL OF A The development of the mouse analog of the transfer learning task (Chapter 2) allows for exploration of whether transfer learning may be sensitive to early disease pathology us ing transgenic mouse model s of chapter, the APPswePS1 transgenic mouse model of early amyloid deposition 1 39 (developed by Jankowski et al., 2004) was used to investigate transfer learning in relation to early AD like pathology. More specifically, the questions investigated were: 1) can transfer learning deficits be detected in a mouse model of AD ( APPswePS1 ) at an age in which plaque deposition is robust in the hippocampus (12 months); 2) can transfer learning impairments be detected prior to visible pathology (3 months); 3 ) how does learning performance compare to performance on another hippocampal dependen t task (Morris water maze)? Choosing a Mouse Model of AD The interaction between the hallmarks of AD (amyloid plaque and neurofibrillary tangles) is complex and not well understood. Transgenic mouse models are a useful tool, providing an opportunity to foc us on a single aspect of neuropathology, which results in better understanding of mechanisms underlying symptoms of disease 140 Furthermore, mouse models offer the ad vantage of observation of pathology within a short lifespan, along with the possibility of a controlled environment 140 Data acquired with transgenic mouse models of AD demonstrate that some A forms can be toxic to neurons and synapses 31,43,47,141 142 that their accumulation precedes other pathological features of the disease (e.g., neurofibrillary tangles) and that they may initiate the progressive d egeneration characteristic of AD 24,36,143 144

PAGE 64

64 Furthermore, animal model studies also indicate th at amyloid plaque accumulation can produce cognitive impairment even in the absence of other pathologies or neuronal loss 108,145 147 These findings support the a myloid beta cascade theory of AD. This theory posits that the deposition of the amyloid parenchyma is a necessary event in Alzheimer's disease pathology and that A process that ultimately leads to the development of other AD pathology ( e.g., neurofibrill ary tangles), glial activation, synaptic dysfunction and ultimately, neuronal loss 148 151 Substantial research efforts are aimed at determining how these pathologies interact, and the mechanisms of how amyloid may be initiating this cascade. The amyloid theory, together with evidence from biomarkers (discussed in Chapter 1) and transgenic studies, suggests that amyloid accumulation is an early event in AD, and that assays that are sensitive to A pathology should detect early, perhaps even asym ptomatic stages of the disease. The APPswePS1 mouse m odel: plaques 40 and 1 42 peptides which are product s of amyloid precursor protein (APP) 152 APP has the characteristic of a cell surface receptor and is expressed in many tissues, in particular in synapses, a s a part of normal metabolism. Although its primary function remains unclear, APP is believed to be implicated in synaptic formation and repair, cell signaling and adhesion 151 Furthermore, mutations on the APP gene and other genes (e.g., PS1 and PS2 ) that affect APP processing can result in overprod uction of A 42, which is a major component of neuritic plaques observed in AD 153 155 As illustrated in Figure 3 1A, t he APP is cleaved secretase both of which release a soluble extracellular fragment ( sAPP). Cleavage of the NH 2 terminus of APP by

PAGE 65

65 secretase (BACE1 or BACE2) 156 158 bound C terminus fragment ( secretase cleavage takes place within the the full secretase cleaved fragment is then cleaved by the secretase complex. If the cleavage is is produced. However, if the cleavag e is after A oligom ers and extracellular fibrils (fibrillar A 2 ) The APPswePS1 double transgenic mouse is a widely used model that develops predictable and progressive age related AD like pathology without frank neuronal loss The APPswe PS1 double transgenic mic e express a chimeric mouse/human amyloid precursor protein (APP; Mo/HuAPP695swe) and the mutant human presenilin 1 (PS1 dE9) both of which are implicated in early onset autosomal dominant forms of AD. The Swedish mutation ( Figure 3 1 B; K595N/M596L), which is linked to familial AD, increases the number of A deposits by stimulating the secretase pathway 159 These mice develop insoluble A deposits (plaques) throughout the brain, including hippocampus, beginning at approximately 6 months with numerous widespread deposits reported by 12 months of age 139,154,160 161 (see T able 4 1) This widespread amyloid pathology is highly reminiscent of the extensive amyloid plaque deposition th at characterizes the AD brain 52,162 Figure 3 3 shows plaq ue deposition in the hippocampu s of APPswePS1 mice at 3, 6 and 12 months of age (Thioflavin S stained plaques Methods described in Chapter

PAGE 66

66 4 ). At 3 months, A 139,144 but large neuritic plaques are mostly absent. However, by 12 months of age, amyloid deposition is robust. Based on the progressive pathology, experiments in Chapter 3 were designed to determine if transfer learning is s ensitive to different aspects of hippocampal amyloid pathology. Specifically, mice were tested on the transfer task at both 3 (pre plaque) and 12 (robust plaque deposition) months of age and further tested on another standard hippocampal dependent task, th e Mor ris water maze, for comparison. Methods Subjects Experiment 1: Is Transfer Learning Sensitive to R obust Plaque Pathology in Aged (12 months old) APPswePS1 M ice? Subjects were a cohort of 12 month old, female Tg (APPswe, PSEN1dE9; APPswePS1; n=11) and age matched non transgenic littermates of B6C3F1/J background strain (NTg; n=7) obtained from The Jackson Laboratory, Maine, USA. Experiment 2: Is Transfer L earning S ensitive to A Elevations Prior to the O nset of P laque D eposition in APPswePS1 M ice ? Subje cts were a second cohort of female APPswePS1 (n= 8) and age matched NTg littermates of B6C3F1/J background (n= 9). Mice were initially tested for transfer learning at 3 months of age an d retested at 12 months of age. Transfer learning task: a pparatus and t ask parameters The testing apparatus for the mouse transfer learning task is illustrated in Figure 2 1, and is described in detail in Chapter 2. Rodent chocolate flavored food pellets (20 mg, AIN 76A from TestDiet) were used for the food reward. To disguis e the odor of the reward, crushed chocolate food pellets were sprinkled over the surface of each pot. The position (left or right) of the rewarded pot varied pseudo randomly across trials. For the

PAGE 67

67 first four trials of every new discrimination problem, mice were allowed to dig in both pots until they obtained the reward (i.e. they were allowed to self correct if they dug in the incorrect pot). On these trials, only their first choice was scored (as correct or incorrect). On trials thereafter, mice were remov ed from the test chamber after only one dig (either correct or incorrect). A dig in a pot was scored if a mouse displaced the digging medium with either its paws or nose. Shaping Shaping took place in the testing box and consisted of training the mice to d ig in two pots, each filled with mixed digging media and with a chocolate pellet reward buried in the bottom. Mice were considered shaped to dig when, after being placed in the apparatus, they reliably retrieved the reward from both pots in less than 3 min for 12 consecutive trials. The day after reaching criterion for shaping, mice were tested on the transfer learning task. Testing Testing was performed as described in Chapter 2. Briefly, mice were trained on a series of three concurrent discrimination pr oblems (learning phase). In each problem learned to discriminate between pots with two stimulus features (odor and digging medium). One of the two pots was baited, with either the odor or the digging medium as the relevant feature. Mice were considered to have acquired the three discrimination problems when they achieved six consecutive choices of the correct (+ = baited) pot, including two correct of each of the three different discrimination pairs. After reaching criterion performance, mice were immediat ely assessed for transfer learning. The mice were presented with 30 trials in which only the stimulus feature that was irrelevant to the reward was changed and the 3 new combinations were presented pseudo randomly.

PAGE 68

68 O dor detection threshold testing A nosmia has been shown to emerge as a consequence of chronological age and 163 Thus, after evaluating transfer learning, mi ce for transfer learning were assessed for their ability to detect and respond to decreasing concentrations of odorants (all APPswePS1 and NTg mice) Odor detection threshold testing was performed immediately following their initial testing in the transfer task (within a week of transfer testing). Cohorts were tested in the same apparatus used for the transfer learning. First, mice were presented with a novel odor discrimination problem (s andalwood vs. mineral oil) using full strength odorant applied directly to the rims of two pots filled with mixed digging media. Sandalwood was the positive odor that predicted the food reward and mice were trained until reaching criterion performance (6 c onsecutive correct trials). Once achieving criterion performance, the mice were tested in a series of discrimination problems during which the sandalwood odor was systematically diluted (1/10, 1/100, 1/1000, 1/10000). Mice were given 16 trials at each dilu tion and the percent error was used as the measure of perf ormance. Experiment 3: Is Water Maze Performance Impa ired in 12 Month O ld APPswePS1 M ice? To directly compare the sensitivity of the rodent assessment of transfer learning to spatial learning abilit y, the APPswePS1 and NTg mice from both cohorts (at approximately 13 months of age) were tested in hidden (hippocampal dependent) and visible cued (non hippocampal dependent) versions of the Morris water maze task. Apparatus The water maze is described in detail in Chapter 2. In short, d uring the spatial reference memory assessment (hidden platform training), a retractable escape platform

PAGE 69

69 (12 cm diameter, HVS Image, UK) was located in the southwest quadrant of the maze surface. During cue training, the tank was analyzed using a computer based video tracking system (Water 2020, HVS Image, UK). Initially, mice were trained in the non hippocam pal dependent cued version of the task in order to assess visual acuity and sensorimotor abilities. Beginning the day immediately following completion of cue training, mice received 6 consecutive days of training (4 trials per day) to find a hidden, statio nary platform, to assess hippocampal dependent spatial reference memory. The fourth swim of days 2 and 6 of hidden platform training were probe trials, during which the platform was retracted to the bottom of the tank for the first 30 sec of the 60 sec tri al. Water maze analysis The water maze data were analyzed using a computer based video tracking system, Water 2020, developed by HVS Image (Hampton, UK). Primary performance measures analyzed were path length (the total distance from the start position to the platform in centimeters) on hidden and visible training trials and percent time in the target quadrant on probe trials. Separate two factor repeated measures ANOVA (genotype x day) were used to evaluate differences across visible platform training tria ls, hidden platform training trials and probe trials. Finally, swim speed (centimeters per second) between groups was also analyzed using a one factor ANOVA (genotype) on hidden and visible trials. Statistical analyses Total number of trials and errors to achieve criterion (6 correct trials on a row) on the last portion of the learning phase where all 3 concurrent pairs are presented was

PAGE 70

70 compared between experimental groups. Due to similar results when total number of trials and errors were analyzed, only n umber of errors was reported and shown. During the transfer phase, in which a fixed number of trials were presented (30 total trials), percent error was the performance measure analyzed. Unpaired t tests were used for analyzes of both phases. Y axis for th e learning phase bar graphs were made equal throughout the document so that performance differences could be easily visualized. For the odor detection thres hold testing, an unpaired t test was used to compare total number of trials and errors during the si mple discrimination of full strength odor versus no odor (mineral oil). Again, errors and total trials to criterion were analyzed, and due to no difference in results, only errors were reported and shown. A repeated measures ANOVA (age x genotype) across o dor dilutions was then used to compare lesion group performance (mean % correct at each dilution) on the progressive dilutions of the full strength odorant. Statistics were performed with StatView software (version 5.01), and in all cases p<0.05 was consi dered significant. The water maze data were analyzed using a computer based video tracking system, Water 2020, developed by HVS Image (Hampton, UK). Primary performance measures analyzed were path length (the total distance from the start position to the platform in centimeters) on hidden and visible training trials and percent time in the target quadrant on probe trials. Separate two factor repeated measures ANOVA (lesion condition/ blocks x day) were used to evaluate differences across visible platform t raining trials, hidden platform training trials and probe trials. Finally, swim speed (centimeters per second) between groups was also analyzed using a one factor ANOVA

PAGE 71

71 (genotype) on hidden and visible trials respectively. Statistics were performed as desc ribed above for each task using St atView software (v ersion 5.01). Results Experiment 1 Transfer learning performance: The goal of experiment 1 was to determine if aged APPswePS1 mice demonstrated transfer learning impairments relative to age matched NTg c pathological changes associated with AD. A pairwise comparison revealed no difference due to genotype on the initial discrimination learning phase of the task (Figure 3 4A; t (17) = 0.36, n.s.). However, the APPswePS1 mice were significantly impaired relative to age matched NTg mice in their ability to perform the discrimination problems when the irrelevant feature was changed during the transfer phase of the task (Figure 3 4B, main e ffect of genotype; t (17) =2.21, p <0.05). Experiment 2 Transfer learning performance: A second cohort of APPswePS1 and NTg mice was assessed longitudinally for transfer learning beginning at 3 months, followed by reassessment at 12 months of age. There were no differences due to genotype on the initial discriminations (Figure 3 5A; t (15) = 0.02, n.s.), nor were they present on the transfer phase at 3 months of age (Figur e 3 5B; t (15) = 0.03, n.s.). These mice were then retested for transfer learning at 12 mo nths of age. As in Experiment 1, APP swe PS1 mice learned the initial discriminati ons on par with NTg mice (Figure 3 6A; t(15) = .24, n.s.) but they were significantly impaired on the transfer phase of the task (Figure 3 6B; t(15) = 2.95, p<0.05).

PAGE 72

72 The simi lar magnitude in transfer learning deficits observed in 12 month old Tg mice across experiments 1 and 2 suggested that there may be minimal savings associated with the transfer learning task. Such a finding could have significant implications for the use o f this task in within subjects experimental designs that are ideal for therapeutic intervention studies. In order to directly compare the magnitude of transfer learning performance of 12 month old APPswePS1 mice with (experiment 2) and without (experiment 1) prior experience in the task, the percentage of correct trials of each mouse from each cohort was divided by the mean of its respective age matched NTg group. A t test indicated that there was no difference in performance between performance deficits in 12 month APPswePS1 mice assessed cross sectionally and longitudinally (Fig ure 3 7; t (18) = 1.38, n.s.). Odor detection testing: In order to determine whether a decreased ability to detect odors with age and/or transgene was a factor in the transfer de f icit observed in the APPswePS1 mice at 12 months of age, odor detection abilities were assessed at different ages in mice from Experiments 1 (12 months) and Experiment 2 (3 months) using identical procedures. In a separate session following transfer learn ing testing, mice were trained to criterion on one additional olfactory discrimination problem. Both young and aged APPswePS1 and NTg mice learned the simple discrimination problem comparably (Figure 3 8A). A repeated measures ANOVA (age x genotype x odor dilutions) revealed no significant differences due to age (F (1, 30) = 0.71, n.s.) or genotype (F (1, 30) = 0.20, n.s.); also, no interaction was observed between age and genotype (F (1, 30) = 0.77, n.s.). Moreover, Figure 3 8B shows that the odor detectio n threshold did not differ across groups, with all groups showing similar declines in

PAGE 73

73 accuracy as concentrations of the odorant were decreased (repeated measures ANOVA; age: F (1,30) = 0.38, n.s.; genotype: F (1,30) = 0.22, n.s.; age X genotype interactio n: F (1,30) = 0.057, n.s. ). Experiment 3 Following all other testing procedures threshold testing, the aged mice from Experiment 2 were trained in visible (cued) and hidden platform versions of the water maze. Repeated measures ANOVAs were used for all ana lyses. As shown in Fig. 3 9A, aged (12 months) APPswePS1 and NTg mice showed comparable path length to reach the visible platform, across days (F (1, 18) = 0.08, n.s.) and there was no interaction between genotype and day (F (1, 18) = 0.16, n.s.). Figure 3 9B shows path length across training days 1 through 6. Both groups improved performance across training (F (5, 90) = 2.92, p<0.05) but there was no main effect of genotype (F (1, 18) = 0.169, n.s.) and no interaction between day and genotype (F (5, 90) = 0.56, n.s.). Probe trial performance is shown in Fig. 3 9C. Percent time in the target quadrant significantly increased for both groups from the early probe trial on day 2 and the last probe trial on day 6 (F (1, 18) = 12.16, p<0.05) but no main effect of genotype (F (1, 18) = 0.55, n.s.) nor interaction between genotype and probe trial (F (1, 18) = 0.26, n.s.) was observed. Finally, no main effect of genotype was found on swim speed during the first trial of cue training (F (1, 18) = 1.35, n.s.) nor on the first hidden platform training trial (F (1, 18) = 2.02, n.s. ) two measures not confounded by learning. Discussion The goal of the studies in Chapter 3 was to determine if transgenic mice that have pathology associated with AD show de ficits in transfer le arning. The APPswePS1 mouse, a widely used model of progressive amyloid deposition, starts developing

PAGE 74

74 plaques in the hippocampal region at around 6 months of age 139,144,164 By 12 months of age, this deposition is robust, and APPswePS1 mice tested for transfer learning at this age were selectively impaired in transfer learning in comparison to age matched NTg mice. Even at the most advanced ages tested here (12 13 months), APPswePS1 mice were able to learn three com pound discriminations concurrently that included combinations across perceptual sets (i.e., on some problems odor was relevant to the correct choice and on others digging medium signaled the correct choice). These data show that APPswePS1 mice are able to form associations across at least two stimulus dimensions. In contrast, when the irrelevant stimulus dimension was altered, aged APPswePS1 mice performed significantly worse than did NTg age matched controls. This deficit was observed both in the initial c ross sectional study conducted on 12 month old mice (Experiment 1) and in the longitudinal study in which transfer learning was comparable to NTgs at young ages (3 months) but deficits of a similar magnitude as in Experiment 1 emerged in the APPswePS1 mice compared to NTgs when these groups were re tested at 12 months of age (Experiment 2). Olfactory deficits are observed in people with mild cognitive impairment (MCI) and 163,165 As anosmi a has been reported at advanced age and in AD, and odor was a key stimulus feature in th is task, mice were assessed for odor detection threshold abilities. There were no age or pathology associated differences in olfactory detection abilities (Fig ure 3 6). Moreover, the salience and difficulty of odor and digging medium problems appeared co mparable as no significant differences in errors or trials to criterion were observed in any group (data not shown). These data

PAGE 75

75 confirm that procedural aspects of the transfer task did not account for the age related deficits observed in t he APPswePS1 mice In Experiment 2, APPswePS1 mice at 12 months of age that were impaired in transfer learning were not impaired in the water maze. As both tasks have been shown to be dependent on the hippocampus 109 110,166 167 it is perhaps surprising that no decrements in performance were observed on the spatial learning assessment. One explanation for this lack of deficit, mentioned in chapter 2, is that mice use different strategies to search for the hidden platform in the wat er maze, and unlike rats, do not need to rely sole on the hippocampus to find it 116 As such, spatial learning (at least as assessed in the water maze) may not be particularly sensitive to hippocampal pathologies in mice. Indeed, ibotenic acid lesioned mice (in which there was virtually complete removal of the hippocampus) showed some decrement in the hidden version of the t ask ( training trials) but had remarkably intact probe trial performance, which is generally considered the more sensitive measure of hippocampal function in this task. These data suggest that transfer learning is a more sensitive detector of hippocampal dy sfunction, at least in the APPswePS1 transgenic mouse model of amyloid deposition 139 O ne of the challenges of translational research has been an inability to extend cross sectional experimental designs traditionally used in rodent experimentation to the interpretation of longitudinal experimental designs typically used in human clinical trials. Tasks such as the water maze are not easily adaptable for re assessment due to significant savings across sessions, even in animals with significant memory impairments In contrast, the longitudinal study presented here demonstrates that the

PAGE 76

76 transfer learning task can be used in a within subject, test retest manner in mice. Direct comparison between 12 month old APPswePS1 mice tested cross sectionally and longitudinally showed that mice tested at 3 months of age were impaired at 12 months, with no savings from prior exposure to the task. Thus, the transfer le arning test is ideal for characterizing hippocampal function in mouse models as pathology progresses with age. The current series of experiments s how that trans fer learning is sensitive to presence of amyloid plaques in the hippocampus of a transgenic mous e model of AD, and suggest that this deficit could be age dependent. This task revealed robust deficits in cognitive function in animals that did not show deficits on more standard assessments of hippocampal function ( i.e. the spatial reference memory ver sion of the Morris water maze task). These data and the highly similar parameters of the human and rodent tasks provide a behavioral paradigm that should be useful for translational research between rodents and humans, resolving a high priorit y need for cu rrent AD research.

