|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
Effects of A-beta Immunothera py and Omega-3 Fatty Acid Administration in Alzheimers Transgenic Mice by Maren T. Jensen A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Biology College of Arts and Sciences University of South Florida Major Professor: Gary Arendash, Ph.D. Huntington Potter, Ph.D. Sidney Pierce, Ph.D. Brian Livingston, Ph.D. Date of Approval: March 6, 2006 Keywords: Active Vaccinati on, Fish Oil, DHA, Cognitive Performance, Plaque Deposition Copyright 2006, Maren T. Jensen
Acknowledgements I thank Dr. David Morgan for providing the animals utilized in the A immunotherapy study and Dr. Marcia Gordon for her involveme nt in the animal vaccinations and A immunostaining. I would also like to thank Purina Mills Test Diet for providing the low omega-3, omega-3-deficient, high n-3 and high n-6 fatty acid diets and Barrow-Agee Laboratories for analyzing all four of the fatty acid diets. In a ddition, I would like to thank Dr. Norman Salem and Dr. Nahed Hussein for their determination of fatty acid levels from the frontal cortex.
i Table of Contents List of Tables................................................................................................................. ....iii List of Figures................................................................................................................ .....v Introduction................................................................................................................... ......1 Alzheimers Disease Background.................................................................................1 AD Behavioral Deficits.................................................................................................2 AD Pathology................................................................................................................5 Cerebral Atrophy, Neuronal Loss...........................................................................5 Neuritic Plaques......................................................................................................7 Neurofibrillary Tangles.........................................................................................12 Genetics of AD...........................................................................................................16 Diagnosis of AD.........................................................................................................18 Risk Factors for AD....................................................................................................25 Treatment of AD.........................................................................................................29 Animal Models............................................................................................................36 PDAPP Transgenic Mice......................................................................................38 APPsw Transgenic Mice.......................................................................................44 APP/PS1 Transgenic Mice....................................................................................52 AD Vaccination..........................................................................................................55 In Vitro Studies.....................................................................................................55 Active A Immunotherapy...................................................................................56 Passive A Immunotherapy..................................................................................64 A Immunotherapy in Humans............................................................................69 Omega-3 Fatty Acids.................................................................................................70 General Background.............................................................................................70 Dietary Manipulation of n3 and n-6 Fatty Acids.................................................75 Human Studies Involving Fatty Acids..................................................................83 Neurochemical Effects of n-3 Fatty Acids............................................................89 Effects of n-3 Fatty Acids on 2nd Messengers......................................................97 Behavioral Effects of n-3 Fatty Acids in Animals..............................................104 Statement of Purpose................................................................................................115 Materials and Methods....................................................................................................117 Life-long Vaccination Study.....................................................................................117 Omega-3 Fatty Acid StudySurvival Analysis........................................................121
ii Omega-3 Fatty Acid StudyBehavioral, Pathologic and Neurochemical Assessments........................................................................126 Behavioral Testi ng Procedures.................................................................................136 Statistical Analysis....................................................................................................143 Results........................................................................................................................ .....146 Life-long Vaccination Study.....................................................................................146 Sensorimotorand Anxiety-based Tasks............................................................147 Cognitive-based Tasks........................................................................................151 A Histopathology..............................................................................................167 FA and Correlation Analyses..............................................................................169 DFA.....................................................................................................................172 Omega-3 Fatty Acid StudySurvival Analyses........................................................176 Omega-3 Fatty Acid Study: Behavioral, Pathologic and Neurochemical Assessments........................................................................181 Sensorimotorand Anxiety-Based Tasks............................................................184 Cognitive-based Tasks........................................................................................188 Cytokine Levels..................................................................................................211 Hippocampal A levels.......................................................................................211 Fatty Acid Brain Tissue Levels...........................................................................216 Correlations and Multimetric Analyses..............................................................222 Discussion..................................................................................................................... ..238 A Vaccination Discussion.......................................................................................238 Omega-3 Discussion................................................................................................248 Survival Analysis................................................................................................249 Behavioral, Pathologic and Ne urochemical Assessments..................................251 Overall Conclusions..................................................................................................267 References..................................................................................................................... ..269
iii List of Tables Table 1. Summary of behavioral st udies involving n-3 deficiency or supplementation to rodents . ..107 Table 2. Percent fatty acid composition of total fat for standard chow, low omega-3 diet and omega-6 deficient diets ... ...125 Table 3. Percent fatty acid composition of total fat for standard chow, high omega-3 and high omega-6 diets .... ...130 Table 4. A summary of transgen ic and immunotherapy at both 4-6 and 15-16 month test points ... ..146 Table 5. Total A and compact A for Tg+/Con and Tg+/A at 17 months of age within fr ontal cortex and hippocampus 168 Table 6. Factor Loadings of be havioral measures, with and Without pathologic measures ... .171 Table 7. Summary of discri minant function analyses . 174 Table 8. A summary of transgenic and dietary behavioral effects (high n-3 vs. high n-6) in NT and Tg+ mice .182 Table 9. A summary of transgenic and dietary behavioral effects (high n-3 vs. standard diet ) in NT and APP/PS1 mice .183 Table10. Correlations for all NT mice combined between cortical fatty acid levels and behavioral measures or for only high n-3 and high n-6 NT mice between fatty acid levels, plasma cytokines and behavioral performance . 224 Table 11. Correlations for combined standard and high n-3 APP/PS1 mice between behavioral measures cortical fatty acid levels and hippocampal A levels or for combined high n-3 and high n-6 APP/PS1 mice between hippocampal A levels, plasma cytokines and brain fatty acid levels ... .227
iv Table 12. Factor loadings of behavi oral measures, with and without pathologic measures ..230 Table 13. Summary of discriminant function analyses including NT and Tg+ mice ..236 Table 14. Summary of discriminant function analyses including NT and APP/PS1 mice for the standard and high n-3 groups . .237