PAGE 77

77 Figure 3 1. Amyloid precursor protein (APP) processing. In (A) APP cleavage sites. Alpha secretase cleavage is invo l ved in the non amyloidogenic pathway and results in the re lease of soluble APP fragments. Cleavage by secretase initiates the amyloidogenic pathway and further cleavage by secretase results in the production of A 42. In (B) APPswePS1 mouse model. Double transgenic mice present a mutant human presenilin mutation 1 (PS1d9) and the swedish mutation (K595 N/M596L), whic h result in overproduction of A 42 protein.

PAGE 78

78 Figure 3 2. Amyloid aggregation at the synapses. After the APP is cleaved through the monomeric state does not appear to n eurotoxic, while oligomers and fibrillary plaques have been implicated in synaptic dysfunction and impairment in plasticity 168

PAGE 79

79 Figure 3 3 Photomicrographs show Thioflavin S stained coronal sections of the hippocampus in 3, 6 and 12 month old APPSwePS1 mice. Note that although non fibrillar A age 139,1 44 large neuritic plaques are mostly absent. At 6 months of age, p laque deposition begins to emerge, and is robust by 12 months in APPswePS1 mice.

PAGE 80

80 Figure 3 4 P erformance of aged (12 months) APP swe PS1 ( blue bars) a nd age matched NTg mice ( red bars ) on the transfer learning task. Panel (A) shows errors to criterion (mean S.E.M) in the concurrent discrimination learning phase of the task and panel (B) shows percent error (mean S.E.M) in the transfer phase of the task. Note that while there is no di fference in performance between NTg and APPswe PS1 mice in number of errors on t he initial discriminations, APPswe PS1 were significantly impaired relative to NTg mice on the transfer phase of the task. See text for statistical analysis.

PAGE 81

81 Figure 3 5 Per formance of young (3 months) APPswePS1 (blue bars) and age matched NTg mice (red bars) on the transfer learning task. Panel (A) shows errors to criterion (mean S.E.M) in the learning phase of the task and panel (B) shows percent error (mean S.E.M) in t he transfer phase of the task. N o difference in performance between NTg and APPswe PS1 mice in number of errors on t he initial discriminations or percent error in the transfer phase of the task. See text for statistical analysis.

PAGE 82

82 Figure 3 6 P erformanc e of the same APPswe PS1 ( blue bars) and NTg mice (red bars) shown in Figure 3 5, re tested for transfer learning at 12 mo. Panel (A) shows errors to criterion (mean S.E.M) in the learning phase and panel (B) shows percent error (mean S.E.M) in the tran sfer phase of the task. No significant differences were observed in errors to criterion (mean S.E.M) during initial discrimination learning (A); however, a significant and robust deficit in transfer learning was observ ed in APPswe PS1 mice at this age. Se e text for statistical analyses.

PAGE 83

83 Figure 3 7. Previous exposure to the task did not alter the ability to detect a significant deficit in transfer learning. This was observed by comparison of performance of cross section and longitudinal testing of 12 m onth old APPswePS1 cohorts on the transfer phase of the task. The magnitude of the deficit is virtually identical, providing strong evidence that this task is well suited for within subjects studies.

PAGE 84

84 Figure 3 8. O dor detection threshold testing in yo ung and aged APPswe PS1 (blue bars) and NTg (red bars) mice. Errors to criterion (mean S.E.M) to learn a novel odor discrimination pair are shown in (A) and percent error (mean S.E.M) of responses to decreasing dilutions of the odorant (B). As expected, g concentrations of the odorant nearing chance performance at a 1:10000 dilution. No effects of age or genotype were observed in the ability to detect odorants. See text for statistical ana lysis.

PAGE 85

85 Figure 3 9. Water maze performance in aged APPswePS1 (blue bars) and NTg (red bars) mice. Panel (A) shows that both groups were able to find a visible platform comparably, demonstrating a lack of sensorimotor or motivational differences between groups. Panel (B) shows that the path from the start position to the stationary hidden platform decreased across the 6 training days for both groups, although there was no main effect of genotype nor any interaction of day with genotype. Panel (C) shows p erformance on probe trials early (probe 1) and late (probe 2) in the spatial memory version of the task. All mice spent significantly more time in the target quadrant (containing the platform) with increased training although in agreement with training tri al data, no effect of genotype was observed. See text for statistical analysis.

PAGE 86

86 CHAPTER 4 EARLY DETECTION AND POSSIBLE MECHANISMS Data from human studies suggest that assessment of transfer learning could be a sensitive method for early detection of cogni tive dysfunction in individuals at risk of 1,11 Data from experiments in Chapter 3 support this assertion, as transfer learning deficits were evident in a mouse mod el that recapitulates early AD like pathology (i.e., accumulation of amyloid beta plaques). Notably, while transfer learning deficits were robust in the APPswePS1 mice at an age in which plaque deposition is substantial (i.e., 12 months of age), deficits w ere not detected in this mouse model at an age that precedes the accumulation of neuritic A plaque deposition (3 months of age). However, the previous study did not include intervening ages of mice, nor were A or other pathological measures compared with behavioral performance. In addition, it is possible that transfer learning may be associated with secondary pathological mechanisms mediated by, or downstream, of amyloid Indeed, A can be toxic to both cellular function and signaling, and synaptic dysfu nction is among the strongest neuropathological correlates of cognitive function in AD 31,169 170 Experiment s in Chapter 4 w ere designed to more directly examine the relationship synaptic function, and transfer learning using two transgenic mouse models of amyloid pl aque deposition that have distinct time courses associated with pathology and that have been used previously to investigate distinct mechanisms whereby A could be detrimentally affect ing hip pocampal function 153,171 176 Table 4 1 shows the main characteristics of the two mouse models used in Chapter 4. The APPswePS1 mouse model which is discussed in detail in Chapter 3, is a double transgenic model th at express es a chimeric mouse/human amyloid precursor

PAGE 87

87 protein ( APP ; Mo/HuAPP695swe) and the mutant human presenilin 1 (PS1 dE9), both of which are implicated in early onset forms of AD. These mice develop deposits (plaques) throughout the brain by processing of APP (discussed in detail in Chapter 3). Furthermore the presenilin mutation increases the A 42 to A 40 ratio, facilitating amyloid deposition. Plaques accumulate in hippocamp us, beginn ing at approximately 6 months of age, with numerous widespread deposits reported in this region by 12 months of age 139,154,160 161 This widespread amyloid pathology in hippocampus and throughout the neocortex is analogous to the extensive amyloid plaque deposition that characterizes the AD brain (SeeT able 4 1) 52,162 Pa renchymal and vascular amyloid pat hology, plaque associated dystrophic neurites and microglial activation are all observed in the APPswePS1 model 139 Behavioral deficits in the AP PswePS1 model are observed as early as 6 months in working memory and executive tasks, such as reversal learning 177 178 However, deficits in declarative memory, or hippocampal dependent tasks, have gene rally not been observed reliably until after significant plaque deposition is evident. These deficits generally emerge between 12 14 months of age 103 106,111,179 or even later in the case of spatial learning 103,105 Notably, performance in a modified Barnes Maze, which is also an assessment of spatial learning and memory 140 is an exception, as deficits have been reported as early as 7 months of age in the APPswePS1 mice 106 The second model used in Chapter 4 is the Tg SwDI mouse model. This mouse is made with a construct containing the human amyloid beta precursor protein, APP gene, with the Swedish K670N/M671L, Dutch E693Q and Iowa D694N mutations (triple mutation) 171 These mutations result in an increase in the A 42 to A 40 ratio with

PAGE 88

88 visible vascular plaque deposition starting at 3 months. Although not characterized extensively, in accord with the more aggressive pathology in this model, behavioral deficits have generally been obs erved earlier in Tg SwDI model relative to APPswePS1. For example, Xu et al. (2007) reported that the Tg SwDI mice can present deficits in the Barnes Maze as early as 3 months of age 180 significantly before ages at which deficits have been observed in the APPswePS1 model 108 109,181 Together, these data support that the differing magnitude or downstream targets of A pathology in these models are reflected at the behavioral level. In addition to A accumulation, changes in synaptic activity and reductions in synaptic markers have been well documented in early stages of AD, and are among the features of the disease that best predict cognitive dysfunction in dementia 16 22 Indeed, recent studies suggest that synaptic dysfunction is apparent well before prominent frank neuronal loss, preceding significant cognitive deficits in humans (See Fig. 1 2) 39,182 Synaptic dysfunction may reflect a numb er of alterations in cellular machinery and signaling. These include loss of synaptic contacts (due to loss of neurons or dendrites); decreases in number of transmitter vesicles released from pre synaptic terminal (e.g., glutamate); reduction in expression of synapse related proteins (reduced mRNA expression); and/or loss of mechanisms responsible for restructuring dendritic spines (such as deficiencies in post synaptic mechanisms or changes in dendritic morphology) 120 In particular, alterations in synaptic protein levels in pre and post synaptic terminals have been linked to synaptic dysfunction in AD 3,38 39,183 186 Synaptophysin is a synaptic vesicle prote in of approximately 38kDa and because it is present in virtually all cells that participate in synaptic transmission (discussed in more

PAGE 89

89 detailed below), it is often used as a marker of pre synaptic integrity 58 Indeed, l oss of the pre synaptic protein synaptophysin in the hippocampus has been shown to be an early event in AD and this change correlates well with cognitive impairment in humans 3,38,186 187 and rodents 183 (See Figure 1 1 and Table 4 1) PSD 95 (105 kDa) is a major postsynaptic scaffold protein at hippocampal excitatory s ynapses and it is often used as a marker of synaptic strength 1 88 PSD 95 levels also decline in AD, but generally much later in the time course of the disease compared to synaptophysin 189 suggesting that the pre synaptic circuitry may be affected early in the disease The extent to which synaptic alterations may be affecting cognition at very early stages of amyloid beta deposition in AD is unknown. T he lack of early and reliable assessments of hippocampal function is a major limitation to understanding the effects of progressive pathological changes, synaptic alterations and hippocampal dysfunction. The recently developed rodent t ransfer learning task may be a useful tool for behavioral characterization and further exploring such relationships. The goal of experiments in Chapter 4 was to examine the relationship between early AD pathology ( synaptic function) and transfer learning using APPswePS1 and Tg SwDI mice. Methods Subjects Breeder pairs were acquired from the Jackson Laboratory, Maine, USA and bred approved barrier facility. For APPswePS1, a cohort was formed by hemizygous crossing between B6C3F1/J and B6C3 Tg(APPswe,PSEN1dE9)85Dbo/J). For Tg SwDI, a cohort was formed by homozygous crossing of C57BL/6 Tg(Thy1 APPswDuIowa) BWevn/J. F 1 pups from hemyz ogous APPswePS1 breeders were genotyped at the Genetic

PAGE 90

90 Analysis Lab, ICBR, at 2 months of age and controls for Tg SwDI were female age matched C57BL/6J female mice acquired from Jackson Laboratory. For behavioral testing, mice were transferred to the McKni ght Brain Institute (University of Florida College of Medicine) at least one month prior to the onset of behavioral assays where they received a d lib access to food and water. Mice were maintained on a 12 h light/dark cycle (lights on at 0800) and climate controlled at 25 o C. All testing was conducted during the light cycle and mice in the study were screened daily for health problems as were sentinel mice that were housed in the same room as experimental mice M ice were individually housed throughout testi ng and one week prior to behavioral testing, mice were food restricted to 85% of their free feeding weight and handled for 3 5 minutes daily to habituate to the experimenter All animal procedures were conducted in accordance with approved institutional a nimal care procedures and NIH guidelines The number of mice in each cohort is s pecified under each experiment. Experiment 1: To W hat Extent are Transfer Learning Deficits A ssociated with A P athology in APPswePS1 and Tg SwDI M ice? Amyloid beta and transfer learning were assessed in cohorts of 3 and 6 month old APPswePS1 and Tg SwDI mice which reportedly have distinct time courses of A associated pathology 139,144,171,190 Transfer learning task Subjects used to assess transfer learning performance were APPswePS1 [ 3 month ol d (n= 7) and 6 month old (n= 8)], age matched non transgenic littermates of B6C3F1/J background strain [3 month old (n= 7) and 6 month old (n= 6)], Tg SwDI [ 3 month old (n=8) and 6 month old (n=6 ) ], and age matched non transgenic background

PAGE 91

91 strain C57BL/6J [3 month old (n= 6) and 6 month old (n= 6)] mice The testing procedures used for assessing transfer learning are identical to that shown in Figure 2 1, and de scribed in detail in Chapter 2. Tissue harvest logy was assessed in APPswePS1 [ 3 month old (n=5), 6 month old (n=7) ] and Tg SwDI [ 3 month old (n= 3) and 6 m onth old (n=6) ] mice. Mice were perfused with PBS and brains were dissected from the cranium and hemisected along the midline. The hippocampus was dissected from one hemisphere and frozen in nitrogen and stored at Linked Immunosorbent Assay (ELISA). The remaining intact hemisphere was prepared for histological analysis of plaque deposition by Thioflavin S staining. Hemibrains were immersion fixed for 48 h at 4C in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer, pH 7 .4 ,cryoprotected in 20% sucrose dissolved in PBS for 48 hours, rapidly frozen on powdered dry ice, and stored at 80C until further processing. Homogenates were prepared from the dissected h ippocampi in order to quantify A 40 and 42 levels using ELISA T issue was first homogenized using a tissue homogenizer in radioimmunoprecipitation assay ( RIPA ) b uffer ( 50mM Tris HCl pH 7.4, 150 mM NaCl, 1% NP 40, 1% s odium d eocycholate, 0.5% SDS) in the presence of (0.5 M). Sample s were then sonicated and centrifuged at 13,000g for 20 minutes at 4 o C. Protein concentration was determined using the Pierce BCA Kit according to the manufacturer 's protocol (Rockford, IL, USA) and an iMARK Microplate Absorbance Reader (BioRad) was used to detect protein concentrations at 550 nm. A portion of the r esultant supernatant was collected and

PAGE 92

92 stored in the 80 C freezer for use in immunoblotting assays described below in Experiment 2. Another portion of the supernatant was further vacuum centrifuged at 100,000 g for one hour at 4 C, and the supernatant from this centrifugation was collected, representing the RIPA soluble fraction for the A analyses with ELISA. Enzyme linked immunosorbent a ssay (ELISA) g capture and AB5 HRP detection were used (generously provided by Drs. Yona Levites and Todd Golde) Plates were then removed from plates and blocked with ACE blocking solution (1% Block Ace, 0.05%NaN3, 10X PBS at pH 7.4) at 4C overnight. RIPA Homogenates aliquots were further diluted 1:5 for A 40. After ACE blocking was removed, NaH 2 PO 4 NaN 3 EDTA, NaCl, BSA, CHAPS lysis buffer at pH 7.0) was immediately added to the plate to prevent drying to duplicate wells. The plate was then incubated overnight at 4C. Fluid was discarded at 1:1000 concentration (diluted in d etection buffer; NaH2PO4, Na3PO4, thimerosal, EDTA, NaCl, BSA at pH 7.0) and incubated overnight at 4C. Plates were washed with PBST, PBS, and TMB (developing solution; 1 M tris Base, 0.5 M Na2HPO4, peroxide) was added to each well. Stopping buffer (85% p hosphoric acid) was added when optimal development was achieved (color in at least 3 standards). Plate was read at

PAGE 93

93 based on the initial weight of brain tissue. Thioflavin(S) staining Numbers of A plaques were evaluated in the opposite hemisphere of brains used for ELISA analysis of A 40 and 42. Cryo protected hemi brains were sectioned at 40 microtom e. Sections (1 6 series) were matched across mice and were collected into 0.1 M TBS, and mounted on glass slides. After 24h, slides were processed for Thioflavin S (Sigma) staining. Slides were rehydrated in running water, incubated for 5 min in KMnO4 (0.2 5%), for 5 min in K2S2O5 /Oxalic Acid (1%), and for 8 min in Thioflavine S (0.02%). Next, slides were differentiated twice in ethanol (80% for 1 minute each time) followed by water (5 minutes). Prolong Gold with DAPI was used during cover slipping. Images of staining in 6 sections from each animal were captured using the Scanscope XT image scanner (Aperio Technologies Vista, CA). ImageScope was used to assist in quantification of total plaque number (sum) in the se 6 sections of the hippocampus 191 193 The final images and layouts were created using Photoshop CS2 (Adobe, San Jose, CA). Statistical A nalyses Transfer learning assessment For the learning phase, the number of errors to reach criterion on the last bl ock of trials (that included all three intermixed problems) was the primary measure of interest. For the transfer phase, in which a fixed number of trials were presented, percent error was the performance measure analyzed. Total and percent error on lear ning and transfer phases, respectively, were compared for each mouse strain using separate two

PAGE 94

94 factor ANOVAs [ age (3 vs 6 months) X genotype (APPswePS1vs B6C3F1/J control or Tg SwDI vs C57BL/6J control) ] To directly compare performance between APPswePS1 a nd Tg SwDI, performance on trans fer learning was expressed as a percent increase in errors relative to the mean performance of the appropriate age matched non transgenic control mice (NTg) and was compared with two factor ANOVA [ age (3 vs 6 months) X genot ype (APPswePS1 vs Tg SwDI) ] A pathology Levels of transgenic models using two factor ANOVAS [ age (3 vs 6 months) X genotype (APPswePS1 vs Tg SwDI) ] Separate one factor ANOVAs were used to compare the sum of plaques in t he hippocampus from 6 sections matched across subjects, so that regions counted were equivalent, between 3 and 6 month old APPswePS1 mice (note that this analysis was not done in Tg SwDI mice). For pla que counts, one way ANOVA (age X total plaque number) w as used to compare number of plaques in hippocampus of or sum of hippocampal plaque number with transfer learning performance. For all experiments, statistics were performed using StatView software (v ersion 5.01) and p<0.05 was co nsidered significant. Experiment 2: Do Alterations in Markers of Synaptic Integrity Accompany Transfer Learning D eficits APPswePS1 and Tg SwDI M ice ? Substantial evidence links A pathology and cognitive decline in early AD to synaptic loss and dysfunction 29,31,34,49,153 In Experiment 2, expression of pre and post synaptic proteins was evaluated in APPswe and Tg SwD I mice to determine the extent

PAGE 95

95 to which impairments in synaptic integrity accompany transfer learning deficits in these two models. Western blotting Hippocampal expression of synaptophysin and PSD 95 was evaluated by Western blotting using APPswePS1 [ 3 month old (n=5), 6 month old (n=7), 12 month old (n=3)], B6C3F1/J [ 3 month old (n=5), 6 month old (n=5) 12 month old (n=3)]; and Tg SwDI [ 3 month old (n=3) and 6 month old (n=5)] and C57BL/6J [ 3 month old (n=5) and 6 month old (n=5)] mi ce. Each experiment was designed to address specific questions defined a priori to ensure that all samples used for comp arisons were processed together on a single gel. This approach served to minimize the effect of inter assay variability Experiments were performed at room temperature P roteins were denatured and reduced in Laemmli mercaptoethanol (Fisher, Pitts burgh, PA, USA) and heated at 100 C for 10 minutes. Twenty of protein per lane were electrophoretically separated on a Tris glycine gel (4 15%) at 200 V for 35 minutes and transferred to Millipo re Immobilon FL PVDF membrane pre incubated in blotting buffer containing trizma base and glycine using a semidry Minnie Gennie electrophoretic transfer apparatus ( Idea Scientific MN USA) for 1 hour at 350 mA Blots were washed 3 times with tris buffered saline (TBS; pH 7.4) then blocked for 1 hour in blocking with 5% milk in TSBT (Tris buffered saline/ with 0.01%) The following antibodies were us ed: mouse monoclonal [SY38] to s ynaptophysin (ABCAM, 1:5,000 overnight incubation), rabbit polyclonal to PSD 9 5 (ABCAM, 1:5,000), chicken polyclonal antibody to glial fibrillary acid protein (GFAP ; EnCor, 1:2,000), and a mouse monoclonal antibody to glyceraldehyde 3 phosphate dehydrogenase (GAPDH; EnCor, 1:10,000). All primary antibodies were incubated overnight a t 4 C with rotation. Binding

PAGE 96

96 was detected using appropriate LI COR IRDye secondary antibodies conjugated to 80 0CW (1:10,000). All secondary incubation s w ere done at room temperature for 1 hour with rotation. Infrared signal was detected wi th the Odyssey I mager system (LI COR ) and measured wi th Image Studio software (LI COR ). Synaptophysin, PSD 95, and GFAP values were expressed as a ratio of GAPDH measured in the same lanes in order to control for loading variat ion. Statistical analyses Levels of synaptop hysin, PSD 95 and GFAP were compared for each mouse strain transgenic strain using one factor ANOVAS [ age (3, 6 and 12 months; APPswePS1) or age (3 vs 6 months); for Tg SwDI) ] Age matched controls were analyzed the same way [age (3,6 and 12 months; B6C3F1 /J) or age (3 vs 6 months; C57BL/6J). Pairwise comparisons were used to compare total p rotein levels in the hippocampal homogenates GAPDH was used as a loading control as there was no age related change in expression of this protein ( pair wise comparisons (age expression levels). Values on x axis were arbitrary and do not infer comparison between graphs. For all experiments, statistics were performed using StatView software (v ersion 5.01) and p<0.05 was considered significant. Experiment 3: Is There O the r Evidence for Synaptic Dysfunction in 6 Month Old APPswePS1 M ice? Data from Experiment 2 (described below) revealed a small but significant reduction in synaptophysin expression in the hippocampus of 6 month old APPswePS1 mice. To determine if this reduc tion had functional implications, synaptic physiology was evaluated in hippocampal slices from APPs wePS1 and age matched NTg mice.

PAGE 97

97 Subjects were non behaviorally characterized APPswePS1 (n=3) and age matched NTg (n=3) mice at 6 months of age. Mice were rap idly decapitated without anesthesia, hippocampi were dissected, and slices ( alvear fibers using a v ibratome (Technical Products International, St. Louis, MO). Slices were transferred to a standard interface recording ch amb er that was continuously perfused (1 ml/min) with oxygenated artificial cerebrospinal fluid (ACSF) containing (in mM) 124 NaCl, 2 KCl, 1.25 KH 2 PO 4 2 MgSO 4 2 CaCl 2 26 NaHCO 3 and 10 glucose. Slices were maintained at 30 0.5 C, and humidified air (95% O 2 5% CO 2 ) was blown over the slices. Electrophysiological recordings: Figure 4 1A shows schematic of e xtracellular synaptic field potentials recording from CA3 CA1 synaptic contacts Recording was done with glass micropipettes (4 concentric bipolar stimulating electrodes (outer pole: stainless steel, 200 m diameter; inner pole: platinum/iridium, 25 m diameter, FHC, Bowdoinham, ME) were positioned approxima tely 1 mm from either side of the recording electrode localized in the middle of the stratum radiatum. A single diphasic stimulus pulse of 100 sec was delivered from a stimulator (SD 9 Stimulator, Grass Instrument Co, West Warwick, RI) to the Schaffer col lateral commissural pathway, in order to evoke field potentials at 0.033 Hz. The signals were amplified, filtered between 1 Hz and 1 kHz, and stored for off line analysis. For analyses, two cursors were placed around the initial descending phase of the exc itatory post synaptic potential (EPSP Figure 4 1B ) waveform, and the maximum slope (mV/ms) of the EPSP was determined by a computer algorithm that found the maximum change across all sets of 20 consecutively recorded points (20 kHz sampling

PAGE 98

98 rate) between the two cursors. For examination of synaptic plasticity, responses were collected for at least ten minutes prior to patterned stimulation to ensure a stable baseline before induction of synaptic plasticity. Long term potentiation (LTP) was induced by emplo ying four 1 second trains of 100 Hz with each train 10 seconds apart. Changes in transmission properties induced by patterned stimulation were calculated as the percent change from the averaged response collected during baseline. For synaptic plasticity st udies, the mean percent changes in the slope of the extracellular synaptic response were measured 55 60 min after 5 Hz stimulation for both control and tetanized pathways. Statistical A nalyses A two factor repeated measures ANOVA was used to compare the e ffects of genotype on synaptic plasticity. Post hoc analyses were conducted using Scheffe tests, with significance set at p < 0.05. Results P ath ology in APPswePS1 and Tg SwDI M i ce? Initial behavioral analyses compared transfer learning (learning and transfer phases) within each mouse model Transfer learning performance in APPswePS1 mice Transfer learning pe rformance is shown in Figure 4 2 On the learning phase, no main effects of age (F (1 24) = 2.28, n.s.) or genotype (F(1,24)= 2.996, n.s) were observed, nor was there an interaction ( age X genotype; F(1,24)= 1.94, n.s). In contrast, comparisons of performance o n the transfer phase (Figure 4 2 A), revealed significant main effe cts of age (F (1, 24 ) = 13.88, p <0.05) and genotype ( F ( 1, 24) = 9.36, p<0.05)

PAGE 99

99 as well as a significant interaction between age and genotype (F (1,24) = p<0.05). To further explore the nature of this interaction, pair wise comparisons were conducted. The se analyses revealed that 6 month old APPswePS1 were significantly impaired relative to both age matched NTg controls (t test: t (12) = 3.850, p<0.05) and 3 month old APPswePS1 mice (t (13) = 3.882, p<0.05). Notably, accuracy of transfer learning perform ance did not differ between 3 month APPswePS1 and 3 month NTg (t (12) = 0.47, n.s.). Transfer learning performance in Tg SwDI mic e In agreement with data from the APPswePS1 model, a two factor ANOVA revealed no main effects of age (F (1, 22)= 0.73, n.s.) o r genotype (F (1, 22) = 4.0, n.s.) on the learning phase of the task. In contrast, comparisons of performance on the transfer phase revealed main effects of age (F (1, 22) = 6.81, p<0.05) and genotype (F (1, 22) = 22.55, p<0.05). Unlike in the APPswePS1 mo del, however, there was no interaction ( age X genotype; F (1, 22) = 1.48, n.s. Fig. 4 2 B] Post hoc comparisons indicated that the lack of interaction was due to the fact that performance of Tg SwDI mice on the transfer phase was significantly impaired com pared to NTg mice at both 3 and 6 months of age (3 months: t (12) = 2.7, p<0.05; and 6 months (t (10) = 3.886, p<0.05).Notably, pair wise comparison also revealed that 6 month old Tg SwDI were significantly impaired compared to 3 month old Tg SwDI (t (12) = 2.26, p<0.05). Comparison of transfer learning acro ss APPswePS1 and Tg SwDI models Transfe r learning performance was directly compared across models and results are shown in Figure 4 2 C. For these comparisons, learning (total errors to criterion) and tra nsfer (% error) performance for each transgenic subject was calculated as a percentage increase in error from the mean performance of the appropriate age

PAGE 100

100 matched non transgenic control NTg (i.e., percent increase in error relative to age matched NTg ). Perf ormance on the learning phase was first com pared using a two factor ANOVA [ age (3 vs 6 months) X genotype (APPswePS1 vs Tg swDI) ] There were no main effects of age (F (1, 25) = 0.2, n.s.), genotype (F (1, 25) = 0.86, n.s.) nor was there an interaction bet ween age and genotype. In contrast, comparisons of performance on the transfer phase revealed differe nces between transgenic models. A two factor ANOVA [ age (3 vs 6 months) X genotype (APPswePS1 vs Tg swDI) ] indicated that there was no main effect of age ( F (1, 25) = 0.959, n.s.) nor was there an interaction between age and genotype (F (1, 25) = 1.012, n.s.). However, there was main effect of genotype (F (1, 25) = 11.546, p<0.05), which indicated that Tg SwDI performance on the transfer phase of the task wa s significantly worse APPswePS1mice. Post hoc comparison between strains at each age revealed that 3 month old Tg SwDI had a lower accuracy in performance compared to 3 month old APPswePS1 ( unpaired t test: t (13) = 2.66, p<0.05). The same result was foun d with 6 month old cohorts, with 6 month od Tg SwDI presenting low accuracy in the transfer phase compare to APPswePS1 (unpaired t test: t (12) = 2.27, p<0.05). levels (ELISA) Figure 4 3A shows A 42 levels in APPswePS1 and Tg SwDI mice. A two facto r ANOVA (age x genotype) performed on A 42 levels measured from APPswePS1 and Tg SwDI mice revealed a significant main effect of age (F (1,16)= 50.89, p<0.05) and genotype (F(1,16)= 46.33, p<0.05) as well as a significant interaction (F (1,16)= 37.82, p< 0.05). Further analysis showed that levels of A 42 were higher within strain, with 6 month old APPswePS1 showing higher levels (t (10)= 2.62, p<0.05) and 6 month old Tg SwDI (t(7)= 5.55, p<0.05) than 3month old cohorts of each strain. Post hoc

PAGE 101

101 compariso ns indicated that A 42 levels in Tg SwDI mice were higher re lative to APPswePS1 (trend at 3 months: t (6) = 2.13, p=0.07; 6 months: t (10) = 8.25, p<0.05 ). SwDI mice were two fold greater at 3 months and seven fold greater at 6 months. A 40 levels (ELISA) Similar results were observed for A 40. A two factor ANOVA (age X genotype) performed on A 40 levels measured from APPswePS1 and Tg SwDI mice revealed a s ignificant main effect of age ( F (1,16)= 60.18 4, p<0.05), genotype (F (1,16)= 58.02, p<0.05) and an interaction of age and genotype (F(1, 16)= 42.48, p<0.05) Further APPswePS1 showing higher levels (t (10)= 2.65, p<0.05 ) and 6 month old Tg SwDI (t (7)= 2.78, p<0.05) than 3 month old cohorts of each particular strain. Post hoc comparisons between strains SwDI relative to APPswePS1 was only evident at 6 months of age (3 months: (t(6)= 1.62, n.s.; 6 months: t (10)= 9.27, p<0.05 ) Figure 4 3B and C show representative Thioflavin S staining in the hippocampus of 6 m onth old and APPswePS1 (Fig ure 4 3B) and Tg SwDI (Fig ure 4 3 C). Note, that in agreement with relatively higher A 42 le vels in Tg SwDI mice, a larger number of plaques were also observed in Tg SwDI relative to APPswePS1 mice. This pattern of increased plaque deposition in Tg SwDI relative to APPswePS1was qualitatively evident across animals at both 3 month and 6 month ages

PAGE 102

102 Relationship between A pathology and transfer learning performance in APPswePS1 mice. in hippocampal homogenates and 42, pmoles/g) and counts of Thioflavin S stained fibrillary plaques, in 3 month old (n=4) and 6 month old (n=6) APPswePS1 4 A) wer e highly correlated with the sum number of pla ques quantified in 6 sections of the hippocampus (matched across mice) Moreover, correlation analyses indicated that, among APPswePS1 mice, both of these measures of A pathology was strongly related to performance on the transfer phase (A 42 and transfe r lea rning: r=.79, p<0.05 (Figure 4 4 B); plaque number and transfer learning: r=0.80, p<0.05 (Figure 4 4 C). A Experiment 2: Do Alterations in Markers of Synaptic I ntegrit y Accompany Transfer Learning Deficits APPswePS1 and Tg SwDI M ice? To investigate synaptic integrity, hippocampal expression of synaptophysin and PSD 95 was evaluated by western blotting. GFAP (glial fibrillary acidic protein) expression was also investig ated in these same mice, as this increase in astroglial marker s has been observed in pathological and normal aging 175,194 196 GAPDH was used as a loading control for the APPswe PS1 and Tg SwDI cohort comparisons, a s there was no effect of age (APPwePS1: F (1 1 1) = 0.17 p >0.05 ; Tg SwDI: F ( 1, 8 )= 0. 04 p >0.05). E ffects of chronological age on synaptophysin and PSD95 expression The effects of chronological age on synaptophysin and PSD 95 were compared in B6C3/J mice (the age matched NTg mice for APPswePS1 cohort) and C57BL/6 mice

PAGE 103

103 (the age matched NTg for the Tg SwDI cohort) separately One factor ANOVAs revealed no main effect of age in synaptophysin (B6C3/J: F(1,10)= 0.13, n.s.; C57BL/6: F(1,8)= 0.72, n.s.) or PSD95 (B6C3/J: F(1,10)= 1.84, n.s.; C57BL/6: F(1,8) = 0.08, n.s.) expression in either strain. Notably, GFAP expression evaluated in these same samples also did not differ with age in the non transgenic mouse strains (one factor ANOVA; B6C3/J: (F (1 10) = 1.21 n.s.; C57BL/6 : F (1, 8) = 0.70, n.s. ). Synaptic protein expression in APPswePS1 mice and Tg SwDI The effects of progressive A pathology on hippocampal synaptophysin and PSD 95 expression were compared separately in APPswePS1 and Tg SwDI mice using one factor ANOVAs (age). APPswePS1 model Representative immunoreactive bands of synaptophysin, PSD 95, GFAP and GAPDH observed for 3 6 and 12 month old APPswePS1 hippocampal samples are shown in Figure 4 5 A. A one factor ANOVA revealed a significant main effect of age in APPswePS1 mice (F (1, 11) = 4.49 p <0.05) such that synaptophysin expression was significantly decreased ( 25 %, Fi g 4 5 B) in 6 and 12 month old APPswePS1 relative to p <0.05 ). Notably, no differences were evident between 6 and 12 month old APPswePS1 mice (n.s.). In addition to decreased synaptophysin expression, there was also a trend toward reduced PSD 95 expression in APPswePS1 mi ce with increasing age (Fig ure 4 5 C), although this did not quite reach statistical significance (F (1,11)= 3.42, p=0.07) Subsequent planned comparisons performed on PS1swePS1 mice of different ages revealed that, although 3 and 6 month old levels of PSD 95 did not differ statistically (3

PAGE 104

104 months: t (8)= .833, ns), PSD 95 expression was significantly reduced ( 35%) in 12 month old relative to 3 month old APPswePS1 ( t(6)= 1.99, p<0.05). Notably, PSD 95 express ion did not differ 3 and 6 month old groups (t (10) = 0.85, n.s.) As expected, given that higher levels of pathology in APPswePS1 mice were observed with increased age, a one factor ANOVA (age) revealed a main effect of age on GFAP expression (F (1 11) = 3.921, p=0.05 ; Fi g. 4 5 D ). Post hoc comparisons indicated that while GFAP expression did not significantly differ between 3 or 6 month old APPswePS1 mice (n.s.), significantly greater expression (+40%) was observed in 12 month old relative to 3 month old (t(6) = 2.58, p<0.05) and 6 month old (t(8)= 1.58, p<0.05) APP swePS1 mice (Figure 4 5 D). Tg SwDI model Representative immunoreactive bands of GFAP and GAPDH observed for 3 and 6 month old Tg SwDI hippocampal homogenates are shown in Figure 4 6 A Separate t tests revealed no main effect of age on synaptophysin (t (6) = 0.3, n.s.) or PSD 95 (t (6) = 0.25, n.s.) expression in Tg SwDI mice. Notably, as shown in Figure 4 6 B, GFAP was markedly elevated (+40%) in 6 month relative to 3 month old Tg SwDI mice (t (6) = 2.18, p<0.0 5). These effects of transgene on GFAP expression were subsequently confirmed in a separate gel comparing the Tg and NTg mice from this model. An unpaired t test revealed a main effect of genotype (t (8) = 2.47, p<0.05), indicating that elev ated GFAP was evident by 3 months of age in Tg SwDI mice Experiment 3: Is There Other Evidence for Synaptic Dysfunction in 6 Month Old APPswePS1 M ice? To measure the influence of over expression of amyloid beta on synaptic transmission, extracellular syna ptic field potential recordings were performed at CA3