v List of Figures Figure 1. Flow diagram showing AD pathogenesis progressing to cognitive impairment 15 Figure 2. Fatty acid synthesis within liver from dietary intake of -linolenic acid and linoleic acid to ultimately give rise to docosahexaenoic acid and arachidonic acid 74 Figure 3. General protocol time line for life-long vaccination study .119 Figure 4. General protocol time line for omega-3 fatty acid studySurvival Analysis .123 Figure 5. General protocol time line for omega-3 fatty acid study Behavior, Pathologic and Neurochemical Assessments .. ..128 Figure 6. Comparison of sensorim otor function in NT, Tg+/Con, and Tg+/A mice at 4 -6 and 15-16 behavioral test points ..149 Figure 7. Anxiety/emotionality, as determined in the elevated plus-maze by percent time in open arms and arm entries ..150 Figure 8. Y-maze arm entries and percent spontaneous alternation for adult and aged NT, Tg+/Con, and Tg+/A mice .152 Figure 9. Morris water maze acquisiti on at adult (4-6 months) and aged (15-16 months) time points for NT, Tg+/Con, and Tg+/A mice ...153 Figure 10. Probe trial testing fo r reference memory retention, assessed on the day following the completion of water maze acquisition .155 Figure 11. Spatial learning/memory, as determined by the circular platform task latency to find the escape hole across 4 blocks consisting of 2 days each 157
vi Figure 12. Spatial learning/memory, as determined by the circular platform task the number of errors, indexed as the number of head pokes into holes which did not contain the escape hole, across four 2-day blocks ...158 Figure 13. Platform recognition test ing for the ability to search/ identify a variably-placed and conspicuously marked platform over 4 days of testing, with latency to swim to the platform being measured ..160 Figure 14. Working memory in the RAWM task at the 4-6 month test point, with errors be ing evaluated for T1, T4, and T5 over three 3-day blocks 162 Figure 15. Working memory in the RAWM task at the 4-6 month test point, with latency to find the hidden platform being evaluated for T1, T4, and T5 over three 3-day blocks ...163 Figure 16. RAWM testing for worki ng memory at the 15-16 month test point, with number of e rrors indicated for T1, T4, and T5 over four 3-day blocks of testing ..165 Figure 17. RAWM testing for worki ng memory at the 15-16 month test point, as indexed by es cape latency, across four 3-day blocks of testing .166 Figure 18. Canonical score plots of step-wise forward discriminant function analyses used to compare the overall cognitive performance of NT, Tg+/Con, and Tg+/A groups ...175 Figure 19. Survivability analysis show ing days on the omega-3-deficient diet where the F1 PS1 mice resulted in no mortality for up to 3 months after starting on the experimental diet ...178 Figure 20. Survivability analysis show ing days on the omega-3-deficient diet where all of the F2 mice resulted in similar mortality after starting on the experimental diet ...179 Figure 21. Survivability analysis co mparing F1 PS1 mice to F2 PS1 mice showing a significant diffe rence in survival between the two generation within the same genotype 180 Figure 22. Comparison of sensorimotor function in NT and Tg+ mice fed either standard, high n-3 or high n-6 diets ...186
vii Figure 23. Anxiety/emotionality, as dete rmined in the Elevated Plus-maze by percent time in open arms and arm entries ...187 Figure 24. Y-maze arm entries and pe rcent spontaneous alternation for standard, high n-3 and high n-6 diet groups within NT and Tg+ mice .190 Figure 25. Morris water maze acquis ition for standard, high n-3 and high n-6 groups for NT and Tg+ mice ...191 Figure 26. Probe trial testing for re ference memory retention, assessed on the day following the completion of water maze acquisition ...192 Figure 27. Probe trial testing for re ference memory retention, assessed on the day following the completion of water maze acquisition, for both NT and APP/PS1 groups ..193 Figure 28. Spatial learning/memory, as determined by the circular platform task by number of h ead pokes/ errors into non-escape holes and latency to find the es cape hole across 8 days of testing, with the last da y of testing shown below ..195 Figure 29. The last day of circular platform show s a strong effect of diet by combining both genotypes .196 Figure 30. Platform recognition test ing for the ability to search/ identify a variably-placed and conspicuously marked platform over 4 days of testing, with latency to swim to the platform being measured ..199 Figure 31. Platform recognition as assessed within NT and APP/PS1 standard and high n-3 diet groups ..200 Figure 32. Working memory in the RAWM task in NT and Tg+ mice being evaluated for T1, T4 and T5 errors over three 3-day blocks ...203 Figure 33. RAWM errors in overall T4 and T5 for both NT and Tg+ groups .204 Figure 34. RAWM errors in overall T4 and T5 for both NT and APP/PS1 groups .205
viii Figure 35. Working memory in th e RAWM task in NT and Tg+ mice being evaluated latency to find the platform for T1, T4 and T5 over three 3-day blocks 208 Figure 36. RAWM latency to find the hidden platform in overall T4 and T5 for both NT and Tg+ groups ..209 Figure 37. RAWM latency to find the hidden platform within the last block of testing for both NT and APP/PS1 on either the standard or high n-3 fatty acid diets ...210 Figure 38. Standardized mean signal intens ity of 10 different cytokines in both NT and APP/PS1 mice on either the high n-3 or high n-6 fatty acid diets ..213 Figure 39. Standardized mean signal intens ity of 10 different cytokines for all mice on either the high n-3 or high n-6 fatty acid diets despite genotype 214 Figure 40. A levels (pmol/g) as measured by ELISA of insoluble and soluble A 1-40 as well as insoluble and soluble A 1-42 ..215 Figure 41. Cortical levels of tota l saturated and monounsaturated fatty acid levels of standard, high n-3 and high n-6-fed NT and APP/PS1 mice 217 Figure 42. Cortical levels of n-6 and n-3 fatty acids for standard, high n-3 and high n-6 NT and APP/PS1 mice ..220 Figure 43. Total n-6 and total n-3 fatty acid levels in standard, high n-3 and high n-6 NT and APP/PS1 mice ..221 Figure 44. Canonical scores plot of stepwise-for ward DFA using all 19 behavioral measures for Tg+ groups .234 Figure 45. Canonical scores plot of direct entry DFA utilizing the seven cognitive measures derived from factor 1 for both NT and Tg+ diet groups ..235
ix Effects of A-beta Immunot herapy and Omega-3 Fatty Acid Administration in Alzheimers Transgenic Mice Maren T. Jensen ABSTRACT Major therapeutics against Alzheimers disease (AD) are targeted towards reducing -amyloid in the brain and improving cognitive performance. Transgenic mouse models of AD have become critical in the development of such therapeutics to protect against or treat AD. Th is dissertation examined the pot ential protective effects of both active A immunotherapy and dietary omega3 fatty acid administration to AD transgenic mice. First, immunization with A 1-42 from 2-16 months of age provided protection against cognitive impairment in APP/ PS1 transgenic mice well into older age. At both adult (4-6 month) and aged (15-16 month) test points, an extensive 6-week behavioral battery was admini stered that measured multiple sensorimotor and cognitive domains. A immunotherapy either partially or completely protected APP/PS1 mice from impairment in reference lear ning/memory, working memory and/or recognition/identification at these test points. However, behavioral protection at the later test point occurred with out any reduction in A deposition within the brain. Therefore, the cognitive benefits of A immunotherapy most likely involved neutralization or removal of A oligomers from the brain. In additi on to immunotherapy, this dissertation
x also examined the behavioral and neurochemi cal effects of a high omega-3 (n-3) or high omega-6 fatty acid (n-6) diet to NT and AD transgenic (Tg+) mice from 2 through 9 months of age. The same 6-week behavior al test battery, as described above, was administered between 7-9 months of age. In NT mice, dietary n-3 or n-6 fatty acids did not result in any beneficial e ffects on cognitive performance. In Tg+ mice, a high n-3 diet improved some, but not most, cognitive skills in comparison to standard-fed Tg+ mice; whereas a diet high in n-6 fatty acids di d not lead to widespread deficits in learning or memory. In fact, there was no differen ce in overall performance on any behavioral task between Tg+ mice given a high n-3 or hi gh n-6 diet. Administra tion of dietary fatty acids did not result in any significan t changes in soluble or insoluble A levels within the brains of Tg+ mice and plasma cytokine le vels in Tg+ mice were largely unaffected. Notably, neither the high n-3 nor high n-6 diet increased cortical levels of n-3 or n-6 fatty acids, respectively, within Tg+ mice. Howeve r, NT mice on a high n-3 or high n-6 diet did show significant elevations in cortical n-3 or n-6 fatty acid levels, respectively, suggesting that Tg+ mice have a deficit in incorp oration of dietary fatt y acids in the brain. Collectively, these results show that life-long administration of active A immunotherapy provides clear cognitive protection well into ol der age, whereas long-term dietary omega3 fatty acid administration does not provide extensiv e cognitive benefit. Both studies underscore the value of using AD transgen ic mice in determining the efficacy of prophylactics against AD.