PAGE 105

105 CA1 hippocampal synapses in slices obtained from 6 month old APPswePS1 and age matched NTgs (n indicates the number of slices). I nput output (IV) curves, paired pulse facilitation (PPF), and long term p otentiation (LTP) were recorded. As shown in Figure 4 7 A (IV Curve) a sample t test showed that the magnitude of EPSPs (excitatory post synaptic potentials) is significantly decreased (t (16) = 5.56, p<0.05) in APPswePS1mice ( 1.350 0.058 n=11 ) compared to controls ( 1.897 0.092, n=11 ) indicat ing a significant decrease in baseline synaptic transmission in the APPswePS1 mice A repeated meas ures ANOVA showed, in Figure 4 7 B, that fib er potential was not different between APPswePS1 (0.520 0.133, n=11) a nd control mice (0.696, 0.219, n=1; F (1, 20) = 1.88, n.s.). Figure 4 7 C shows that paired pulse stimulation resulted in a similar level of PPF for APPswePS1 (121.52 2.41, n= 11) than age matched control mice (126.85 2.16, n=11; F (1, 18) = 0.18, n.s.). Figure 4 8 A shows LTP in APPswePS1 and control mice. High frequency stimulation induced LTP in slices obtained from both APPswePS1 (t (8)= 3.16, p<0.05) and controls ( t(7)= 5.33 ; p<0.05). However, as s hown in Figure 4 8 B, a one way ANOVA revealed that the re was a tendency for a decrease in the amplitude of LTP in APPswePS1 mice (123.97 5.36, n=9) compared to age matched controls (131.90 5.55, n=8; F (1,15)= 3.76, p=0.07) Further analysis demonstrated that the last five minutes of LTP induction APPswePS1 s howed decreased EPSP (117.38, 5.58, n= 11) than controls (133.36, 6.42; F (1, 20) = 5.82, p< 0.05 not shown ). D iscussion The overall goals of experiments in Chapter 4 were to extend results from Chapter 3, and further examine the relationship between A D related A pathology and

PAGE 106

106 hippocampal dependent transfer learning Two different transgenic models which overproduce A but with differing time courses, were used to compare the emergence and severity of transfer learning deficits. Synaptic integrity was also evalu ated to begin to examine the mechanism through which A might media te transfer learning deficits. Transfer Learning D eficits in APPswePS1 In agreement with the findings from Chapter 3, no deficits were observed in the APPswePS1 mice rela tive to age matche d NTg at 3 months of age. Notably, the findings in Chapter 4 significantly extended those of Chapter 3 by showing that robust transfer learning deficits emerge in the APPswePS1 mice by 6 months of age age at which plaques begin to develop These transfer learning impairments represent some of the earliest cognitive deficits that have been detected in this widely used A over expressing mouse model of AD. While APPswePS1 mice have been shown to exhibit modest deficits on a version of the Barnes maze at 7 months of age 197 the vast majority of hippocampal dep endent performance deficits do not emerge until much later ages. Indeed, the earliest spatial learning impairments reported in this model are at 14 months of age, and even at later ages, the presence and magnitude of spatial deficits varies substantially a cross studies 109,166 These discrepancies in the emergence of impairments in different hippocampal dependent tasks may reflect the extent to which either A directly or indirectly (through t ask ) mediate different aspects of hippocampal function For example, transfer learning oligomers, where as spatial learning performance may depend more heavily on 198 Alternatively,

PAGE 107

107 and perhaps more likely, the robustness in transfer learning impairments at an early age in the APPswePS1 model may reflect the fac t that this task is less subject to alternative performance strategies than are spatial learning tasks 116,125,129 Indeed, as discussed in Chapter 3, although the water maze is an accurate measure of hippocampal fu nction in rats, it is likely that other non hippocampal strategies can be effectively recruited to improve performance on the water maze task. This has been an issue particularly when assessing spatial learning and memory in mice, as they have a higher ten dency to float and use the other strategies for locating the platform 116, 129 These include, for example, swimming at a fixed distance from the side of the tank (i.e, thigmotaxis) by varying the distance from the wall systematically, mice can effectively reduce their search time for the platform without ever truly learning the spatial location 116 This p ossibility, then, suggests that the early hippocampal dysfunction associated with the APPsweP S1 model can be overcome or masked during performance in spatial tasks by non hippocampal supported strategies, and it is only when brain regions supporting these alternative strategies are also compromised (>12 months of age) that deficits emerge 109,199 In either case, evidence from experiments in this dissertation support that transfer learning may be a particularly useful task for detecting ear ly hippocampal dependent deficits in the APPswePS1 mouse model of AD and, more broadly, may be sensitive to accumulating hippocampal A which has been identified as the earliest event in the AD pathological cascade 25,36,200 Comparison of Transfer Learning D eficits in APPswePSI and Tg SwDI M ice To further assess the role of A pathology on transfer learning, performance in the APPswePSI mouse model was compared to performance in a second A over

PAGE 108

108 expressing model, t he Tg SwDI mouse. While both of these strains exhibit progressive A pathology, previous work has shown, and the current experiments support, that the amyloid deposition observed is accelerated and more robust in the Tg SwDI model 171 The prediction for Experiment 1 in Chapter 4 was that if transfer learning deficits were particularly sensitive to overall levels of A pathology, either soluble or ins oluble fibrillar A deficits should be evident at earlier ages in Tg SwDI mice and should be more pronounced overall relative to APPswe PSI mice. Unlike the APPswePS1 mice, robust deficits in transfer learning were evident in T g swDI mice at the earliest age examined (3 months). Analysis of hippocampal A 40 and 42 further indicated that A levels (RIPA soluble) were significantly (two fold) greater in 3 month Tg SwDI mice relative to 3 month APPswePS1 mice, supporting that the earlier performance deficit s in Tg SwDi mice may be consequent to more robust A pathology. Indeed, across ages, both A 42 le vels and transfer learning deficits were of a greater magnitude in Tg SwDI relative to APPswePS1 mice. Together, these findings demonstrate that performance on the hippocampal dependent mouse transfer learning task developed in Chapter 2 is sensitive to early stage of A pathology in which soluble A levels are prominent but fibrillary A plaques are just beginning to accumulate in two separate mouse models o f A over expression. At ages examined in Chapter 4, both A 42 and the numbers of Thioflavin S stained plaques (fibrillar A ) in the hippocampus strongly correlated with the magnitude of transfer learning performance deficits among APPswePS1 mice. It is notable that such strong relationships may not be evident at more advanced stages of A pathology in mice or humans. As discussed in Chapters 1 and 2, in addition to utilizing the

PAGE 109

109 n the transfer phase (particularly across many trials) may also reflect non hippocampal dependent new learning of the re configured problems (i.e., even complete hippocampal lesions do not completely abolish performance on this phase of tas k see Chapter 2 ). It would be expected that while transfer learning impairment should be readily detectable at any point after reaching a critical threshold of A pathology, the magnitude of such rendering the task insensitive to further estimation of degree of pathological burden. Synaptic Integrity and Transfer Learning Performance The mechanism whereby either soluble A or fibrillar A plaques is detrimentally influencing transfer learning performance is not known. Previous studies have shown a decline in synaptic integrity, as reflected by a reduction in the expression of the pres ynaptic protein synaptophysin, is a strong correlate of early cognitive deficits in AD 3,24 25,201 Furthermore, synaptic deficits have been extensively reported in several APP transgenic mouse models 29 30,153,202 209 H ippocamp al expression of synaptophysin is present in all neuronal synapses 186 and, as such, it is often used as a reliable general synaptic marker PSD 95 is a major post synaptic scaffolding protein found in excitatory synapses 188 and changes in expression of this protein have also been observed in AD 29,49 In Experiment 2, synaptophysin and PSD 95 expression were exam ined in 3, 6 and 12 month old APPswP1 and 3 and 6 month old Tg swDI mice. Notably, the results from the current experiments indicate that a reduction in synaptophysin accompanied the earliest age at which transfer learning deficits were observed in the APP swePS1 model (6 months). This time point is also the age in which deposition of fibrillar A

PAGE 110

110 plaques is observed. The decrease in expression of synaptophysin suggests that changes in the pre synaptic circuitry are occurring early in the APPswePS1 model and could be mediated by A pathology 205 Although it is important to note that changes in protein levels do not equa te directly to changes in function, or synaptic plasticity, this decrease does suggest impaired integ rity of the synaptic circuitry. As mentioned above, data from mouse models support the findings in APPswePS1. For example, Rutten et al. (2005) have demons trated age related synaptophysin loss in a model of APP that expresses the Swedish and London familial mutations ( KM670/ 671NL) an d the human mutant presenilin 1 (PS1 M146L ; APP751 SL /PS1 M146L mice ) 210 Although these mutations are different than the ones in APPswePS1 mouse model they result in similar early overproduction of A and an inc rease in GFAP levels Rutten and colleagues found that synaptic deficits, observed as a significant decrease in synaptophysin immunoreactive pre synaptic buttons occur as early as 4.5 months, the age at which insoluble A deposits begin to devel op in the APP751 SL /PS1 M146L mice 210 The early transfer learning deficits ob served in Experiment 1 are likely due action of both earlier and intermediary soluble forms of A late fibrillar plaques (see Figure 3 2) 31,198,211 212 A recent study in cerebellar Purkinje cells from another APP/PS1 mouse model ( KM670/671NL and PS1L166P mutations) demonstrated that electrophysio logical measures of synaptic plasticity were affected by the presence of soluble A before the occurrence of fibrillar plaques 198 Together, these results are in accordance to the data acquired in APPswePS1 that show decrease in synaptophysin expression at the time of amyloi d deposition, at a point in which fibrillar A is beginning to deposit, and levels of soluble A are high.

PAGE 111

111 In addition, results from studies from Chapter 4 suggest that changes in expression of PSD 95 are related to late stages in the progression of A pat hology (at 12 months of age in APPswePS1). Previous studies have also reported a decrease in PSD 95 later in the disease progression in both humans and in mouse models of AD 29,49,189 Curiously, som e studies report an increase in PSD 95 expression levels through the course of amyloid pathology deposition 213 214 In those studies, the increase is suggested as a natural compensatory response mechanism to neurotoxicity caused b y A a term used to describe the mechanism explaining the reason why presence of pathology in some individuals is not reflect ed on their measure of cognitive dysfunction 215 216 However, the majority of s tudies with mouse models of amyloid deposition report deficits in post synaptic circuitry and reduction in post synaptic protein expression 217 219 The temporal differences observed in the changes in pre and post synaptic protein levels in the APPswePS1 mice suggest that amyloid pathology may be affecting the pre and post synaptic cir cuitry differently. However, this is not likely, and the differ ence in detection is probably due to differences in overall amounts of these proteins in the hippocampus: synaptophysin is in every synapse, while PSD 95 is found in excitatory synapses only 29 Another possibility is that projecting neurons, to which these pre synaptic circuitries belong to, are being affected by A prior to detriment to the post synaptic hippocampal neurons. Data from human studies have shown that plaque deposition in the entorhinal cortex is an early event in AD 220 221 occurring prior to hippocampal pathology 187,222 The layers II and III of the enthorhinal cortex project directly to the

PAGE 112

112 hippocampal pyramidal cells in the stratum moleculare a nd stratum lacunosum, providing the pre synapt ic input to this circuitry 183,192 Indeed, loss of synaptophysin immunoreactivity in these layers correlates with spatial learning deficits in aged rats 102 Further detailed analysis of plaque deposition in the enthorhinal cortex of APPswePS1 is needed to determine whether this region is indeed affected early in these mice. Furthermore, regional analyses of synaptophysin expression in t he molecular and lacunosum layers of the hippocampus would allow for determination of whether impairment in the enthorhinal cortex, mediated by A toxicity 31,46 47,49,223 could be resulting in the observed pre synaptic changes. There was no evidence for changes in synaptophysin or PSD 95 expression in Tg SwDI mice, despite early and robust amyloid depositi on and transfer learning deficits. These data suggest that the mechanism by which amyloid is detrimentally influencing hippocampal function may be independent of synaptic integrity (at least as assessed by synaptophysin and PSD 95 expression). However, fun ctional measures of synaptic integrity in the Tg SwDI model were not evaluated in the current study. It is notable that GFAP levels were significantly elevated at the earliest ages in which transfer learning performance deficits were observed in Tg SwDI (3 months), where as GFAP expression was not significantly elevated until 12 months of age in APPswePS1mice. GFAP is expressed by astrocytes 224 and increased GFAP expression occurs in a wide variety of conditions including, inflammation, brain injury, normal and pathological aging 175,194 196,225 The Tg swDI model is mostly used as a model of cerebral microvascular deposition, whic h is 180 This type of vascular pathology has been associated with localized neuro inflammatory responses which has been

PAGE 113

113 previously linked to cognit ive impairment 226 For example, Xu et al. (2007) suggest that the early cog nitive deficits observed in the Tg SwDI mice (Barnes maze, 3 months) are mediate by detrimental inflammatory mechanisms resultant from vascular pathology, which are evidenced by presence of activated astrocytes and microglia at the time d eficits are observed 180 As such, it may be that alternative factors associated with these inflammatory responses (e.g., oxidative stress) are mediating synaptic cognitive impairment in Tg SwDI. Synaptic Function in APPswePS1 Mice The reduction in synaptophysin in the APPswePS1 mice at 6 months of age was explored further with hippocampal slice electrophysiology in Experiment 3. Results from recording studies showed that b aseline synaptic transmission was decreased in the APPswePS1 mice compar ed to age matched NTgs. The difference in baseline transmission was not due to loss of fib ers in AP PswePS1 (no differences in fiber volley comparisons between groups) This result supports that the integrity of synaptic circuitry may be compromised, a finding that is consistent with results from Experiment 2 showing the loss of synaptophysin expression at 6 months 227 228 Paired pulse facilitation (PPF), is a type of short term plasticity believed to result from residual calcium increase, which in turn increases the probability of neurotransmitter release after p re synaptic action potential. Notably, PPF was not altered in 6 month APPswePS1 mice, indicating that this short term plasticity mechanism is intact at this age. In contrast, there was a strong trend toward a deficit in a measure long term plasticity long term potentiation (LTP) in APPswePS1 mice. Hippocampal LTP has bee n implicated as the molecular mechanism sub serving hippocampal information processing and memory formation 120 P revious studies hav e reported a significant

PAGE 114

114 decrease in LTP in APPswePS1 mice at 12 months of age 229 230 A tendency towards a decrease in the amplitude of LTP in 6 month old APPswePS1 mice was observed and the prediction (based on previous studies and the re d uction of PSD 95 protein levels at 12 months) would be for LTP to be significant impaired in APPswePS1 at 12 months of age. The experiments in Chapter 4 s how that transfer learning is sensitive to the early different time courses of pathological progression. Furthermore, these deficits were correlated with both measures of soluble and fib rillar (plaques) at early ages SwDI and APPswePD1) and electrophysiological data (APPswePS1 only) demonstrated that transfer learning may be a marker for early hippocampal dysfunc tion independe nt of the mechanisms mediating the deficits.

PAGE 115

115 Figure 4 1. Recording of field excitatory postsynaptic potentials (fEPSP) at hippocampal CA3 CA1 synapses in brain slices. (A) Electrical stimuli are delivered via a bipolar stimulation electr ode from the pre synaptic fibers of extracellular recording electrode. (B) EPSP waveforms before and after LTP.

PAGE 116

116 Figure 4 2 Transfer learning performance in 3 and 6 month old APPswePS1 and Tg SwDI mice. (A) Performance on the transfer phase (percent error) in APPswePS1 (blue) and B6C3/J mice (red). (B) Performance on the transfer phase (percent error) in Tg SwDI (green) and C57BL/6J (pink) mice. Note that deficits were evident in Tg SwD I mice by 3 months of age but not in APPswePS1 mice until 6 months of age. In (C), percent increase in transfer learning performance compared directly across models indicates performance in the Tg SwDI mice is significantl y worse than in APPswePS1 mice. Se e text for statistical analysis.

PAGE 117

117 Figure 4 3 SwDI than in APPswePS1 hippocampus. (A) So (mean SE). A strong trend towards SwDI relative to APPswePS1 mice was eviden t at 3 month of age. This difference was magnified by 6 months of age, at which SwDI mice (green bar) were more than 7 fold that of APPswePS1 mice (blue bar). Right panels show visible amyloid deposits, stained with Thioflavin S, in co ronal hippocampal slices taken from representative (B) APPswePS1 and (C) Tg SwDI mice at 6 months of age. In SwDI, more robust plaque deposition was evident in Tg SwDI compared to APPswePS1 mice. 500 mm

PAGE 118

118 Figure 4 4 Line ar regression plots showing relationship between A 42 levels and transfer learning performance in APPswePS1. (A) Linear regression plot showed a positive correlation between A 42 levels and plaque (Thioflavin S positive) counts in the hippocampus of APPsw ePS1 mice. (B) Transfer learning performance expressed as a percentage decline from the mean performance of age matched NTgs) is significantly correlated to A 42 levels in the hippocampus and in (C) transfer learning is significantly correlated to plaque numbers in the hippocampus. See text for statistics.

PAGE 119

119 Figure 4 5 Synaptic protein expression in APPswePS1 mice. (A) Representative immunoreactive bands of synaptophysin, PSD 95, GFAP and GAPDH observed in the hippocampus of 3, 6 and 12 month old APPsw ePS1 mice. (B) Synaptophysin immunoreactivity (mean S.E.) was significantly reduced in both 6 and 12 month old mice compared to 3 month old mice. (C) PSD95 immunoreactivity (mean S.E.) was decreased in 12 month old mice relative to 3 month old mice. (D ) GFAP immunoreactivity (mean S.E.) was significantly increased in 12 months mice relative to 3 and 6 month old mice. See text for statistics.

PAGE 120

120 Figure 4 6 GFAP expression in Tg SwDI mice. (A) Representative immunoreactive bands of GFAP for 3 and 6 m onth old Tg SwDI mice (B) GFAP expression (mean S.E.) was significantly elevated in 6 month compared to 3 month old Tg SwDI mice See text for statistics.

PAGE 121

121 Figure 4 7 Baseline transmission and paired pulse facilitation. (A) Input output curves for APPswePS1 (blue circles) and age matched NTgs (red circles). E ach point represents the mean the SEM IV curve is significantly decreased in APPswePS1 mice, demonstrating a impairment in synaptic transmission. (B) There was no difference in mean fiber pot ential by voltage for APPswePS1 (blue bar) and NTg (red bars), demonstrating no difference in number of stimulated axonal fibers. (C) There was no difference in mean pair pulse facilitation (PPF) ratio between APPswe PS1 and controls, suggesting that short term plasticity was not affected in APPswePS1 mice. See text for statics.

PAGE 122

122 Figure 4 8 Long term potentiation in APPswePS1 (blue circles) and age matched NTgs (red circles). Each point represents the mean the SEM. (A) Time course changes in the slope of EPSPs for the NTgs and APPswePS1 mice pathways before and after delivery of high frequency stimulation (100 Hz, 100 pulses, for 1 sec, repeated 4 times with 10 sec apart.) (B) A strong tendency for decrease in LTP (LTP path) in APPswePS1 mice (blue bar ) compared to NTg (red bar) was observed. (C) Bar diagram showing the average magnitude of LTP during the last 5 min of recording for NTg and APPswePS1 mice demonstrated a significant effect showing reduction in LTP induction. See text for statistics.

PAGE 123

123 Ta ble 4 1. Table summarizes APPswePS1 and Tg SwDI mouse models Transge n ic Line Mutation A deposits Hippocampal Deficits Synaptic Deficits Inflammation markers Olfactory changes APPswePS1 Swedish/ Human PS1 6 mon 7 mon (modified Barnes maze) 12 mon (LTP, sp ine a lterations) 14 mon microgliosis 231 ; 16 mon increase in GFAP 232 Not reported Tg SwDI Sw edish/ Dutch/ Iowa 3 mon 3 mon (Barnes maze) Not reported 3 mon 233 Not reported

PAGE 124

124 CHAPTER 5 CONCLUSION The main objective of this dissert ation was to develop a mouse assessment of transfer learning which has been shown to be predictive of cognitive decline in non demented elderly humans 1 Such a rodent task, if truly analogous to the human task, would bridge the ga p between rodent and human studies, facilitating comparisons of results acquired from mechanistic preclinical studies to clinical diagnosis and evaluation of potential therapies. Furthermore, assessment of transfer learning in transgenic mouse models would pathology in relation to early changes in cognitive function. Specifically, the goals were to 1) to develop a mouse analog of the human task and to determine if the hippocampus is critical for transfer learning in the mouse version of the task, 2) to test the utility of the model of AD; and 3) to investigate the role of early AD synaptic dys function) in mediating transfer learning deficits in two mouse models of AD. Implications of Chapter 2 and Future Studies In Chapter 2, a mouse transfer task highly analogous to a human assessment of transfer learning, was developed for mice. Findings fro m experiments in this chapter validated the design of the mouse transfer task by showing that mice can readily learn a series of compound discrimination problems within a single session and then transfer this information when irrelevant stimuli were altere d. Furthermore, bilateral hippocampal lesions in mice significantly impaired transfer learning while performance in the initial associative learning phase was unaffected. These findings support the utility of the transfer task as an assessment of hippocamp al function in mice as it allows for

PAGE 125

125 dissociative characterizing of hippocampal dependent behavior (transfer learning) from simple stimuli association (learning phase). One important advantage the mouse transfer task offers over other hippocampal dependent tasks is that, while mild to moderate motor impairment affects performance and behavioral measures in the water maze, moderate motor impairments should not prevent an animal from performing the transfer task, as time to complete the task is not integral t o performance measures (% accuracy is used). Future studies should evaluate other neuro anatomical substrates required for successful performance on the transfer learning task. Certainly, brain regions beyond hippocampus (e.g., prefrontal cortex) may also be important for accurate transfer learning. With respect to prefrontal cortex, for example, it would be expected that age/pathology associated changes in this region might largely influence the initial learning phase of the task. Notably, deficits related to prefrontal cortical pathology have also been detected in 6 month old APPswePS1 67,177,234 and prefrontal cortical lesions have been shown to affect perceptual attention in rats, which may be important for the e ncoding of initial stimuli during acquisition of associations 235 237 Analysis of performance of pre frontal cortex lesioned mice in the transfer task would allow determination of whether this region is critical t o the learning phase as predicted 83 80 This data would also confirm the dissociation o f cognitive constructs involved in the flexible processing of information n ecessary for transfer learning. Implications of Chapter 3 and Future Studies In Chapter 3, transfer learning was assessed in a transgenic mouse model of amyloid pathology (APPswePS1 mouse model of AD). Deficits in transfer learning were unaffected by prior experience on the task and were observed in concurrence with the

PAGE 126

126 deficits were not observe d prior to this plaque deposition (3 months). Furthermore, transfer learning was more sensitive than a standard hippocampal dependent test (water maze), as the 12 month old APPswePS1 mice were not impaired. Findings from Chapter 3 demonstrate that transfer learning may be a useful detection method of assessment of hippocampal dysfunction in mouse models of AD, perhaps more so than other standard hippocampal tests for mice (e.g., water maze). Future studies should also include the transfer learning assessmen t of additional AD transgenic models (e.g., tau mouse model) and mouse models of other neurodegenerative diseases that may benefit from hippocampal assessment. The ability to use the transfer learning task in within subject studies suggests that this asses sment is ideal for the evaluat ion pharmacological treatments. Implications of Chapter 4 and Future Studies Studies from Chapter 4 began to delve into the mechanisms that may underlie the hippocampal defi amyloid pathology (Tg SwDI) was assessed and compared to performance of APPswePS1 mice. Both models plaque deposition (6 months in APPswePS1 and 3 months in Tg SwDI). Findings from Chapter 4 provide additional support for the idea that transfer learning is sensitive to transfer task can be used as an assessment of hippocampal function independent of the mechanisms underlying deficits. Transfer learning impairment was observed in 6 month old APPswePS1 and 3 month old Tg SwDI, ages at which plaque begins to develop in each model. Furthermore, the transfer

PAGE 127

127 learning deficits observed in APPswePS1 appeared to be progressive, as no deficits were present at 3 months but were observed at 6 months of age. However, a ceiling effect on how much impairment can be detected was noted. For example, both 6 and 12 month old APPSwePS1 mice had the same magnitude of percent error in t he transfer phase of the task (s ee Chapter 3 Figure 3 4B and 4 1C). Comparison of performance between 3 and 6 month old Tg SwDI demonstrated the same ceiling ef fect; as again, the magnitude of error was not different between ages (when comparing % increase in error relative to age matched NTg). This ceiling effect may be addressed by changing the measure of transfer learning performance during the transfer phase. For example, errors and trials to criterion or a smaller number of total trials, which may be more sensitive to the earlier portion of transfer task when mice are adapting to the transfer phase from the learning phase, could be used. Further discussion of results and implications of APPswePS1 and Tg SwDI transfer learning assessment and pathology will be discussed separately below. APPswePS1: I mplicatio ns and Future Studies observed prior to the development of the pathol ogy. Moreover, the APPswePS1 model does not show overt neuronal loss, thus, synaptic dysfunction was a likely candidate as the me diator of transfer learning deficits. The lack of ne uronal loss is often criticized as a weakness of this model However, although it is still unknown if the neuronal loss in AD is directly caused by A accumulation, i f this is the case the hypothesized t ime frame in which this process wo uld take place is 2 to 3 years the typical lifetime of a mouse 238 Thus, the synaptic perturbations that may be mediated by A would be characteristic of early stages of pathological progression, and the APPswePS1 synaptic

PAGE 128

128 deficits would be reflecting primary synaptotoxicity (early) rather than secondary consequences of neuronal degeneration (late) An alysis of synaptic function and synaptic integrity suggested that pre synaptic mechanisms may be affected prior to the observance of changes related to post synaptic mechanisms. This early synaptic dysfunction is most likely mediated by toxicity resulting from the accumulation of mu the synapses is a research topic of extreme importance as the understanding of this In mice, correlations o f cognition, synaptic function and smaller forms of amyloid (such as 46 47,111,211,239 240 not a good correlate of cognitive dysfunction, accumulat precedes synaptic changes and has been shown to predict cognitive impairment 142,207 amyloid beta derived diffusable ligands, are a 142 Previous studies suggest that the hippocampal region is particularly vulnerable to the detrimental 169,212,241 242 For example, t here is evidence that A can bind to pre and postsynaptic elements, interacting functionally and structurally with several membrane bound receptors such as 7 nicotinic acetylcholine, NMDA, AMPA, insulin, RAGE within many other receptors 43,49,243 F urthermore in humans and APP transgenic en shown to disrupt long term potentiation 30 31,49,211,238, 244 245 a nd to produce long lasting decrease of baseline synaptic strength 120 Oligomers have also been shown to r educ e synaptic activation 246 248 and to pr omote long term depression (LTD). Shankar et al. (2007) suggests that A oligomers may

PAGE 129

129 dependent signaling cascades toward pathways involved in the induction of LTD 169 Another way in which A could enhance LTD is by blocking glutamate uptake at the synapses, resulting in desensitization of the receptors, and ultimat ely synaptic depression 241 LTD, in turn, could med iate reduction of post synaptic densities late in 241 (e.g., reduce PSD 95 expression), as previous studies have shown that LTD results in progr essive loss of den dritic spines Thus, in accordance with literature and previous studies, the results from Experiment 2 in APPswePS1 mice support that pre synaptic dysfunction occurs early ultimately resulting in synaptic failure which could contribute to the loss of hipp ocampal function and transfer learning impairment in the absence of frank neuronal loss 139 It is also important to note that in this mod el, onl y pre synaptic alterations are required for observance of transfer learning impairment, which suggests that transfer learning is indeed an early predictor of further pathological related dysfunction. Future studies would include the analyses of synaptic de nsity by stricter methods such as counting of synaptic puncta and analysis of integrity of post synaptic structure. Also, further pus at the time of transfer learning impairment. Tg SwDI: Implications and Future Studies Additional findings from Chapter 4 with Tg SwDI mice confirm that transfer learning impairment can be used as an assessment of early amyloid pathology deposition in the hippocampus independent of the mechanism underlying the deficits. Surprisingly, in lieu of findings in APPswePS1, the loss of synaptic proteins was not observed in the Tg SwDI mice at either 3 or 6 months of age. Overt neuronal loss has

PAGE 130

130 not been shown in this model, and further, because synaptic function was not evaluated in Chapter 4 it is not possible to conclude that synaptic impairment is not occurring at the age transfer learning deficits were observed. Also, the Tg SwDI mice have not been as exte nsively characterized as the APPswePS1 m ice so limited amount of data is available for comparisons. However, these findings, indicate that the difference in genetic mutations between APPswePS1 and Tg SwDI possibly even due to the influence of the backgrou nd strain in which the mice were developed, can result in a difference in the mechanisms by which amyloid deposition affects the neura l circuitry in the hippocampus. The Tg is present not only in AD (about 80% cases), but other diseases, and it can be defined as a condition on its own: cerebral amyloid angiopathy 171 It is hypothesized that inflammatory resp onses and mechanisms are at the center of vascular amyloidosis 171 172,249 However, it is stil l not known if inflammation is a cause, contributor, or secondary phenomenon to A accumulation. I t is not sur prising that GFAP levels were elevated as early as 3 months of age in this model (compared t o 12 months in APPswePS1 mice), as early increase has also been observed in previous studies with increases of GFAP at 3 months of age 250 GFAP, glial fibrillary acid protein, is a marker of astrocyte reactivity and is expressed by numerous cell types but is parti cularly used as a marker for astrocytes, possibly mediating neuron .astrocyte interactions 224 Within its many suggested functions, astrocytes are believed to play a role in the repair and scarring process of the brain following traumatic injuries, secretion and recycling of neurotransmitters, synaptic remodeling, and modulation of oxidative stress 251 As such,

PAGE 131

131 changes in astrocytes (e.g., quantity, activation states) can result in extremely deleterious consequences for the nervous system 252 253 High levels of GFAP have been reported in AD 194 196,254 .Reactive astrocytes are usually observed around plaques, encircling A deposits in a manner similar t o glial sca rring 253,255 Astrocytes express receptors that bind A such as RAGE, membrane associated proteoglycans, within many others, which may account for the accumulation of asctrocytes near A aggregation 256 Furthermore there is evidence that A can directly affect astrocyte activity. Chow et al. (2010) showed that A exposure disrupt s astrocyte calcium homeostasis, resulting in increases in GFAP in astrocyte neuron cocultures 257 Also, exposure of astrocyte cultures to A result in increases in IL 1 TNF iNOS and NO production 258 Microglia activation and changes could be a secondary mechanism by which increased astrocyte reactivity may be mediate or appear co ncurrently to. P revious associated astrocytes upregulate the expression of pro inflammatory cytokines acting as chemoattractants for microglia 172,259 260 Currently, significant research is focus plaques are believed to be the site of local inflammatory responses 226,260 Accumulating evidence indicates that neurotoxic glial activation may be involved in the early neurodegenerative process. P ost mortem analysi s of AD brains reveal microglia clustered around plaques combined with high levels of oxidative stress and neuroinflammation 260 M icroglial NADPH oxidase, for example, has been implicated in the progressive nature of AD through the production of reactive oxygen species (ROS) 230 Together, mechanisms of oxidative stress and inflammation could

PAGE 132

132 influence LTP and plasticity by affecting the calcium equilib rium at the synapse 261 resulting in synaptic dysfunction 172 and ultimately loss of hippocampal function 102 As such, at the time transfer learning deficits are observed, the presence of various forms of A could be resulting in the increase in levels of GFAP observed, indicating the presence of inflammatory response s that could involve microglial activation and oxidative stress. I t is important to note, that these initial assessments of synaptic integrity from experiments in Chapter 4 must be interpreted lightly, as the number of mice in each cohort was very small. Further studies are needed to det ermine the exact mechanisms that are mediating transfer learning impairment in the Tg SwDI mouse model. Additional s tudies of synaptic function, inflammatory responses and markers of oxidative stress in both models would allow for better comparison between the two hypothesized distinct Together, findings from APPswePS1 and Tg SwDI studies suggest that transfer learning is sensitive to ear accumulation, and it is an excellent tool for analyses of progressive hippocampal dysfunction in mice. Furthermore, results from Tg SwDI indicate that early amyloid deposition appears to mediate transfer learning deficits in the absence of changes in synaptic protein levels. Final W ords The studies in this dissertation demonstrate that transfer learning should be an ideal assessment of hippocampal function in mouse models of neurodegenerative disease. The high level of analogy between the m ouse and human test paradigms provides a higher comparative power between results from mechanistic studies of neurobiology in mouse models to humans who suffer from the disease. Finally, the

PAGE 133

133 ability for within subject testing suggests that this task may b e an excellent tool for conducting longitudinal preclinical studies.

PAGE 134

134 LIST OF REFERENCES 1 Myers, C. E., Kluger, A., Golomb, J., Ferris, S.H., de Leon, M.J., Schnirman, G., Gluck, M.A. Hippocampal Atrophy Disrupts Transfer Generaliz a tion in Nondemented Elderly. J. Geriatr. Psychiatry Neurol. 15 82 90 (2002 ). 2 Association, A. 2012 Alzheimer's disease facts and figures. 131 168. 3 Masliah, E., Mallory M, Hansen L, DeTeresa R, Terry RD. Quantitative synaptic alterations in the human neocortex during normal aging. Neurology 43 192 197 (1993). 4 Francis, P. T., Palmer, A.M., Sims, N.R., Bowen, D.M., Davison, A.N., Esiri, M.M., Neary, D., Snowden, J.S., Wilcock, G.K. Neurochemical studies of early nfluence on treatment. N Engl J Med 313 7 11 (1985). 5 Mouton, P. R., Martin, L. J., Calhoun, M. E., Dal Forno, G. & Price, D. L. Cognitive decline strongly correlates with cortical atrophy in Alzheimer's dementia. Neurobiology of Aging 19 371 377 (1998 ). 6 Banchera, C. et al. formation of neurofibrillary tangles in Alzheimer's disease Brain Res. 477 90 99 (1989). 7 Ferri, C. P. et al. Global prevalence of dementia: a Delphi consensus study. The Lancet 366 2112 2117 (2005). 8 Alagiakrishnan, K., Gill, S. S. & Fagarasanu, A. Genetics and epigenetics of Alzheimer's disease. Postgrad Med J. (2012). 9 Rabin, L. A. et al. Predicting Alzheimer's disease: neuropsychological tests, self reports, and i nformant reports of cognitive difficulties. J Am Geriatr Soc 60 1128 1134 (2012). 10 Gluck, M. A., Myers, C.E., Nicolle, M.M., Johnson, S. Computational models of the hippocampal region: implications for prediction of risk for Alzheimer's disease in non demented elderly. Curr Alzheimer Res 3 247 257 (2006). 11 Myers, C. E., Kluger, A., Golomb, J., Gluck, M. A. & Ferris, S. Learning and Generalization Tasks Predict Short Term Cognitive Outcome in Nondemented Elderly. J Ger Psych. Neurol. 21 93 104 (2008 ). 12 Johnson, S. C., Schmitz, T.W., Asthana, S., Gluck, M.A., Myers, C.E. Associative Learning Over Trials Activates the Hippocampus in Healthy Elderly but not Mild Cognitive Impairment. Aging, Neuropsychology, and Cognition 15 129 145 (2008).