Introduction I. Alzheimers Disease Background Alzheimers disease, first diagnosed by Alois Alzheimer in 1906, was not widely accepted as a disease affecting the elderly until the late 1960s. Within this period Alzheimers disease (AD) arose as the most common form of seni le dementia (Selkoe, 2001). The prevalence of AD is most common among people over 85 years of age (Munoz & Feldman, 2000). Primary clinical symptoms of AD include short-term and long-term memory loss, paranoia, delusions, decreased language abi lity and an overall cognitive impairment. The two hallmark pa thologic features of AD include neuritic plaques and neurofibrillary tang les. Silver staining was firs t used in the early 1900s to indicate neurofibrillary tangles (NFTs) within neuronal cell s. Following that discovery, in 1930 Divry used Congo Red staining to loca te senile plaques composed of amyloid (Maccioni et al., 2001). In the 1960s Kidd and Terry first described in detail the composition of neuritic plaques and NFTs, which are composed of paired helical filaments (Selkoe, 2001). General risk fact ors for the disease include age, genetic mutations, and lifestyle influences such as high blood pressure, unhealthy diet, changes in hormones, and educational attainment/cognitive activity level.
2 II. AD Behavioral Deficits During early AD, patients experience a loss of episodic, semantic and some implicit memory function (Becker et al ., 1996; Salmon & Fennema-Notestine, 1996; Greene & Hodges, 1996). A loss of episodic me mory appears similar to general amnesia so that patients are unable to learn new information; this denotes an impairment of storage and consolidation of new information (Becker et al., 1996). There is also an inability to shift information fr om short-term into long-term stores for later retrieval. Semantic memory is normally linked to longterm memory function of object, facts, concepts and words and their meanings, however in cases of early AD, there is a mild impairment in this type of memory function. A decrease in categor y fluency (generation of words from set categories, such as animals or fruits) is evident in early stages linked to a deficit of storage within the neural networks of the inferolateral areas of the temporal neocortex (Salmon & Fennema-Notestine, 1996). There is also an impairment of reading (comprehension of word meaning) due to the decrease in semantic memory function (Mathias, 1996). There are, however, no impa irments in the lexical routes of reading, such as phonemic or syntactic skills at this early stage. Deficits in visual attention, extrapyramidal signs (bradykinesea, rigid ity or tremors), ideomotor and ideational apraxias as well as nociceptive reflexes (snout and palmomental refl exes) are present in the early stages of AD (Coslett & Saffra n, 1996; Kidron & Freedman, 1996). Despite discrepancies in language studi es, language impairments have been shown to be evident in the early stages of AD (Becker et al., 1996). Most of the above mentioned deficits pr ogress as AD develops with addition of new cognitive impairments. From mild to moderate AD, 9-12% of subjects experience
3 hallucinations or delusions, with the range for severe AD between 40-60% (Schneider & Dagerman, 2004). It is proposed that these tw o symptoms of psychosis represent distinct subtypes of AD (Cook et al., 2003). AD subj ects with psychosis undergo a more rapid cognitive deterioration than those with no psychotic symptoms (Cook et al., 2003; Schneider & Dagerman, 2004). Psychotic sy mptoms, representing hallucinations and delusions, are mainly a product of a frontal lobe dysfunction (Schneider & Dagerman, 2004). However, degeneration of the majo r brainstem aminergic nuclei results in symptoms of depression for moderate AD persons, affecting between 30-50% of AD cases. These subjects also have a relative preservation of the cholinergic system in comparison to non-depressive AD cases (Zubenko et al., 2003). Also present in moderate AD is an impairment of divided attention wh ere attention can not be split between more than one task (Morris, 1996a). This is a ro le of executive function attributed to long-term memory where the deficit is nonglobal. There are also deficits in pyramidal signs related to changes in white matter (Kidron & Fr eedman, 1996). These include hyperactive jaw jerk reflex, extensor plantar responses, hyperre flexia and ankle clonus. Also present in about 10% of all moderate AD cases is moto r impersistence (inabi lity to sustain a voluntary movement) which is a deficit in sustained attention. More prominent in moderate AD is motor persistence (perseverati on of movement) which is associated with the frontal system (Kidron & Freedman, 1996). The late stages of AD sh ow progression of cognitive impairment from the moderate stage of AD as well as impairment in procedural memory (Becker et al., 1996). Almost complete loss of both episodic and sema ntic memory occurs in the late stages of AD. In addition, late stage AD subjects that are severely impaired exhibit more of a
4 continuing loss of cognitive function compar ed to AD subjects with less impairment (Auer et al., 1994). Also evid ent are severe impairments of executive function, praxis and visuospatial behaviors. Severe deficits in language functioning can be seen by this stage including loss of fluency, echolalia, pali lalia, non-verbal utterances and in some cases mutism (Boller et al., 2002). Seizures are also present in about 9.8-64% of all autopsy-verified late stage AD cases (Kid ron & Freedman, 1996). Some cases of AD (20-41%) progress to a bedridden state, mostly due to the incidence of contractures that prevents subjects from ambulatory behavior s (Souren et al., 1 995). Late stage AD patients can also exhibit a vegetative stat e, although because of differences in neurological evaluations there is a strong disagreement as to the incidence of a vegetative state in AD subjects (Volicer et al., 1997). Aside from add itional symptoms of late stage AD, death is more common in severe AD cases due to pneumonia compared to nondemented control subjects (Beard et al., 1996). However in a retrospective study by Kammoun et al. (2000), there we re no significant correlations between the incidence of AD and death due to bronchopneumonia compared to non-demented elderly control subjects; the authors did identify dementia as an underlying ca use of death. In contrast, Hui et al. (2003) did show a significant co rrelation between incr eased mortality with increased cognitive decline in AD subjects over a four-year period, but the authors did not state the cause of mortality. As previously mentioned a definitive diagnosis of AD is made only after death. This includes the presence of a large number of neurofibrillary tangle (NFT)-containing neurons and amyloid plaques within specifi c brain regions (Selkoe, 2001). Typically, NFTs are found primarily in neocortex and medial temporal structures including the
5 hippocampus, amygdala and parahippocampal gyr i. Subcortical regions such as the thalamus, mamillary bodies, nucleus basalis of Meynert, substantia nigra and locus coerulus also contain NFTs later in the diseas e process. In the AD brain, neuritic plaques appear first, and to their greatest exte nt in the cerebral co rtex and hippocampus; subcortical structures such as the putamen of the thalamus, the locus coerulus and the hypothalamus exhibit neuritic plaques to a le sser extent later in the disease process (Selkoe, 2001). III. AD Pathology Cerebral Atrophy, Neuronal Loss. Alzheimers disease results in an average of 7-8% decrease in total brain weight through atrophy (Morris, 1996a). Global atr ophy is characterized by widened sulci, narrowed convolutions, decreased white matter a nd ventricular enlargement, specifically the lateral ventricles. The greatest loss of brain weight occu rs in the cerebral hemispheres. Diffuse atrophy is associat ed with loss of cortex and hippocampal pyramidal neurons associated with the neocor tex. The presence of NFTs in the cortex and hippocampus may account for the substantial cell loss in the nuclei that project to the neocortex, such as cholinergic neurons of th e nucleus basalis of Meynert, noradrenergic neurons from the locus coerulus and serotonergic neurons of the raphe nuclei. There is a 40% decrease in the number of cells within the subiculum of the hippocampus (Morris, 1996c). Primary sensory and motor neurons, as well as thalamic projections experience minimal cell loss with AD. In normal aged in dividuals, there is no neuronal loss within layer II of the entorhinal cortex, with mini mal loss within layer III (Trillo & Gonzalo,