PAGE 135

135 13 De Leo n, M. J., George, A.E., Stylopoulod, L.A., Smith, G., Miller, D.C. Early Marker for Alzheimer's DIsease: The atrophic Hippocampus. The Lancet (1989). 14 de Leon, M. J., Golomb, J., George, A.E., Convit, A., Tarshish, C.Y., McRae, T., De Santi, S., Smith, G., Ferris, S.H., Noz, M. The radiologic prediction of Alzheimer disease: the atrophic hippocampal formation. American journal of neuroradiology 14 897 906 (1993). 15 Mark, J. W., Claudia, H. K., Walter, F. S., Gay, L. R. & Juan, C. T. Hippocampal neuron s in pre Neurobiology of aging 25 1205 1212 (2004). 16 Guerrero Berroa, E. et al. The MMSE orientation for time domain is a strong predictor of subsequent cognitive decline in the elderly. Int J Geriatr Psychiatry 24 1429 1437 (2009). 17 Belleville, S., Sylvain Roy, S., de Boysson, C. & Mnard, M. C. Characterizing the memory changes in persons with mild cognitive impairment. Prog Brain Res 169 365 375 (2008). 18 de Paula, J. J. et al. The Tower of London Test: differen t scoring criteria for diagnosing Alzheimer's disease and mild cognitive impairment. Psychol Rep 110 477 488 (2012). 19 Kurt, P., Yener, G. & Oguz, M. Impaired digit span can predict further cognitive decline in older people with subjective memory compla int: a preliminary result. Aging Ment Health 15 364 369 (2011). 20 Silva, D. et al. Comparison of Four Verbal Memory Tests for the Diagnosis and Predictive Value of Mild Cognitive Impairment. Dement Geriatr Cogn Dis Extra 2 120 131 (2012). 21 Blacker, D. et al. Neuropsychological measures in normal individuals that predict subsequent cognitive decline. Arch Neurol 64 862 871 (2007). 22 Maruff, P. et al. Subtle memory decline over 12 months in mild cognitive impairment. Dement Geriatr Cogn Disord 18 3 42 348 (2004). 23 Wiechmann, A., Hall, J. R. & O'Bryant, S. E. The utility of the spatial span in a clinical geriatric population. Neuropsychol Dev Cogn B Aging Neuropsychol Cogn 18 56 63 (2011). 24 Jack, C., Vemuri, P. & Wiste, H. J. EVidence for order ing of alzheimer disease biomarkers. Archives of Neurology 68 1526 1535 (2011).

PAGE 136

136 25 Jack, C. R. et al. Introduction to the recommendations from the National Institute on Aging Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's dis ease. Alzheimer's & dementia : the journal of the Alzheimer's Association 7 257 262 (2011). 26 Knopman, D. S. et al. Short term clinical outcomes for stages of NIA AA preclinical Alzheimer disease. Neurology 78 1576 1582 (2012). 27 Mattsson, N. et. al. CSF Biomarkers and Incipient Alzheimer Disease in Patients With Mild Cognitive Impairment. JAMA 302 385 393 (2009). 28 Sjgren, M., Andreasen, N. & Blennow, K. Advances in the detection of Alzheimer's disease use of cerebrospinal fluid biomarkers. Clini ca Chimica Acta 332 1 10 (2003). 29 Almeida, C. G., Tampellini, D., Takahashi, R. H., Greengard, P., Lin, M. T., Snyder, E. M., Gouras, G. K. Beta amyloid accumulation in APP mutant neurons reduces PSD 95 and GluR1 in synapses. Neurobiol. Dis. 20 187 19 8 (2005). 30 Breyhan, H., Wirths, O., Duan, K., Marcello, A., Rettig, J., Bayer, T.A. APP/PS1KI bigenic mice develop early synaptic deficits and hippocampus atrophy. Acta Neuropathol. 117 677 685 (2009). 31 Cerpa W, D. M., Inestrosa NC. Structure functi on implications in Alzheimer's disease: effect of Abeta oligomers at central synapses. Curr. Alzheimer Res. 5 233 243 (2008). 32 DeKosky, S. T., Scheff, S.W. Synapse loss in frontal cortex biopsies in erity. Ann. Neurol. 27 457 464 (1990). 33 Dickson, D. W. et al. Correlations of synaptic and pathological markers with cognition of the elderly. Neurobiology of Aging 16 285 298. 34 Dickson, D. W. et al. Correlations of synaptic and pathological marker s with cognition of the elderly. Neurobiol. Aging 16 285 298 (1995). 36 Jack, C. r., Vemuri, P., Wiste, H. J. & et al. Shapes of the trajectories of 5 major biomarkers of alzheimer disease. Archives of Neurology 1 12 (2012). 37 Sperling, R. A. et al. T oward defining the preclinical stages of Alzheimer's disease: recommendations from the National Institute on Aging Alzheimer's Association workgroups on diagnostic guidelines for Alzheimer's disease. Alzheimers Dement. 7 280 292 (2011).

PAGE 137

137 38 Robert, D. T. et al. Physical basis of cognitive alterations in alzheimer's disease: Synapse loss is the major correlate of cognitive impairment. Annals of Neurology 30 572 580 (1991). 39 Terry, R. D., Masliah, E., Salmon, D.P., Butters, N., DeTeresa, R., Hill, R., Ha disease: synapse loss is the major correlate of cognitive impairment. Ann. Neurol. 30 572 580 (1991). 40 Sze, C. I., Bi, H., Kleinschmidt DeMasters, B. K., Filley, C. M. & Mar tin, L. J. disease brains. Journal of the Neurological Sciences 175 81 90 (2000). 41 Kamenetz, F., Tomita, T., Hsieh, H., Seabrook, G., Borchelt, D., Iwatsubo, T., Sisodia, S., Malinow, R. APP processing and synaptic function. Neuron 37 925 937 (2003). 42 Roselli, F., Tirard, M., Lu, J., Hutzler, P., Lamberti, P., Livrea, P., Morabito, M., and Almeida, O. F. Soluble beta amyloid1 40 induces NMDA dependent degradation of po stsynaptic density 95 at glutamatergic synapses. J. Neurosci. 25 11061 11070 (2005). 43 Lacor, P. N., Buniel, M. C., Furlow, P. W., Sanz Clemente, A., Velasco, P. T., Wood, M., Viola, K. L., and Klein, W. L. Abeta oligomer induced aberrations in synapse composition, shape, and density provide a molecular basis for loss of J. Neurosci. 27 796 807 (2007). 44 Sheline, Y. I. et al. APOE4 allele disrupts resting state fMRI connectivity in the absence of amyloid plaques or decreased CSF Abeta42. J. Neurosc. 30 17035 17040 (2010). 45 Sperling, R. A., Jack, C. R. & Aisen, P. S. Testing the Right Target and Right Drug at the Right Stage. Science Translational Medicine 3 111 133 (2011). 46 Ferreira, S. T., Marcelo, N. N. V. & Fernanda, G. D. F. Soluble protein oligomers as emerging toxins in alzheimer's and other amyloid diseases. IUBMB Life 59 332 345 (2007). 47 Glabe, C. G. Common mechanisms of amyloid oligomer pathogenesis in degenerative disease. Neurobiology of Aging 27 570 575 (2006). 48 Gruden, M. A., Davidova, T.B., Malisauskas, M., Sewell, R.D., Voskresenskaya, N.I., Wilhelm, K., Elistratova, E.I., Sherstnev, V.V., Morozova Roche, L.A. Differential neuroimmune markers to the onset of Alzheimer's disease neurodege neration and dementia: autoantibodies to Abeta((25 35)) oligomers, S100b and neurotransmitters. J Neuroimmunol. 186 181 192 (2007).

PAGE 138

138 49 Koffie, R. M. et al. Oligomeric amyloid beta associates with postsynaptic densities and correlates with excitatory synap se loss near senile plaques. Proceedings of the National Academy of Sciences 106 4012 4017, (2009). 50 42) globulomer) suppress spontaneous synaptic activity by inhibition of P/Q type calcium currents. J Neurosci. 28 788 797 (2008). 51 Oddo, S., Caccamo, A., Tran, L. & al., e. Temporal profile of amyloid beta (Abeta) oligomerization in an in vivo model of Alzheimer disease. A link between Abeta and tau pathology. J Biol Chem 281 1599 1604 (2006). 52 qualitative and quantitative analysis. Dementia 1 (1990). 53 Terry, R. D., Masliah, E., Salmon, D. P., Butters, N., DeTeresa, R., Hill, R., Hansen, L. A., Katzman, R. Physical basis disease: synapse loss is the major correlate of cognitive impairment. Ann. Neurol. 30 572 580 (1991). 54 Portet, F. et al. MCI Working Group of the European Consortium on Alzheimer's Disease (EADC) Mild cognitive impairment (MCI) in medical practice: a critical review of the concept and new diagnostic procedure. Report of the MCI Working Group of the European Consortium on Alzheimer's Disease. J Neurol Neurosurg Psychiatry 77 714 718 (2006). 55 Dubois, B. et al. Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS ADRDA criteria. Lancet Neurol 6 734 746 (2007). 56 Lowndes, G. J., Saling, M.M., Ames, D., Chiu, E., Gonzalez, L.M., Savage, G.R. Recall and recognition of verbal paired asso ciates in early Alzheimer's disease. Journal of the International Neuropsychological Society 14 591 600 (2008). 57 Bizon, J. L. & Nicolle, M. M. Rat models of age related cognitive decline. In Handbook of Models for Human Aging (Academic Press, 2006). 58 Kandel, E. R., Kupfermann, I. & Iversen, S. Principles of Neural Science Fourth Edition edn, 1227 1245 (McGraw Hill, Health Professions Division, 1991). 59 Kluger, A., Ferris, S.H., Golomb, J., Mittelman, M.S., Reisberg, B. Neuropsychological Pred iction of Decline to Dementia in Nondemented Elderly. J. Geriatr. Psychiatry Neurol. 12 (1999). 60 Henke, K. A model for memory systems based on processing modes rather than consciousness. Nat Rev Neurosci 11 523 532 (2010).

PAGE 139

139 61 Eric R. Kandel, James H. Schwartz & Jessell, T. M. Principles of Neural Science 1151 1163 (McGraw Hill, Health Professions Division). 62 Gluck, M. A., Myers, C.E. Hippocampal mediation of stimulus representation: A computational theory. Hippocampus 3 491 516 (1993). 63 Gluc k, M. A., Myers, C.E. Representation and association in memory: A neurocomputational view of hippocampal function. Curr. Dir. Psychol. Sci. 4 23 29 (1995). 64 Gluck, M. A., Myers, C.E. Gateway to Memory: An Introduction to Neural Network Models of the Hippocampus and Learning. Cambridge, MA: MIT Press. (2001). 65 Gluck, M. A., Guest Editor Hippocampal Computation and Memory (Special issue of Hippocampus). 6 (1996). 66 Myers, C. E., Gluck, M.A. Cortico Hippocampal Representations in Simultaneous Odo r Discrimination: A Computational Interpretation of Eichenbaum, Mathews, and Cohen (1989). Behavioral Neuroscience 110 685 706 (1996). 67 Montgomery, K. S. et al. Novel age dependent learning deficits in a mouse model of Alzheimer's disease: Implications for translational research. Neurobiology of Aging (2011). 68 Barense, M. D., Bussey, T.J., Lee, A.C., Rogers, T.T., Davies, R.R., Saksida, L.M., Murray, E.A., Graham, K.S. Functional specialization in the human medial temporal lobe. J. Neurosci. 25 1023 9 10246 (2005). 69 Quamme, J. R., Yonelinas, A.P., Norman, K.A. Effect of unitization on associative recognition in amnesia. Hippocampus. Hippocampus 17 192 200 (2007). 70 Eichenbaum, H., Mathews, P. Further Studies of Hippocampal Representation Durin g Odor Discrimination Learning. Behavioral Neuroscience 103 1207 1216 (1989). 71 Wilson, C. R., Charles, D.P., Buckley, M.J., Gaffan, D. Fornix Transection Impairs Learning of Randomly Changing Object Discriminations. The Journal of Neuroscience 27 1286 8 12873, doi:10.1523/JNEUROSCI.3536 07.2007 (2007). 72 Sutherland, R., Rudy, J. Configural association theory: The role of the hippocampal formation in learning, memory and amnesia. Psychobiology 17 129 144 (1989).

PAGE 140

140 73 Rudy, J., Sutherland, R. Configura l association theory and the hippocampal formation: An appraisal and reconfiguration. Hippocampus 5 375 398 (1995). 74 Rudy, J. W. & O'Reilly, R. C. Conjunctive representations, the hippocampus, and contextual fear conditioning. Cogn Affect Behav Neuros ci. 1 66 82 (2001). 75 Moses, S. N., Ryan, J.D. A comparison and evaluation of the predictions of relational and conjunctive accounts of hippocampal function. Hippocampus 16 43 65 (2006). 76 Ikonen, S., McMahan, R., Gallagher, M., Eichenbaum, H., Tanil a, H. Cholinergic system regulation of spatial representation by the hippocampus. Hippocampus 12 386 397 (2002). 77 Bunsey, M., Eichenbaum, H. Conservation of hippocampal memory function in rats and humans. Nature 379 255 257 (1996). 78 Eichenbaum, H., Fagan, A., Mathews, P. & Cohen, N. J. Hippocampal System Dysfunction and Odor Discrimination Learning in Rats: Impairment or Facilitation Depending on Representational Demands. Behav Neurosci 102 331 339 (1988). 79 Rudy, J. W. & Sutherland, R. J. Config ural association theory and the hippocampal formation: An appraisal and reconfiguration. Hippocampus 5 375 389, doi:10.1002/hipo.450050502 (1995). 80 Myers, C., Shohamy, D., Gluck, M., Grossman, S., Kluger, A., Ferris, S., Golomb, J., Schnirman, G., Schw artz, R. Dissociating hippocampal versus basal ganglia contributions to learning and transfer. Journal of Cognitive Neuroscience 15 185 193 (2003). 81 Myers, C. E. et al. Learning and generalization deficits in patients with memory impairments due to anterior communicating artery aneurysm rupture or hypoxic brain injury. Neuropsychology 22 681 686 (2008). 82 Eichenbaum, H., Yonelinas, A.R., Ranganath, C. The Medial Temporal Lobe and Recognition Memory. Annu Rev Neurosci. 30 123 152 (2007). 83 Myer s, C. E., Gluck, M. A., Granger R. Dissociation of hippocampal and entorhinal function in associative Learning: A computational approach. Psychobiology 23 116 138 (1995). 84 Pascalis, O., Hunkin, N.M., Bachevalier, J., Mayes, A.R. Change in backgroun d context disrupts performance on visual paired comparison following hippocampal damage. Neuropsychologia 47 2107 2113 (2009).

PAGE 141

141 85 Philipson, O. et al. Animal models of amyloid Beta related pathologies in Alzheimer's disease. FEBS J. 277 1389 1409 (2010) 86 Quon, D. et al. Formation of beta amyloid protein deposits in brains of transgenic mice. Nature 352 239 241 (1991). 87 Yamaguchi, F. et al. Transgenic mice for the amyloid precursor protein 695 isoform have impaired spatial memory. Neuroreport 2 7 81 784 (1991). 88 Wirak, D. O. et al. Deposits of amyloid beta protein in the central nervous system of transgenic mice. Science 19 323 325 (1991). 89 Sandhu, F. A., Salim, M. & Zain, S. B. Expression of the human beta amyloid protein of Alzheimer's dis ease specifically in the brains of transgenic mice. J Biol Chem 266 21331 21334 (1991). 90 Morrissette, D. A., APArachikova, A., Green, K. N. & Laferla, F. M. Relevance of transgenic mouse models to human Alzheimer's disease. J Biol Chem 284 6033 6037 ( 2009). 91 Epis, R. et al. Searching for new animal models of Alzheimer's disease. Eur J Pharm 626 57 63 (2010). 92 Kirwan, C. B., Gilbert, P.E., Kesner, R.P. The role of the hippocampus in the retrieval of a spatial location. Neurobiology of Learning an d Memory 83 65 71 (2005). 93 Addis, D. R., Moscovitch, M. & McAndrews, M. P. Consequences of hippocampal damage across the autobiographical memory network in left temporal lobe epilepsy. Brain 130 2327 2342 (2007). 94 Morris, R. G., Garrud, P., Rawlins J.N., O'Keefe, J. Place navigation impaired in rats with hippocampal lesions. Nature 297 681 683 (1982). 95 Bizon, J. L., LaSarge, C.L., Montgomery, K.S., McDermott, A.N., Setlow, B., Griffith, W.H. Spatial reference and working memory across the lifes pan of male Fischer 344 rats. Neurobiology of Aging 30 646 655 (2009). 96 Candi, L. L. et al. Deficits across multiple cognitive domains in a subset of aged Fischer 344 rats. Neurobiology of aging 28 928 936 (2007). 97 Gallagher, M., Bizon J.L.,Hoyt E.C.,Helm, K.A., Lund, P.K. Effects of aging on the hippocampal formation in a naturally occurring animal model of mild cognitive impairment. Experimental Gerontology 38 71 77 (2003).

PAGE 142

142 98 LaSarge, C. L., Montgomery, K.S., Tucker, C., Slaton, S., Griffit h, W.H., Setlow, B., Bizon, J.L. Deficits across multiple cognitive domains in a subset of aged Fischer 344 rats. Neurobiology of Aging 28 (2007). 99 LaSarge CL, N. M. C. o. d. c. r. m. o. h. a. I. A. M. o. H. C. A. B., J. L., & A.G. Woods, ed), pp 73 102 New York, NY: Humana Press. Comparison of different cognitive rat models of human aging. In: Animal Models of Human Cognitive Aging. 73 102 (2009). 100 Norris, C. M., Halpain, S. & Foster, T. C. Reversal of Age Related Alterations in Synaptic Plasticity by Blockade of L Type Ca2+ Channels. The Journal of Neuroscience 18 3171 3179 (1998). 101 Kumar, A. & Foster, T. C. Intracellular calcium stores contribute to increased susceptibility to LTD induction during aging. Brain Research 1031 125 128 (2005). 1 02 Smith, T. D., Adams, M. M., Gallagher, M., Morrison, J. H. & Rapp, P. R. Circuit Specific Alterations in Hippocampal Synaptophysin Immunoreactivity Predict Spatial Learning Impairment in Aged Rats. J. Neurosci. 20 6587 6593 (2000). 103 Bizon, J. L., P rescott, S., Nicolle, M.M. Intact spatial learning in adult Tg2576 mice. Neurobiology of Aging 28 440 446 (2007). 104 Dumont, M., Strazielle, C., Staufenbiel, M., Lalonde, R. Spatial learning and exploration of environmental stimuli in 24 month old femal e APP23 transgenic mice with the Swedish mutation. Brain Res. 1024 113 121 (2004). 105 Lassalle, J. M., Halley, H.Dumas,S., Verret, L., Francs, B. Effects of the genetic background on cognitive performances of TG2576 mice. Behavioural Brain Research 191 104 110 (2008). 106 Reiserer, R. S., Harrison, F.E., Syverud, D.C., McDonald, M.P. Impaired spatial learning in the APPSwe + PSEN1DeltaE9 bigenic mouse model of Alzheimer's disease. Genes Brain Behav. 6 54 65 (2007). 107 Minkeviciene, R. et al. Ag e related decrease in stimulated glutamate release and vesicular glutamate transporters in APP/PS1 transgenic and wild type mice. Journal of Neurochemistry 105 584 594 (2008). 108 Savonenko, A., Xu, G. M., Melnikova, T., Morton, J. L., Gonzales, V., Wong M. P., Price, D. L., Tang, F., Markowska, A. L., Borchelt, D. R. Episodic like relationships to beta amyloid deposition and neurotransmitter abnormalities. Neurobiol. Dis. 18 602 617 (2005).