6 1992). By contrast, in mild to severe AD cases there is a 50-90% loss of cells within the entorhinal layer II (Mor rison & Hof, 1997). There is al so a substantial loss of neurons within region CA1 of the hippocampus in mild to severe AD cases (Price et al., 2001). In addition, Fukutani et al. (2000) suggest that ne uron loss in CA1 and entorhinal cortex is a direct result of neurofibrilla ry degeneration. Neuron loss within CA1 and subiculum of the hippocampus in severe AD has also been shown to correlate to formation of neurofibrillary tangles (Ro ssler et al., 2002). Specifically, pyramidal cells of the entorhinal cortex, CA1 and subiculum regi ons of the hippocampus are susceptible to neuronal degeneration and can lead to globa l disruption between association cortices (Morrison & Hof, 1997). In addition to neur onal loss in hippocampal regions, substantial loss occurs in the cortex. AD brains experien ce cortical atrophy in regions of the left hemisphere more than the right hemisphere (Gee et al., 2003; Thom pson et al., 2003). Affected regions include the anterior temporal, posterolate ral temporal and dorsolateral prefrontal regions (Gee et al., 2003), spari ng the neurons of the sensorimotor cortex, occipital poles and cerebellum (Karas et al ., 2003; Thompson et al., 2003). Overall, the total brain volume changes that occur with AD subjects are related to the accumulation of cortical NFTs (Silbert et al., 2003). Measurements of regional cerebral blood flow (rCBF) by single photon emission computed tomography (SPECT) and positr on emission tomography (PET) reveal decreases in brain functioning in vivo SPECT and PET scans show maximal changes within the parietal and temporal brain regi ons and are consistent with neurodegeneration (Morris, 1996c). Correlations can be found between rCBF and neuropsychological functioning so that language dys function correlates with d ecreased left parietal and
7 temporal lobe rCBF. In PET studies ther e is a significant de crease in the resting activation of the parietal lobes of AD cases compared to normal controls. The greatest reductions are within regions that are impa ired in early AD such as the middle and inferior temporal regions as well as angular and superior pa rietal gyri (M orris, 1996c). Implicated in the neurodegeneration of AD is the formation of amyloid plaques and development of intra-neuronal neurofib rillary tangles as discussed below. Neuritic Plaques. Amyloid precursor protein (APP) is a single transmembrane glycoprotein translated from chromosome 21. APP is a shor t-lived protein that is processed within the endoplasmic reticulum and subsequently modi fied through secretory pathways (Selkoe, 2001). Post-translational pro cessing of APP within the e ndoplasmic reticulum includes glycosylation and proteolytic cleavage by sp ecific enzymes (Selkoe, 2001; Turner et al., 2003b). APP is oriented so that the N-term inus is the extracellular domain and the Cterminus resides within the cytoplasm (Maccioni et al., 2001). There are three major isoforms of APP includi ng residues ending with 695, 751 and 770. APP695 mainly occupies neurons, while APP751 and 770 ar e found in both neurons and non-neurons throughout the body. These two isoforms also c ontain a Kunitz class of serine protease inhibitors, also known as a KPI domain. Cleav age of APP occurs so that the derivatives are released into vesicle lumens and the ex tracellular spaces. There are three cleavage sites along the APP peptide. The fi rst cleavage is provided by the -secretase enzyme and releases a large soluble fragment (C83) with the COOH-terminal fragment within the membrane. The second cleavage site results from -secretase to release a 99-residue
8 (C99) into the extracellular spa ce or vesicle lumen. Cleavage by -secretase of the C99 fragment results in a smaller peptide, -amyloid or A composed of exons 16 and 17 from the APP695 isoform (Selkoe, 2 001; Ling et al., 2003). Cleavage by -secretase of the C83 peptide results in a smaller fr agment named p3. APP is anterogradely transported to the axon terminals, the primary site for processing into A After APP is processed, left-over fragments of APP can undergo retrograde transport back to the cell body to be recycled. Recycling within the e ndosomes could lead to proteolytic cleavage by and -secretases to release the A peptide (Selkoe, 2001). Normal APP functioning includes cell adhesion, neuroprotection, neuropr oliferative activati on, intracellular and extracellular communication, as well as a cargo receptor fo r kinesin for microtubule transport and inhibition of seri ne proteases (Maccioni et al ., 2001; Turner et al., 2003b). Deletion of the APP gene does not lead to any significant deleterious effects in in vivo experiments; however, cells cultured from AP P deleted animals exhibit decreased neurite growth and nerve function (Selkoe, 2001). In 1983 Allsop and colleagues identified neuritic plaques as containing a 40-42 amino acid peptide, known as the -amyloid peptide (Turner et al., 2003b). A 40 and A 42 are both cleaved by si milar processes utilizing and -secretases from the amyloid precursor protein. Diffuse plaques from A PP cleavage result from accumulation of A 42; these plaques are modestly fibrillar and nontoxi c. Formation of diffuse plaques occurs mainly within association and limbic corti ces (Roher et al., 2000) Neuritic plaque formation originates with -helix structured monomers of either A 40 or A 42 peptides. Amyloidogenesis proceeds with deprotonati on of specific side chain residues and
9 protonation of additional residues to cause a destabilization of the -helix and conformational changes to a more stable A dimer with a -helical structure (Serpell, 2000). The dimer formation includes a hydrophobic core surrounded by hydrophilic residues. A ribbon-like protofilament is formed as a result of adjacent C-terminal binding between dimers to create a series of antiparallel -sheets. Pairs of these -sheets form a tubular cylinder within the center of each pr otofilament (Roher et al., 2000). Two types of protofilaments exist; the first includes a la rge diameter with periodic twists and the second type appears later in amyloidogenesis and forms no twists (S erpell, 2000). Fibril formation is a concentration dependant nucleation mechanism that utilizes 2-5 protofilaments. A hollow center is fo rmed by hydrophobic residues through an intertwined helical structure with the hydrophilic N-terminus on the outside of the fibril; after the fibril is formed the N-terminus is degraded. Compact (dense) neuritic plaques, from aggregation of A 40 onto A 42 fibrils, are formed mainly in neocortex and hippocampus, as well as within cortical and leptomeningeal vessels (Roher et al., 2000). Dystrophic neurites, found in and around f ilamentous plaques, contain enlarged lysosomes, numerous mitochondria and pa ired helical filaments (Selkoe, 2001). Compact plaque cores are formed with a strong -sheet conformation from simultaneous aggregation and disaggregation of A fibrils. Microglia are as sociated with the plaque perimeter and commonly express CD45 and HLADR antigens. Reactive astrocytes form a ring surrounding the outsid e of the plaque with abundant glial filaments. The presence of activated microglia and reactive astrocytes closely associated with mature amyloid plaques indicates that an inflammatory response has occurred to fibrillar A deposits (Selkoe, 2001). Os kar Fisher first proposed that an inflammatory
10 process occurred in conjunction with amyl oid fibrils in 1910. This proposal came without experimental evidence however it is now known that fibrillar A deposits are associated with a locally induced chronic infl ammatory response (Fig. 1). The absence of T-cells and immunoglobulins from the A plaque induced in flammation reaction suggests that this not a classical immune-m ediated response (Eikelenboom et al., 2002). The inflammation associated with AD is due to the activation of local microglia from accumulation of fibrillar A deposits. A accumulation leads to the increased production of free radicals from microglia and peroxidative injury of proteins and lipids that can lead to selective neuronal dysfunc tion and cell death (Fig. 1) (Selkoe, 2001). Activated microglia secrete cytokines (TNFand IL-1 ) and neurotrophi c factors (TGF1) as well as generate free radicals (NO and superoxide) and fatty acid metabolites (eicosanoids and quilolinic acid) (Liu & Hong, 2002; Cotman et al., 1996). Binding of complement factor C1q to fibrillar A deposits, with co-stimulation of serum amyloid P component (SAP) were shown to activate associ ated microglial clusters to initiate the classical complement cascade, as well as ge nerate production of C5 a (Eikelenboom et al., 2002; Cotman et al., 1996; Selkoe, 2001). A lthough microglia begin to appear during thioflavin positive staining, their major recr uitment occurs after the generation of a proinflammatory anaphylactic peptide, C5a, is activated by fibrillar A (Cotman et al., 1996). Reactive astrocytes are re cruited by activated microglia, apparently is response to microglial cytokine secretion. Astrocytes are also the source of two inflammatory proteins which directly cont ributes to the formation of amyloid plaques (Selkoe, 2001; Potter et al., 2001)-antichymotrypsin (ACT) and apoli poprotein E (ApoE) (Fig. 1). The overexpression of IL-1, ACT and ApoE in specific brain regions showing
11 neuropathology in AD-affected individuals suggests a region-specific inflammatory cascade for AD (Potter et al., 2001). Together ACT and ApoE have been described as amyloid promoters or pat hologic chaperones from in vitro and in vivo studies, supporting their role in the AD inflammatory cascade (Sanan et al., 1994; Wisniewski et al., 1994; Bales et al., 1999). In addition, increased levels of ACT are found in AD patients serum and CSF (Sun et al., 2003) and associated with severity of AD dementia (Dekosky et al., 2003). In order to in crease AD pathology, possi ble by inhibiting A degrading enzymes, the ACT/A signal peptid e variant increases th e amount of mature glycosylated ACT for secretion. The level and location of APP processing to increase toxic C-terminal fragments can also result fr om an affect from ApoE. Nilsson et al. (2004) conclude that ACT and ApoE can act sy nergistically or inde pendently to promote the development of mature amyloid plaques and diffuse A deposits without affecting monomeric A levels. In addition to the inflammatory respons e, amyloid deposits also lead to a condition known as cerebral amyloid angiopathy (CAA) (Fig. 1). CAA is a deposition of amyloid within meningeal and cerebral arteries arterioles, venules or capillaries and can occur in the absence of parenchymal A deposits (Jellinger, 2002; Selkoe, 2001). The primary A species present within blood vessels l eading to CAA was determined to be A 40; recently, it was discovered that A 42 may also be present in amyloid angiopathy (Selkoe, 2001). CAA increases with age and results in thickened vessel walls through A accumulation. Drainage channels and mi croglia take up the extracellular A and deposit it in the vascular lumen to aid in clearan ce. Hyaline necrosis surrounding the amyloid
12 deposits within the blood vessel wall can ru pture and lead to cer ebral hemorrhaging. Cerebral hemorrhages linked to CAA occur pr imarily within frontoparietal cortex and parietal cortex with lesser amounts within temporal and occipital cortices, basal ganglia and cerebellum (Jellinger, 2002). Most subjects with AD do not have cerebral hemorrhage despite the large amounts of amyl oid deposits within their blood vessels (Selkoe, 2001). The -amyloid peptide has many implications as previously described including formation of plaques, its role within the infl ammatory process, and it s link to CAA. In 1991, Hardy and Allsop first described the am yloid hypothesis. The hypothesis defined the root cause of AD as the overproduction and aggregation of A into senile plaques (Turner et al., 2003b). Multiple genetic risk fa ctors such as an increase in the number of ApoE4 alleles, PS1, PS2 and APP mutations (see following sections) all increase the process of A deposition, leading to an increased risk for developing AD. Recent research has lead to two changes for the in itial amyloid hypothesis; the first is that oligomers and fibrils, in a ddition to diffuse/compact A deposits, contribute to the neuronal/cognitive dysfunction of AD, and the sec ond is that changes in the activation of APP fragment molecules can also affect APP processing and A production (Turner et al., 2003b). Neurofibrillary Tangles. Neurofibrillary tangles are found within neurons of the hippocampus, entorhinal cortex, parahippocampal gyrus, amygdala and fr ontal, temporal, pari etal and occipital cortices as well as certain subcortical nuc lei (Selkoe, 2001; Maccioni et al., 2001). Some
13 transgenic mouse models suggest that A stimulates NFT formation as discussed later in the Animal Models section (Fig. 1). Tau is a microtubule-associ ated protein that normally binds tubulin to aid in assembly of microtubule formation and stabilization within neurons (Mudher & Lovestone, 2002). Phosphorylation of tau by protein kinases cdk5 or GSK3 causes it to dissociate from the microtubule and aggregate to form insoluble paired helical f ilaments (PHF) (Maccioni et al., 2001). Formation of PHF within neurons causes microtubule instabilit y such that microtubules are eventually replaced by tangles (Mudher & Lovestone, 2002). Some studies indicate that NFTs, not amyloid plaque burden, show a high correlati on with cognitive d ecline (Selkoe, 2001; Ohm et al., 2003). NFT formation occurs in brai n regions that are implicated in memory. Within the cortex, tangles first form in the la yer II of the entorhinal cortex prior to A plaques (Ohm et al., 2003). Aside from a regional specificity, there is a cellular specificity in NFT formation. In some rare cases of AD only a small portion of cells are prone to form NFTs. In such cases, corti cal pyramidal neurons can be tangle poor and instead can form Lewy bodies composed of -synuclein proteins. Despite the few numbers of tangles that are pr esent, these brain areas do el icit comparable amounts of A plaques in the hippocampus with less depositi on in the neocortex compared to AD brains that exhibit both NFTs and A plaques (Terry et al., 1987). A combination of A plaques and NFTs could account for synaptic terminal loss in AD. Synaptic loss occurs at the terminal dendrite segments mainly within neocortical association areas, hippocampus, fr ontal lobe and temporal lobe A decrease in synaptic density correlates highest w ith cognitive impairment (Coleman & Yao, 2003). Amyloid plaques have been found to alter synaptic inte grity, but not to cause the loss of synapses.
14 PHF within the neuron disrupts axonal transpor t and results in starvation of the synaptic terminal and the eventual loss of the synapse (Scheff & Price, 2003). Studies have yet to show a direct correlation betw een synapse loss and either A plaques or presence of NFTs.