PAGE 143

143 109 Puolivli, J., Wang, J., Heikkinen, T., Heikkil, M., Tapiola, T., van Groen, T., Tanila, H. Hippocampal A beta 42 levels correlate with spatial memory deficit in APP and PS1 double transgenic mice. Neurobiology of Disease 9 339 347 (200 2). 110 Gordon, M. N., King, D.L., Diamond, D.M., Jantzen, P.T., Boyett, K.V., Hope, C.E., Hatcher, J.M., DiCarlo, G., Gottschall, W.P., Morgan, D., Arendash, G.W. Correlation between cognitive deficits and Abeta deposits in transgenic APP+PS1 mice. Neuro biology of Aging 22 377 385 (2001). 111 Chen, G., Chen, K.S., Knox, J., Inglis, J., Bernard, A., Martin, S. J., Justice, A., McConlogue, L, Games, D., Fredman, S.B., Morris, R.G. A learning deficit related to age and beta amyloid plaques in a mouse model of Alzheimer's disease. Nature 408 975 979 (2000). 112 Koistinaho, M., Ort, M., Cimadevilla, J.M., Vondrous, R., Cordell, B., Koistinaho, J., Bures, J., Higgins, L.S. Specific spatial learning deficits become severe with age in beta amyloid precursor p rotein transgenic mice that harbor diffuse beta amyloid deposits but do not form plaques. Proc Natl Acad Sci U S A 98 14675 14680. (2001). 113 Lord, A., Englund, H., Sderberg, L., Tucker, S., Clausen, F., Hillered, L., Gordon, M., Morgan, D., Lannfelt, L., Pettersson, F.E., Nilsson, L.N. Amyloid b transgenic mice. FEBS J. 4 995 1006 (2009). 114 Frick, K. M., Stillne r, E.T. and Berger Sweeney J. Mice are not little rats: s pecies differences in a one day water maze task. Neuroreport 11 3461 3465 (2000). 115 Gallagher, M., Burwell, R., Burchinal, M. Severity of spatial learning impairment in aging: development of a learning index for perform ance in the Morris water maze. Behav Neurosci 107 618 626 (1993). 116 Janus, C. Search strategies used by APP transgenic mice during navigation in the Morris water maze. Learning and Memory 11 337 346 (2004). 117 Whishaw, I. Q. J. A. T. Of mice and mazes: similarities between mice a nd rats on dry land but not water mazes. Physiology of Behavior 60 1191 1197 (1996). 118 Barense, M. D., Fox, M.T., Baxter, M.G. Aged rats are impaired on an attentional set shifting task sensitive to medial frontal cortex damage in young rats. Learning and Memory 9 191 201 (2002). 119 Bissonette, G. B., Martins, G.J., Franz, T.M., Harper, E.S., Schoenbaum, G., Powell, E.M. Double dissociation of the effects of medial and orbital prefrontal cortical lesions on attentional and affective shifts in mice. J. Neurosci. 28 11124 11130, doi:10.1523/jneurosci.2820 08.2008 (2008).

PAGE 144

144 120 Sweatt, J. D. Mechanisms of Memory Vol. 1 29 53 (Elsevier, 2003). 121 DeVito, L. M. E., H. Distinct contributions of the hippocampus and medial wh ere like memory in mice. Behav Brain Res (2009). 122 Baxter, M. G., Bucci, D.J., Gorman, L.K., Wiley, R.G., Gallagher, M. Selective immunotoxic lesions of basal forebrain cholinergic cells: effects on learning and memory in rats. Behav Neurosci 109 714 722 (1995). 123 Franklin, K. B. J. & Paxinos, G. The mouse brain in stereotaxic coordinates 3rd edn, (Elsevier, 2008). 124 Desmedt, A., Aline, M., Ren, G. & Robert, J. The effects of ibotenic hippocampal lesions on discri minative fear conditioning to context in mice: impairment or facilitation depending on the associative value of a phasic explicit cue. European Journal of Neuroscience 17 1953 1963 (2003). 125 David L. Brody & Holtzman, D. M. Morris water maze search str ategy analysis in PDAPP mic. exp. Neurol. 197 330 340 (2006). 126 Montgomery, K. S., Mackey, J., Thuett, K., Ginestra, S., Bizon, J.L., Abbott, L.C. Chronic, low dose prenatal exposure to methylmercury impairs motor and mnemonic function in adult C57/B6 mice. Behav Brain Res 191 55 61 (2008). 127 Cahill, L., Vazdarjanova, A. & Setlow, B. The basolateral amygdala complex is Eur. J. Neurosci. 12 3044 3050 (2000). 128 Brown, R. E., Stanford, L. & chellinck, H. M. Developing Standardized Behavioral Tests for Knockout and Mutant Mice. ILAR Journal: Mouse Behavioral Models in Biomedical Research 41 (2000). 129 Bettis, T. J. & Jacobs, L. F. Sex specific strategies in spatial orientation in C57BL/6J mice. Behavioural Processes 82 249 255 (2009). 130 Janus, C. & Westaway, D. Transgenic mouse models of Alzheimer's disease. Physiology & Behavior 73 873 886 (2001). 131 Carter, C. S., Leeuwenburgh, C., Daniels, M. & Fos ter, T. C. Influence of calorie restriction on measures of age related cognitive decline: role of increased physical activity. J Gerontol A Biol Sci Med Sci. 64 850 859 (2009). 132 Halagappa, V. K. et al. Intermittent fasting and caloric restriction amel iorate age related behavioral deficits in the triple transgenic mouse model of Alzheimer's disease. Neurobiol Dis. 26 212 220 (2007).

PAGE 145

145 133 Chung, H. Y. et al. Molecular inflammation: Underpinnings of aging and age related diseases. Ageing Research Reviews 8 18 30 (2009). 134 Carlini, V. P. et al. Decreased memory for novel object recognition in chronically food restricted mice is reversed by acute ghrelin administration. Neuroscience 153 929 934 (2008). 135 Tucci, V., Hardy, A. & Nolan, P. M. A comparis on of physiological and behavioural parameters in C57BL/6J mice undergoing food or water restriction regimes. Behavioural Brain Research 173 22 29 (2006). 136 Hashimoto, T. & Watanabe, S. Chronic food restriction enhances memory in mice analysis with m atched drive levels. Neuroreport 16 1129 1133 (2005). 137 Orsini, C., Buchini, F., Conversi, D. & Cabib, S. Selective improvement of strain dependent performances of cognitive tasks by food restriction. Neurobiology of Learning and Memory 81 96 99 (2004 ). 138 Kirchner, H. et al. Caloric Restriction Chronically Impairs Metabolic Programming in Mice. Diabetes doi:10.2337/db11 1621 (2012). 139 Jankowsky, J. L. et. a. Mutant presenilins specifically elevate the levels of the 42 residue beta amyloid pepti de in vivo: evidence for augmentation of a 42 specific gamma secretase. Hum. Mol. Genet. 13 159 170 (2004). 140 Crawley, J. N. What's Wrong With My Mouse? 2nd edn, 138 140 (Wiley Interscience, 2007). 141 Lambert, M. P. et al. Vaccination with soluble Aβ oligomers generates toxicity neutralizing antibodies. Journal of Neurochemistry 79 595 605 (2001). 142 protein impair synaptic plasticity and behavior. Behavioural Brain Research 192 106 113 (2 008). 143 Weiner, M. W. et al. The Alzheimer's Disease Neuroimaging Initiative: Progress report and future plans. Alzheimer's & dementia : the journal of the Alzheimer's Association 6 202 211.e207 (2010). 144 Wang, A., Das, P., Switzer, R. C., Golde, T. E. & Jankowsky, J. L. Robust Amyloid Clearance in a Mouse Model of Alzheimer's Disease Provides Novel Insights into the Mechanism of Amyloid The Journal of Neuroscience 31 4124 4136 (2011). 145 Saura, C. A. et al. Conditional Inactivatio n of Presenilin 1 Prevents Amyloid Accumulation and Temporarily Rescues Contextual and Spatial Working Memory Impairments in Amyloid Precursor Protein Transgenic Mice. J. Neurosci. 25 6755 6764, doi:10.1523/jneurosci.1247 05.2005 (2005).

PAGE 146

146 146 Scholtzova, H et al. Memantine leads to behavioral improvement and amyloid reduction in Alzheimer's disease model transgenic mice shown as by micromagnetic resonance imaging. Journal of Neuroscience Research 86 2784 2791 (2008). 147 Wilcock, D. M., Rojiani, A., Rose nthal, A., Subbarao, S., Freeman, M.J., Gordon, M.N., et al. Passive immunotherapy against Abeta in aged APP transgenic mice reverses cognitive deficits and depletes parenchymal amyloid deposits in spite of increased vascular amyloid and microhemorrhage. J. Neuroinflammation 1 24 (2004). 148 Oddo, S., Caccamo, A., Kitazawa, M., Tseng, B. P. & LaFerla, F. M. Amyloid deposition precedes tangle formation in a triple transgenic model of Alzheimer's disease. Neurobiology of Aging 24 1063 1070 (2003). 149 M cCarty, M. F. Toward prevention of alzheimers disease Potential nutraceutical strategies for suppressing the production of amyloid beta peptides. Medical Hypotheses 67 682 697 (2006). 150 McGowan, E., Pickford, F., Kim, J., Onstead, L., Eriksen, J., Yu C., et al. Abeta42 is essential for parenchymal and vascular amyloid deposition in mice. Neuron 47 191 199 (2005). 151 Nathalie, P. & Jean Noe¨, l. O. Processing of amyloid precursor protein and amyloid peptide neurotoxicity. Curr. Alzheimer Res. 5 92 99 (2008). 152 Thinakaran, G., Koo, E.H. Amyloid precursor protein trafficking, processing, and function. J Biol Chem 283 29615 19619 (2008). 153 Chan, S. L., Furukawa, K., Mattson, M.P. Presenilins and APP in Neuritic and Synaptic Plasticity. Neuromol ecular Medicine 2 167 196 (2002). 154 Holcomb, L., Gordon, M.N., McGowan, E., Yu, X., Benkovic, S., Jantzen, P., Wright, K., Saad, I., Mueller, R., Morgan, D., Sanders, S., Zehr, C., O'Campo, K., Hardy, J., Prada, C.M., Eckman, C., Younkin, S., Hsiao, K. Duff, K. Accelerated Alzheimer type phenotype in transgenic mice carrying both mutant amyloid precursor protein and presenilin 1 transgenes. Nature 4 97 100 (1998). 155 Wang, Y., Greig, N. H., Yu, Q. s. & Mattson, M. P. Presenilin 1 mutation impairs ch olinergic modulation of synaptic plasticity and suppresses NMDA currents in hippocampus slices. Neurobiology of Aging 30 1061 1068 (2009). 156 Sinha, S. et al. Purification and cloning of amyloid precursor protein bera secretase from human brain. Nature 4 03 537 540 (1999). 157 Disease. J. Mol. Neurosc. 17 157 170 (2001).

PAGE 147

147 158 Vassar, R. et al. Beta Secretase Cleavage of Alzheimer's Amyloid Precursor Protein by the Transmembrane Asparti c Protease BACE. Science 286 735 741 (1999). 159 Hardy, J., Selkoe, D.J. and Problems on the Road to Therapeutics. Science 297 353 356 (2002). 160 Burgess, B. L., McIsaac, S.A., Naus, K.E., Chan, J.Y., Tansley, G.H., Yang, J., Miao, F., Ross, C.J., van Eck, M., Hayden, M.R., van Nostrand, W., St George Hyslop, P., Westaway, D., Wellington, C.L. Elevated plasma triglyceride levels precede amyloid deposition in Alzheimer's disease mouse models with abundant A beta in plasma. Neurobiology f Disease. 24 114 127 (2006). 161 Wengenack, T. M., Whelan, S., Curran, G.L., Duff, K.E., Poduslo, J.F. transgenic mouse brain. Neur oscience 101 929 944 (2000). 162 Mufson, E. J., Bothwell, M., Kordower, J.H. Loss of nerve growth factor receptor subregions of the basal forebrain. Experimental Neurology 105 (198 9). 163 Djordjevic, J., Jones Gotman, M, De Sousa, K., Chertkow, H. Olfaction in patients with mild cognitive impairment and Alzheimer's disease. Neurobiology of Aging 29 693 706 (2008). 164 van Groen, T., Kiliaan, A. J. & Kadish, I. Deposition of mouse amyloid [beta] in human APP/PS1 double and single AD model transgenic mice. Neurobiology of Disease 23 653 662 (2006). 165 Bahar Fuchs A et al. Olfactory deficits and amyloid disease, mild cognitive impairment, and healthy aging: a PiB PET study. J Alzheimers Dis 22 1081 1087 (2010). 166 Heikkinen, T., Kalesnykas, G., Rissanen, A., Tapiola, T., Iivonen, S., Wang, J., Chaudhuri, J., Tanila, H., Miettinen, R., Puolivli, J. Estrogen treatment improves spatial learning in APP + PS1 mice but does not affect beta amyloid accumulation and plaque formation. Experimental Neurology 187 105 117 (2004). 167 Liu, L., Ikonen, S., Heikkinen, T., Heikkil, M., Puolivli, J., van Groen, T., Tanila, H. Effects of fimbria fornix lesion and amylo id pathology on spatial learning and memory in transgenic APP+PS1 mice. Behav Brain Res 134 433 445 (2002). 168 LaFerla, F. M., Green, K. N. & Oddo, S. Intracellular amyloid [beta] in Alzheimer's disease. Nat Rev Neurosci 8 499 509 (2007).

PAGE 148

148 169 Shankar, G. M., Bloodgood, B.L., Townsend, M., Walsh, D.M., Selkoe, D.J., Sabatini, B.L. Natural oligomers of the Alzheimer amyloid reversible synapse loss by modulating an NMDA type glutamate receptor dependent signaling pathway. J. Neurosci. 27 2866 2875 (2007). 170 Bojarski, L., Herms, J. & Kuznicki, J. Calcium dysregulation in Alzheimer's disease. Neurochemistry International 52 621 633 (2008). 171 Davis, J. et al. Early onset and Robust Cerebral Microvascular Accumulation of Prot ein in Transgenic Mice Expressing Low Levels of a Vasculotropic Protein Precursor. J Biol Chem 279 20296 20306. (2004). 172 Fan, R., DeFilippis, K. & Van Nostrand, W. E. Induction of complement proteins in a mouse mode J Neuroinflamm. 4 22 (2007). 173 Grabowski, T. J., Cho, H. S., Vonsattel, J. P., Rebeck, G. W. & Greenberg, S. M. Novel amyloid precursor protein mutation in an Iowa family with dementia and severe cerebral amy loid angiopathy. Ann Neurol. 49 697 705 (2001). 174 Aucoin, J. S. et al. Selective cholinergic denervation, independent from oxidative stress, in a mouse model of Alzheimer's disease. Neuroscience 132 73 86 (2005). 175 Boom, A. et al. Astrocytic calciu m/zinc binding protein S100A6 over expression in Alzheimer's disease and in PS1/APP transgenic mice models. Biochimica et Biophysica Acta (BBA) Molecular Cell Research 1742 161 168 (2004). 176 Garcia Alloza, M. et al. Characterization of amyloid deposi tion in the APPswe/PS1dE9 mouse model of Alzheimer disease. Neurobiology of Disease 24 516 524 (2006). 177 Zhuo, J. M., Prakasam, A., Murray, M.E., Zhang, H.Y., Baxter, M.G., Sambamurti, K., Nicolle, M.M. An increase in Abeta42 in the prefrontal cortex i s associated with a reversal learning impairment in Alzheimer's disease model Tg2576 APPsw mice. Current Alz. Res. 5 385 391 (2008). 178 Montgomery, K. S. Novel age dependent learning deficits in a mouse model of ons for tra nslational research. (2011 ). 179 Volianskis, A., Kstner, R., Mlgaard, M., Hass, S. & Jensen, M. S. Episodic memory deficits are not related to altered glutamatergic synaptic transmission and plasticity in the CA1 hippocampus of the APPswe/PS1[Delta]E9 d eleted transgenic mice model of [beta] amyloidosis. Neurobiology of Aging In Press, Corrected Proof

PAGE 149

149 180 Xu, F. et al. Early onset subicular microvascular amyloid and neuroinflammation pr otein precursor transgenic mice. Neuroscience 146 98 107 (2007). 181 Gruart, A., Lpez Ramos, J.C., Muoz, M.D., Delgado Garca, J.M. Aged wild type and APP, PS1, and APP + PS1 mice present similar deficits in associative learning and synaptic plasticity independent of amyloid load. Neurobiology of Disease 30 439 450 (2008). 182 Ann. Neurol. 41 7 (1997). 183 Smith, T. D., Calhoun, M.E., Rapp, P.R. Circuit and morphological specificity o f synaptic change in the aged hippocampal formation. Neurobiol. of Aging 20 357 358 (1999). 184 Masliah, E. et al. Topographical distribution of synaptic associated proteins in the neuritic plaques of Alzheimer's disease hippocampus. Acta Neuropathologic a 87 135 142, doi:10.1007/bf00296182 (1994). 185 Masliah, E. et al. Synaptic and neuritic alterations during the progression of Alzheimer's disease. Neuroscience Letters 174 67 72 (1994). 186 Masliah, E., Terry, R. D., DeTeresa, R. M. & Hansen, L. A. I mmunohistochemical quantification of the synapse related protein synaptophysin in Alzheimer disease. Neuroscience Letters 103 234 239 (1989). 187 Sze, C. I., Troncoso, J.C., Kawas, C., Mouton, P., Price, D.L., Martin, L.J. Loss of the presynaptic vesicle protein synaptophysin in hippocampus correlates with cognitive decline in Alzheimer disease. J. Neuropathol. Exp. Neurol. 56 933 944 (1997). 188 Hu, X. et al. BDNF Induced Increase of PSD 95 in Dendritic Spines Requires Dynamic Microtubule Invasions. Th e Journal of Neuroscience 31 15597 15603, doi:10.1523/jneurosci.2445 11.2011 (2011). 189 Shao, C. Y., Mirra, S. S., Sait, B. R. H., Sacktor, T. C. & Sigurdsson, E. M. Postsynaptic degeneration as revealed by PSD 95 reduction occurs after Acta Neuropathologica 122 285 292 (2011). 190 Miao, J. et al. Vascular Degeneration and Neuroinflammation in Transgenic Mice Expressing Hu Am J Pathol. 167 505 515 (2005).