15 Inflammatory Proteins; ACT & ApoE NFTs Free Radicals Oxidative Dama g e A Neuritic Plaques Neuronal and Synaptic loss (neurotoxicity) Cognitive Impairment Astrocyte Activation CAA Figure 1. Flow diagram showing AD pathogenesis progressing to cognitive impairment. Inflammatory Proteins; ACT &ApoE amyloid ( A & A ) Microglial Activation IL-1, TNF
16 IV. Genetics of AD Familial Alzheimers disease (FAD) is an autosomal dominant disease that accounts for 5-50% of all AD cases (Agnew, 1996; Selkoe, 2001). With little to no difference in phenotypic changes between FAD and sporadic AD (SAD) the main difference is the earlier age of onset for FAD (e.g. before 60 years of age), which has been linked to point mutati ons on chromosomes 1, 14 or 21. The presenilin 2 gene from chromosome 1 can be affected by nine different mutations that increase the ri sk of developing AD; while chromosome 14 codes for the presenilin 1 (PS1) gene that can contai n over 100 mutations (Marcon et al., 2004). Mutations of PS1 decrease the age of AD ons et to the early 40s and 50s through a selective increase in -secretase cleavage of peptides C 99 and C83; this mainly leads to an increase of A 42 (Selkoe, 2001). Early-onset AD, from a PS1 deletion mutation, can result in verbal and visual memory impairme nt in addition to defi cits in intellectual functions. These types of impairments were related to temporoparietal hypometabolism that is typically found in AD subjects (Verkkoniemi et al., 2004) One specific type of PS1 mutation, E280A, associates neuronal loss in the CA1 region and epileptic seizures of subjects with FAD (Velez-Pardo et al., 2004 ). The same mutation has been shown to cause an increase of A 42, not A 40, in FAD brains suggesting that PS1 can alter the cleavage of APP at the C-termin al end (Lemere et al., 1996). AD risk can also be affected by muta tions of the APP gene residence on chromosome 21. Downs syndrome patients (trisomy 21) have an overproduction of A 40 and A 42 as well as development of diffuse plaques by age 12 and neuritic dystrophy with NFTs in the 20s and 30s (Sel koe, 2001). The formation of NFTs in
17 Downs syndrome patients most likely occurs after the accumulation of extracellular A deposits and inflammatory response (Mori et al., 2002). Specifically, Hirayama et al. (2003) proposed that NFTs form in Downs syndrome patients as a result of an axonal flow disturbance caused by the accumulation of A 43 within neurons of the cortex. Although most SAD and FAD cases have simila r phenotypes, APP mutations can lead to an increased inciden ce of myoclonus, seizures and extr apyramidal signs (Agnew, 1996). There are up to 9 missense mutations that occur on the APP gene which facilitate cleavage to increase production of A 40 and A 42. A point mutation at APP717 is referred to as the London mutation and t ypically results in an increase of A 42 (Selkoe, 2001). This mutation could result in a switch from valine to is oleucine in which case the onset occurs at approximately 50 years of age. Additional mutations from valine to glycine or valine to phenylalanine decrease th e age of onset to w ithin the 40s (Agnew, 1996). A double mutation at APP670/671 (Swedi sh mutation) increases the production of both A 40 and A 42 through increased -secretase activation (Selkoe, 2001). The production of A from the Swedish mutation is a result of competition between and secretase upon APP within Golgi-derived ve sicles (Haas et al., 1995). Human subjects who possess the Swedish mutation form of AD exhibit reduced glucose metabolism within the temporal lobe which precedes cognitive impair ment (Wahlund et al., 1999), and reduced regional cerebral blood flow within basal and lateral temporal lobes (Julin et al., 1998). Overall, although the SAD and a bove-mentioned FAD cases differ mainly in age of onset, there are also subtle differenc es in pathology as the disease progresses. Apolipoprotein E (ApoE), which is involve d in lipid trafficking, expresses three different alleles; ApoE2 is the least common, ApoE3 is the most common with
18 inheritance of ApoE4 giving ri se to an increased risk for developing AD. Chromosome 19 contains the ApoE4 allele, which is over expressed in many AD cases. Inheriting 1 or 2 copies of the ApoE4 allele increases the ri sk for developing late onset AD in the 60s and 70s (Agnew, 1996; Selkoe, 2001). Studies show that the ApoE4 protein results in a higher A plaque burden although there is no change in the amount of NFTs present. An increase in the number of A fibrils could occur via a decr ease of clearance or increased deposition of A 40 within the cerebral cortex an d microvasculature (Selkoe, 2001). ApoE4 and ACT together promote depositi on of monomeric and/or oligomeric A filaments (Potter et al., 2001), in essence by act ing as pathologic chaperones in binding to A filaments. Glckner et al (1999) proposes that the acti on of ApoE primarily occurs at the early stages of AD. They found changes in hippocampal ApoE expression only within the brains of subjec ts during the beginning stage of AD (Glckner et al., 1999). V. Diagnosis of AD The diagnosis of Alzheimers disease occu rs with complete assurance only after death by the presence of high numbers of am yloid plaques and NFTs within the brain. Prior to death, there are multiple cognitive ex ams designed to predict the presence of a cognitive impairment as an indicator for the development of AD. A diagnosis of AD follows three stages, the first is an initial screening into the patients personal history; this establishes the subtype of the disease. The second stage involves a formal neuropsychological exam which determines th e provisional diagnosis of the specific conditions of AD. The third stage includes a neurological exam and CT scan in order to exclude alternative forms of dementia not re lated to AD (Gray & Sala, 1996). Some of
19 the neurological screening methods used in clude the Mini-Mental State Exam (MMSE), which results in numbered scores administered to the patients. A score ranging from 0 to 17 implies severe cognitive impairment, wh ereas a score between 18 and 23 indicates mild to moderate impairment. Lastly, a score from 24 to 30 denotes an uncertain or absent degree of cognitive impairment (Crum et al., 1993). Additional cognitive exams include a brief telephone questionnaire or a pape r and pencil skills test to screen for early signs of impairment, as well as multiple verbal memory tests. Some of these exams include the Cognitive Capacity Screening Examination, the Mental Status Questionnaire, the Short Portable Mental Status Questionna ire and the Dementia Rating Scale (Gray & Sala, 1996). Neurological screening exams are ab le to detect specific changes in memory function from early to moderate to severe forms of AD. Additional screening methods include computed tomography (CT), magnetic resonance imaging (MRI), magnetic resona nce spectroscopy (MRS ), functional MRI (fMRI), single photon emission computed tomography (SPECT) and positron emission tomography (PET). CT and MRI scans can detect anatomic changes within the brain, whereas MRS imaging measures metabolites as an indicator of brain function. For example, N -acetylaspartate (NAA) is a metabolite f ound within axons, choline (Cho) is in the myelin membrane and creatine (Cr) is involved in neuronal metabolic activity (Lee et al., 2003). PET and SPECT are additional form s of functional neuroimaging for glucose metabolism that can be measured within the brain. In early AD, prior to cognitive decline, CT and MRI scans show cerebral atrophy and ventricular enlargement, including a progr essive increase of co rtical subarachnoid spaces. Some studies using CT scans show at rophy of the medial temporal lobes in AD
20 and volume loss of the hippocampus with MRI compared to non-demented controls (Lee et al., 2003). Glucose metabolic PET im aging with fluorodeoxyglucose (FDG) has the potential to detect early neoc ortical dysfunction before neur opsychological testing reveals dementia. As questionable dementia progresses toward mild AD, PET imaging detects abnormalities in the parietal lobe which exhi bit asymmetrical differences in metabolism (Demetriades, 2002). PET imaging also show s a metabolic decrease in the posterior cingulate cortex, cinguloparietal transitional cortex, middle temporal cortex, inferior temporal cortex as well as angular and superior parietal gyri in early AD compared to age-matched controls (Demetriades, 2002). PET imaging can also detect amyloid deposition with labeled probes in early AD and presumably before cognitive symptoms are present. Multiple compounds have b een tested as possible probes to bind A within the brain such as a thioflavin-T analogue (Mathis et al., 2002), (18)F-labeled IMPY derivative (Cai et al., 2004), BF-108 (Suemoto et al., 2004), stilbene derivatives (Ono et al., 2003) and the Pittsburgh Co mpound-B (Klunk et al., 2004). Mathis et al. (2002) found that the thioflavin-T analogue, [(11)C] -labeled 6, has a high uptake within the brain of transgenic mice and resulted in la beled cerebral plaques and cerebrovascular amyloid deposits. The BF-108 compound wa s found to label senile plaques, neurofibrillary tangle s and vessels laden with amyloi d within temporal cortex and hippocampus of AD patients. BF-108 also labele d plaques within the brains of transgenic mice (Suemoto et al., 2004). In additi on, the Pittsburgh Compound-B (PIB) had a high binding affinity within association cortex, mo st prominently within frontal, parietal and occipital cortices as well as the striatum. In areas such as the pons and cerebellum where there is minimal amyloid pathology, there was similar binding to that of non-AD control
21 subjects. The same study also found an inve rse correlation between PIB retention within cortex and cerebral glucose me tabolism (Klunk et al., 2004). Overall, the use of labeled probes by PET scanning to detect A within the brain prior to cognitive impairment or in early AD is a useful tool for diagnosis a nd shows high affinity for amyloid plaques. In moderate AD, MRS detects a decrease in NAA within temporoparietal, hippocampal and parahippocampal regions. A decrease in the NAA/Cr ratio of the occipital lobes confers a decr ease in the involvement of functional neurons; a decrease within medial temporal lobes corresponds to cognitive decline (Lee et al., 2003). PET imaging studies show a global decrease in ce rebral glucose metabolism, specifically the neocortical structures includi ng the parietal, frontal and pos terior temporal association cortices (Demetriades, 2002). A decreased me tabolism of the parietot emporal cortex is a possible diagnosis for AD because of its high sensitivity and specificity to distinguish AD from other dementias. A substantial decrease is present in AD subjects of the resting activation of parietal lobes with the most am ount of decrease within the occipital cortex (Morris, 1996c). PET studies also exhib it a hypometabolism within primary cortices, medial temporal cortex and posterior cortical regions (D emetriades, 2002). A bilateral hypometabolism of temporoparietal associa tion cortex is present in AD and other dementias (Demetriades, 2002; Lee et al., 2003). PETs reveal a preservation of primary sensorimotor and visual cortices, cere bellum and striatum (Demetriades, 2002). Together, PET and SPECT studies show ma ximal changes within the parietal and temporal regions with a negative correlati on between regional cerebral blood flow and neuropsychological functioning (Morris, 1996c ). Language dysfunctions have been associated with left parietal temporal change s, whereas praxis disturbances correlate to
22 decreased regional cerebral blood flow within bilateral and temporal lobes (Morris, 1996b). In more severe cases of AD, there is a decreased metabolism within the frontal cortex (Demetriades, 2002; Lee et al., 2003). For cases of FAD where subjects ar e homozygous for ApoE4, there is a significant decrease in the metabolic rate of glucose within posterior cingulate, prefrontal, temporal and parietal regions (Demetriades, 2002). In the same subjects prior to cognitive decline there is si gnificant hypometabolism within the temporoparietal region (Lee et al., 2003). PET imaging of bitempor al and biparietal regions in FAD cases involving mutations of the APP gene show decreases in metabolism compared to agematched controls (Demetriades, 2002). Additional biomarkers for diagnosis of AD include cerebrospinal fluid (CSF) testing for the presence of total tau (t -tau), phosphorylated tau (p-tau) and A 1-42. Tau is a protein found within neuronal axons that has six different isoforms and 21 possible phosphorylation sites. The t-tau found within CSF of patients with MCI that progressed to AD is significantly increased compared to those with MCI th at did not develop AD (Blennow & Hampel, 2003). The detection of t-tau has a sensitivity of 90% and specificity for distinguishing MCI from non-demented subjects of 100% (Blennow & Hampel, 2003; Hampel et al., 2003). Disti nguishing early AD from normal aging using ttau from the CSF reveals a sensitivity of 85% and a specificity of 75% (Hampel et al., 2003). In addition to Alzheimers disease, incr eased CSF levels of t-tau are also found in other dementias including vascular deme ntia, frontotemporal dementia, lewy body dementia and semantic dementia (Blennow & Vanmachelen, 2003; Hampel et al., 2003). There are also studies show ing a significant correlation between t-tau and cognitive
23 decline (Kanai et al., 1998; Nishimura et al., 1998), while additional studies show no correlation (Arai et al., 1995; Vigo-Pelfre y et al., 1995; Hulsta ert et al., 1999). Levels of p-tau are used to disti nguish AD from control subjects through monoclonal antibodies specific to different epitopes. Overall, there is a mean sensitivity of 70% and a mean specificity of 94% with a mean increase in p-tau of 250% compared to age-matched controls (Blennow & Vanmach elen, 2003). At baseline, MCI patients have an increased level of p-tau versus cont rol subjects (Buerger et al., 2002a; Andreasen et al., 2003). With progression from MCI to AD, levels of ptau correlate with cognitive decline (Buerger et al., 2002a). Because levels of p-tau can distinguish MCI from nondemented control subjects, using CSF p-tau levels to diagnose AD prior to clinical dementia would improve treatment methods. C SF p-tau levels have also been inversely correlated with the age of onset for SAD and FAD subjects (Thake r et al., 2003) and positively correlated with the intensity of disease progression marked by a presence of ventricular widening during AD advancemen t (Wahlund & Blennow, 2003). CSF levels of p-tau correspond to the levels of p-tau with in the brain and forma tion of tangles in AD; there are, however lower measurable levels of p-tau in vascular dementia, frontotemporal dementia and lewy body dementia as compared to levels present in AD (Blennow & Hampel, 2003). Using p-tau to accurately distinguish AD from ot her types of dementia is challenging. However, AD can be signifi cantly distinguished from frontotemporal dementia by increased CSF p-tau of AD subj ects (Buerger et al., 2002b). Hampel et al. (2004) also found that AD can significantly be distinguished from dementia with Lewy bodies by measuring CSF levels of p-tau. They conclude that because there is significant difference of CSF p-tau between two types of similar dementias, that p-tau must have
24 different roles within AD as compared to dementia with Lewy bodies (Hampel et al., 2004). Overall, there is a decrea se in the level of CSF A 1-42 in AD as detected by ELISA with specific antibodies to A As previously discussed APP results in A by cleavage first by -secretase releasing the C-terminus of APP into the extracellular space followed by cleavage by -secretase to release monomeric A Because of the close association between the CNS and CSF, A can flow into the CSF and levels can be measured for diagnosis. The mean sensitivity and specificity for A 1-42 is 85% for distinguishing AD from age-matched contro ls (Blennow & Vanmachelen, 2003). There is also a strong association be tween decreased levels of A 1-42 in CSF and increased number of plaques within hippocampus a nd neocortex (Blennow & Hampel, 2003). Some studies suggest that th e decreased levels of CSF A are due to increased trapping of brain-produced A into amyloid plaques, although in Creutzfeldt-Jakob disease there are no deposited plaques desp ite a decrease in CSF A (Hampel et al., 2003). Another indicator for developing AD is the inheritance of mutated forms of APP, PS2 or PS1 genes. APP, primarily exons 16 and 17, and PS2 mutations are detected from genomic DNA analysis of peripheral blood leukocytes amplified by polymerase chain reaction (PCR) and sequenced to give the ex act gene. The PS1 mutations are commonly detected by extraction of total RNA from peripheral blood leukocytes for DNA amplification by a reverse transcriptase-PCR (R T-PCR) protocol. The entire PS1 gene is then sequenced to identify any possible mutations.