PAGE 150

150 191 West, M. J., Bach, G., Sderman, A. & Jensen, J. L. Synaptic contact number and size in stratum radiatum CA1 of APP/PS1[Delta]E9 transgenic mice. Neurobiology of A ging 30 1756 1776 (2009). 192 Heinemann, U., D, S., C, E. & T, G. Properties of Entorhinal Cortex Projection Cells to the Hippocampal Formation. Annals of the New York Academy of Sciences 911 112 126 (2000). 193 Radde, R., Bolmont, T., Kaeser, S.A., C oomaraswamy, J., Lindau, D., Stoltze, L., Calhoun, M.E., Ja¨ggi, F., Wolburg, H., Gengler, S., Haass, C., Ghetti, B., Czech, C., Ho¨lscher, C., Mathews, P.M., Jucker, M. Ab42 driven cerebral amyloidosis in transgenic mice reveals early and robust pathology European Molecular Biology Organization 7 (2006). 194 Gomes, F. C. A., Paulin, D. & Moura Neto, V. Glial fibrillary acidic protein (GFAP): modulation by growth factors and its implication in astrocyte differentiation. Brazillian Journal of Medical and B iological Research 32 619 631 (1999). 195 Lovell, M. A., Geiger, H., Van Zant, G. E., Lynn, B. C. & Markesbery, W. R. Isolation of neural precursor cells from Alzheimer's disease and aged control postmortem brain. Neurobiology of Aging 27 909 917 (2006) 196 Simpson, J. E., Ince, P.G., Lace, G., Forster, G., Shaw, P.J., Matthews, F., Savva, G., Brayne, C., Wharton, S.B. Astrocyte phenotype in relation to Alzheimer type pathology in the ageing brain. Neurobiology of Aging (2008). 197 Frye, C. A. & Walf, A. A. Effects of progesterone administration and APPswe+PSEN1[Delta]e9 mutation for cognitive performance of mid aged mice. Neurobiology of Learning and Memory 89 17 26 (2008). 198 Hoxha, E., Boda, E., Montarolo, F., Parolisi, R. & Tempia, F. Excitabili ty and Synaptic Alterations in the Cerebellum of APP/PS1 Mice. PLOS One 7 e34726 (2012). 199 Lalonde, R., Kim, H.D., Fukuchi, K. Exploratory activity, anxiety, and motor coordination in bigenic APPswe + PS1/DeltaE9 mice. Neurosci. Lett. 369 156 161 (2004). 200 Jack, C. R., Jr., Marjanska, M., Wengenack, T. M., Reyes, D. A., Curran, G. L., Lin, J., Preboske, G. M., Poduslo, J. F., Garwood, M. Magnetic Resonance Imaging of Alzheimer's Pathology in the Brains of Living Transgenic Mice: A New Tool in Alzheimer's Disease Research. Neuroscientist 13 38 48 (2007). 201 Schubert, W. et al. Localization of Alzheimer [beta]A4 amyloid precursor protein at central and peripheral synaptic sites. Brain Research 563 184 194 (1991).

PAGE 151

151 202 Bell, K. F., de Kort G. J ., Steggerda S., Shigemoto R., Ribeiro da Silva A., Cuello A. C. Structural involvement of the glutamatergic presynaptic boutons in a transgenic mouse model expressing early onset amyloid pathology. Neurosci. Lett. 353 143 147 (2003). 203 Chan, S. L., F urukawa, K., Mattson, M.P. Presenilins and APP in neuritic and synaptic plasticity: implications for the pathogenesis of Alzheimer's disease. Neuromolecular Med. 2 167 196 (2002). 204 Dickey, C. A. et al. Selectively Reduced Expression of Synaptic Plasti city Related Genes in Amyloid Precursor Protein + Presenilin 1 Transgenic Mice. J. Neurosci. 23 5219 5226 (2003). 205 Gengler, S., Hamilton, A. & Hlscher, C. Synaptic Plasticity in the Hippocampus of a APP/PS1 Mouse Model of Alzheimer's Disease Is Impai red in Old but Not Young Mice. PLOS One 5 e9764 (2010). 206 Giacchino, J., Criado, J. R., Games, D. & Henriksen, S. In vivo synaptic transmission in young and aged amyloid precursor protein transgenic mice. Brain Research 876 185 190 (2000). 207 Hsia, A. Y. et al. Plaque independent disruption of neural circuits in Alzheimer' s disease mouse models. Proceedings of the National Academy of Sciences of the United States of America 96 3228 3233 (1999). 208 Jorgensen, A. B., Dalby, N. O., Mork, A., Veng, L M. Presynaptic glutamatergic Neurodegenerative Dis. 4 241 (2007). 209 Lazarov, O., Lee, M., Peterson, D. A. & Sisodia, S. S. Evidence That Synaptically Released beta Amyloid Accumulates as Extracellular Deposits in the Hippocampus of Transgenic Mice. J. Neurosci. 22 9785 9793 (2002). 210 Rutten, B. P. F. et al. Age Related Loss of Synaptophysin Immunoreactive Presynaptic Boutons within the Hippocampus of APP751SL, PS1M146L, and APP751SL /PS1M146L Transgenic Mice. Am J Pathol 167 161 173 (2005). 211 Gong, Y., Chang, L., Viola, K.L., Lacor, P.N., Lambert, M.P., Finch, C.E., Krafft, G.A., Klein, W.L. Alzheimer's disease affected brain: presence of oligomeric A beta ligands (ADDLs) suggests a molecular basis for reversible memory loss. Proc Natl Acad Sci U S A 100 10417 10422 (2003). 212 Walsh, D. M., Selkoe, D.J. AB Oligomers a decade of discovery. Journal of Neurochemistry 101 1172 1184 (2007).

PAGE 152

152 213 Lakshmana, M. K. et al. Role of Ra nBP9 on amyloidogenic processing of APP and synaptic protein levels in the mouse brain. The FASEB Journal (2012). 214 Leuba, G. et al. Postsynaptic density protein PSD 95 expression in Alzheimer's disease and okadaic acid induced neuritic retraction. Neur obiology of Disease 30 408 419 (2008). 215 Bach, P. et al. Vaccination with A{beta} Displaying Virus Like Particles Reduces Soluble and Insoluble Cerebral A{beta} and Lowers Plaque Burden in APP Transgenic Mice. J Immunol 182 7613 7624, doi:10.4049/jimm unol.0803366 (2009). 216 Francis, P. T., Pangalos, M.N., Stephens, P.H., Bartlett, J.R., Bridges, P.K., Malizia, A.L., Neary, D., Procter, A.W., Thomas, D.J., Bowen, D.M. Antemortem measurements of neurotransmission: possible implications for pharmacother apy J. Neurol. Neurosurg. Psychiatry 56 80 84 (1993). 217 Hsieh, H., Boehm, J., Sato, C., Iwatsubo, T., Tomita, T., Sisodia, S., and Malinow, R. AMPAR removal underlies Abetainduced synaptic depression and dendrit ic spine loss. Neuron 52 (2006). 218 Merino Serrais, P. et. al. Layer specific alterations to CA1 dendritic spines in a mouse model of Alzheimer's disease. Hippocampus 21 1037 1044, doi:10.1002/hipo.20861 (2011). 219 Perez Cruz, C. et al. Reduced Spine Density in Specific Regions of CA1 Pyramidal Neurons in Two Transgenic Mouse Models of Alzheimer's Disease. The Journal of Neuroscience 31 3926 3934, doi:10.1523/jneurosci.6142 10.2011 (2011). 220 Gomez Isla, T., Price, J.L., McKeel, D.W. Jr., Morris, J. C., Growdon, J.H., Hyman, B.T. Profound loss of layer II entorhinal cortex neurons occurs in very J. Neurosci. 16 4491 4500 (1996). 221 Jauhiainen, A. M., Pihlajamki, M., Tervo, S., Niskanen, E., Tanila, H., Hnninen, T., Vann inen, R.L., Soininen, H. Discriminating accuracy of medial temporal lobe volumetry and fMRI in mild cognitive impairment. Hippocampus 19 166 175 (2009). 222 Wakabayashi, K., Honer, W. G. & Masliah, E. Synapse alterations in the hippocampal entorhinal for mation in Alzheimer's disease with and without Lewy body disease. Brain Res. 665 24 32 (1994). 223 Meyer Luehmann, M. et al. Rapid appearance and local toxicity of amyloid beta Nature 451 720 724 (2008).

PAGE 153

153 224 Weinstein DE, Shelanski ML & RK., L. Suppression by antisense mRNA demonstrates a requirement for the glial fibrillary acidic protein in the formation of stable astrocytic processes in response to neurons. J Cell Biol. 112 1205 1213 (1991). 225 Kats unori, Y., Kazuho, H., Akio, K., Jian wen, H. & Masaru, K. Expression of amyloid precursor protein like molecule in astroglial cells of the subventricular zone and rostral migratory stream of the adult rat forebrain. Journal of Anatomy 205 135 146 (2004). 226 Hickman, S. E., Allison, E. K. & El Khoury, J. Microglial Dysfunction and Amyloid Clearance Pathways in Aging Alzheimer's Disease Mice. The Journal of Neuroscience 28 8354 8360, doi:10.1523/jneurosci.0616 08.2008 (2008). 227 Tarsa, L. & Goda, Y. Synaptophysin regulates activity dependent synapse formation in cultured hippocampal neurons. Proceedings of the National Academy of Sciences 99 1012 1 016 (2002). 228 Priller, C. et al. Excitatory synaptic transmission is depressed in cultured hippocampal neurons of APP/PS1 mice. Neurobiology of Aging 30 1227 1237 (2009). 229 Fitzjohn, S. M., Morton, R.A., Kuenzi, F., Rosahl, T.W., Shearman, M., Lewis, H., Smith, D., Reynolds, D.S., Davies, C.H., Collingridge, G.L., Seabrook, G.R. Age relate d impairment of synaptic transmission but normal long term potentiation in transgenic mice that overexpress the human APP695SWE mutant form of amyloid precursor protein. J. Neurosci. 21 4691 4698 (2001). 230 Ma, T. et al. Induced Impairments in Hippocampal Synaptic Plasticity Are Rescued by Decreasing Mitochondrial Superoxide. The Journal of Neuroscience 31 5589 5595 (2011). 231 Hickman, S. E., Allison, E. K. & El Khoury, J. Microglial Dysfunction and Defective Amyloid Clearance Pathways in A J Neurosci 28 8354 8360 (2008). 232 Malm, T. M. et al. Pyrrolidine Dithiocarbamate Activates Akt and Improves Amyloid Burden. J Neurosci 27 3712 3721 (2007). 233 Fan, R ., DeFilippis, K. & Van Nostrand, W. E. Induction of complement proteins J Neuroinflammation 4 22 (2007).

PAGE 154

154 234 Zhuo, J. M., Prescott, S.L., Murray, M.E., Zhang, H.Y., Baxter, M.G., Nicolle, M.M. E arly discrimination reversal learning impairment and preserved spatial learning in a longitudinal study of Tg2576 APPsw mice. Neurobiology of Aging 28 1248 1257 (2007). 235 Schoenbaum, G., Setlow, B., Saddoris, M.P., Gallagher, M. Encoding Changes in O rbitofrontal Cortex in Reversal Impaired Aged Rats J. Neurophysiol 95 1509 1517 (2006). 236 Schoenbaum, G., Nugent, S., Saddoris, M.P., Setlow, B. Orbitofrontal lesions in rats impair reversal but not acquisition of go, no go odor discriminations. Ne uroreport 13 885 890 (2002). 237 Bissonette, G. B. et al. Double Dissociation of the Effects of Medial and Orbital Prefrontal Cortical Lesions on Attentional and Affective Shifts in Mice. J. Neurosci. 28 11124 11130 (2008). 238 Selkoe, D. J. Alzheimer' s Disease Is a Synaptic Failure. Science 298 789 791, ( 2002). 239 Golde, T. E., Schneider, L. S. & Koo, E. H. Anti disease: the need for a paradigm shift. Neuron 69 203 213 (2011). 240 Klunk, W. E. et. al. Imaging brain a compound B. Ann. Neurol. 55 306 319 (2004). 241 Li, S., Hong, S., Shepardson, N.E., Walsh, D.M., Shankar, G.M., Selkoe, D. Soluble Oligomers of Amyloid [beta] Protein Facilitate Hippocampal Long Term Depressi on by Disrupting Neuronal Glutamate Uptake. Neuron 62 788 801 (2009). 242 Walsh, D. M. et al. Naturally secreted oligomers of amyloid [beta] protein potently inhibit hippocampal long term potentiation in vivo. Nature 416 535 539 (2002). 243 Lacor, P. N et al. Synaptic targeting by Alzheimer's related amyloid beta oligomers. J Neurosci 24 10191 10200 (2004). 244 N methyl D aspartate receptor dependent mechanism that is b locked by the Alzheimer drug memantine. J. Bio. Chem. 282 11590 11601 (2007). 245 De Felice, F. G., Velasco, P.T., Lambert, M.P., Viola, K., Fernandez, S.J., Ferreira, S.T., Klein, W.L. Abeta Oligomers Induce Neuronal Oxidative Stress through an N Methyl D aspartate Receptor dependent Mechanism That Is Blocked by the Alzheimer Drug Memantine. J. Biol. Chem. 282 11590 1 1601 (2007).

PAGE 155

155 246 Cerpa W et al. Wnt 5a occludes Abeta oligomer induced depression of glutamatergic transmission in hippocampal neurons. Mol Neurodegener. 5 3 (2010). 247 Li, L., Cheung, T., Chen, J. & Herrup, K. A Comparative Study of Five Mouse Models of Alzheimer's Disease: Cell Cycle Events Reveal New Insights into Neurons at Risk for Death. International Journal of Alzheimer's Di sease 2011 (2011). 248 Lacor, P. N. et al. Synaptic Targeting by Alzheimer's Oligomers. The Journal of Neuroscience 24 10191 10200, (2004). 249 Andersen, K.L. et.al. A Gender differences in the incidence of AD and vascular dementia: T he EURODEM Studies. EURODEM Incidence Research Group. Neurology 53 1992 1997 (1999). 250 Xu, F. e t a l Early onset subicular microvascular amyloid and neuroinflammation protein precurs or transgenic mice. Neuroscience 146 98 107 (2007). 251 Wyss Coray, T. & Mucke, L. Inflammation in neurodegenerative disease: A double edged sword. Neuron 35 419 432 (2006). 252 Belanger, M. & Magistretti, P. J. The role of astroglia in neuroprotection Dialogues Clin Neurosci 11 281 295 (2009). 253 Sofroniew, M. V. & Vinters, H. V. Astrocytes: Biology and Pathology. Acta Neuropathol. 119 7 35 (2010). 254 Tseng, H. C. et al. Injuring neurons induces neuronal differentiation in a population of hippoc ampal precursor cells in culture. Neurobiology of Disease 22 88 97 (2006). 255 Rodriguez, J. J., Olabarria, M., Chvatal, A. & Verkhratsky, A. Astroglia in dememntia and Alzheimer's disease. Cell Death Differ 16 378 385 (2009). 256 Wyss Coray, T. et al. Adult mouse astrocytes degrade amyloid Beta in vitro and in situ. Nat Med 9 453 457 (2003). 257 Chow, S. K., Yu, D., Macdonald, C. L., Buibas, M. & Silva, G. A. Amyloid beta peptide directly induces spontaneous calcium transients, delayed intercellular calcium waves and gliosis in rat cortical astrocytes. ASN Neuro 2 e00026 (2010).

PAGE 156

156 258 White, J. A., Manelli, A. M., Holmberg, K. H., Van Eldik, L. J. & Ladu, M. J. Differential effects of oligomeric and fibrillar amyloid beta 1 42 on astrocyte mediated in flammation. Neurobiol Dis 18 459 465 (2005). 259 Bialowas McGoey, L. A., Lesicka, A., Whitaker Azmitia, P.M. Vitamin E increases S100B mediated microglial activation in an S100B overexpressing mouse model of pathological aging. Glia 56 1780 1790 (2008). 260 Frautschy, S. A., Yang, F., Irrizarry, M., Hyman, B., Saido, T. C., Hsiao, K., Cole, G. M. Microglial response to amyloid plaques in APPsw transgenic mice. Am. J. Pathol. 152 307 317 (1998). 261 Araque, A., Parpura, V., Sanzgiri, R. P. & Haydon, P. G. Tripartite synapses: glia, the unacknowledged partner. Trends in Neurosciences 22 208 215 (1999).

PAGE 157

157 BIOGRAPHICAL SKETCH Karie nn Montgomery graduated with a bachelor of science in b iochemistry (2006) from Texas A&M University. She was accepted to the Behavioral and Cellular laboratory. Findings from her first year project indicated that prenatal exposure to methylmercury in mice, even at the lowest dose examined to date, can have long lasting motor and cognitive consequences for adult s. Based on this publication, Dr American Psychological Association. During her second year of graduate school, she w as also involved in the characterization of aged Fisher 344 rats in behavioral tasks with the goal of developing a rodent model of mild cognitive impairment. During this he long term deleterious effects of cocaine on learning and memory. More recently, Dr Montgomery has established a transfer learning task in mice which is analogous to a disease. This novel task is a promising tool for behavioral characterization of rodent models of neurodegenerative disease, and is well suited for clinically relevant within subject intervention studies. Data from this project was submitted for a NRSA fell owship, which Dr Montgomery was awarded in the fall of 2010. Dr Montgomery transferred to the University of Florida College o f Medicine with the Bizon laboratory in 2010 and joined the laboratory of Dr. Daoyun Ji as a post doctoral fellow in the departme nt of Molecular Biology at Baylor College of Medicine in October of 2012