25 FAD from mutations of the APP gene ar e not prominently used as diagnostic markers for the detection of early-onset AD. Zekanowski et al. (2003) verified the presence of two APP mutations, T714A and V715A, in a population of patients with early-onset AD from Poland. They averaged the age of onset from either mutation around 55 years of age (Zekanowski et al., 2003). A second study found additional APP mutations in non-demented carriers, KN670/ 671ML and E693G (Almkvist et al., 2003). These APP mutant carriers exhibited no clini cal abnormalities compared to non-carriers (Almkvist et al., 2003). Mutations from the PS2 gene are also not major contributors to FAD. However, one mutation of PS2, Q228L decreases the age of onset to around 55 years of age (Zekanowski et al., 2003). Muta tions of the PS1 gene are more common and carriers exhibit much of the same pathology from SAD subjects. This includes evidence of neuritic plaques, amyloid angiopathy, NF Ts (Takao et al., 2002), diffuse plaques (Larner & du Plessis, 2003) and cortical de generation mainly within the primary motor cortex (Miklossy et al., 2003). The averag e age of onset for some of the many PS1 mutations range from 25 years (Miklossy et al., 2003) to around 55 years (Zekanowski et al., 2003). VI. Risk Factors for AD The primary and most significant genetic ri sk factor for the de velopment of lateonset AD is inheritance of at least one copy of the ApoE4 allele. Each ApoE4 allele inherited increases risk and d ecreases the age of onset for developing AD. About 80% of all FAD cases and 64% of SAD cases contain at least one copy of the ApoE4 allele,
26 whereas it is present in only 31% of cont rol subjects (Corder et al., 1993). ApoE is synthesized within the brain a nd secreted by astrocytes. It plays a major role in lipid transport through the CNS (Olsen, 1998). Ho wever, the primary mechanism that ApoE4 affects to increase risk of AD is not th rough lipid transport, but most probably by affecting A deposition. The binding of ApoE4 to A 42 increases A deposition (Olsen, 1998) and density of neuritic plaques (Strittmatter & Roses, 1995; Nilsson et al., 2004). PDAPP transgenic mice, homozygous for murine ApoE4 had increased diffuse A deposition within cerebral cortex compared to PDAPP ApoE knock-out (ApoE-KO) mice (Nilsson et al., 2004). The PDAPP + ApoE transgenic mice also had increased deposition of congophilic-positive compact A plaques within cerebral cortex and hippocampus versus ApoE-KO mice. Together with the plaque pathology, ApoE mice exhibited significant memory deficits ev ident by spatial learning impairments and working memory impairments illustrated by th e Morris Water maze and radial arm water maze tasks (Nilsson et al., 2004). A second significant risk factor for the development of AD is high blood cholesterol levels. Increased total serum cholesterol level in midand late-life is positively associated with carrying the ApoE4 al lele (Kivipelto et al., 2002). Accounting for this factor, a high level of total serum chol esterol is still a significant risk factor for developing MCI and AD (Kivipelto et al., 2002 ; Reitz et al., 2004). The mechanism of how cholesterol levels affect development of AD is unknown; however it is thought that A production is indirectly increased by se rum cholesterol levels, perhaps through stimulation of -secretase processing of APP. Refolo et al. (2000) first determined that a high cholesterol diet results in increased amyloid depos ition in the brains of AD
27 transgenic mice. Specifically, th is study found increased amounts of -C terminal fragments from APP cleavage resulting from -secretase activity. Inversely, the increased -secretase activity most likely resulted in a decrease of -secretase activity which also led to the increased deposition of A Simons et al. ( 1998) also supported the theory that cholesterol affects -secretase activity in cultured hippocampal neurons. In contradiction to the above studies, Burns et al. (2003) suggest that -secretase activity is highly affected by cholesterol accumulation in a mouse model of Niemann-Pick type C disease to increas e production of A 40 and A 42. Also, Wahrle et al. (2002) found an inhibition of -secretase activity in cultured brain cells with cholesterol depletion suggesting that -secretase activity is cholesteroldependent. Also, treatments with cholesterol-lowering drugs as discussed in th e next section have been shown to inhibit A production through stimulation of -secretase and decrease AD risk. In addition to the affects of increased fat and decreased fish consumption on AD risk, reduced dietary intake of fruits and vegetables also cont ribute to the risk factor. To date, there have been no clin ical longitudinal studies perf ormed that administered DHA and/or EPA to test their protective effects against AD development; however some studies have administered fish oil supplements or a combination of omega-3 and omega-6 fatty acids to test their effects. Because of this, it is not known whether DHA and EPA, specifically, are conferring protection against dementia, or whether it is some additional aspect within the fish oil that results in bene ficial effects. Specifically, clinical studies have shown that a high intake of fish, hi gh in the polyunsaturated fatty acids DHA and EPA, is associated with a reduced risk for developing cognitive impairment, particularly
28 AD (Kalmijn et al., 1997a; Kalmijn et al., 1997b, Kalmijn et al., 2004). Kalmijn et al. (1997a) reported that an increased intake of to tal fat, saturated fat a nd cholesterol were all significantly associated with an increased risk for developing dementia, whereas increased fish consumption was the major factor associated with a reduced risk. Kalmijn et al. (1997b) also showed a significant corr elation between increas ed fish consumption and decreased risk for developing dementia The elderly subjects that developed dementia over the 8-year study were shown to have a higher intake of total fat, and the omega-6 fatty acid linoleic acid (LA) with a lower intake of fish, docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) compar ed to those subjects that did not develop dementia. In contrast to the above studies, most recently Morris et al. (2005) reported a weak association between dietary intake of fish and rate of age-related cognitive decline. Instead, the authors suggested that overall fat consumption led to a greater risk of dementia. An additional study by Morris et al. (2003) also suggested that other fats play the primary role predicting the risk associated with developing AD. This study reported that saturated and trans -unsaturated fats were associat ed with developing AD, whereas n3 fatty acids did not contribute or protect against risk of developing AD (Morris et al., 2003). Aside from the possible involvement of nutri ents from fish, dietary vitamin intake plays a role in risk for developing AD. A re duced dietary intake of vitamin B leads to increased plasma homocysteine levels (Selhub & Miller, 1992) and accumulation of DNA damage that contributes to AD risk (K ruman et al., 2002). Memory function is positively associated with intake of vitamins E, C and A (Solfrizzi et al., 2003; Meydani, 2003). Specifically, supplementation with vitamin E and vitamin C resulted in a
29 decreased risk of age-associated diseases including cognitive impairment (Meydani, 2003) and dementia (Haan & Wallace, 2004). Although some clinical studies have reported that vitamins E and C have no prot ective benefits agains t development of AD (Masaki et al., 2000; Laurin et al., 2004), additional clinical studies show a significant reduction of risk for developing AD with s upplementation/intake of vitamins E and C (Morris et al., 1998; Engelhart et al., 2002b; Zandi et al., 2004). Animal studies involving non-transgenic rats (Joseph et al., 1998; Joseph et al., 1999) and mice (ShukittHale et al., 1999) show improved cognitive pe rformance with supplementation of various fruits and vegetables such as strawberries, blueberries or spinach. Supplementation with apple juice has also been shown to improve cognition and decreased oxidative stress in mice (Rogers et al., 2004). Overall, sporadic AD risk factors prim arily include genetic predisposition (ApoE4 allele), high blood choles terol (LDL) levels, high dietary intake of fats, and low dietary intact of fish, fruits and vegetables. Because SAD can not be accurately predicted, it is probably a combinati on of the above mentioned risk factors that all contribute to the development of AD. VII. Treatment of AD All of the US Food and Drug Administra tion approved treatments for Alzheimers disease are aimed at providing relief from symptoms associat ed with cognitive decline. Current treatments include acetylcholine es terase inhibitors: d onepezil (Aricept), galantamine (Reminyl), rivastigmine (E xelon), as well as the NMDA antagonist, memantine. Additional treatments have been tested in animal models, in vitro and
30 clinical trials including anti-inflamma tory drugs, antioxi dants and behavior modifications. Tacrine was the first acetylcholine este rase inhibitor (AChEI) approved for AD symptoms. In a clinical tria l tacrine relieved symptoms of hallucinations, anxiety, apathy and disinhibition in individuals with moderate dementia (K aufer et al., 1996). A second clinical trial found improvements in la nguage production, comprehension, word recognition and delusions in AD subjects with mi ld to moderate dementia (Raskind et al., 1997). Donepezil was approved as a noncompe titive and reversible AChEI for mild to moderate AD symptoms. It is metabolized in the liver by cytochrome P450 enzymes and uridine-diphosphate glucuronosyl transferase. Because of its long half-life (~ 70 hours) only one daily dose is needed (Scarpini et al., 2003). A clinical tria l for 1 year including mild to moderate AD patients resulted in im proved cognition and activities of daily living versus placebo-treated patients (Winblad et al., 2001). An additional study of moderate to severe AD subjects for 24 weeks showed that Donepezil improved cognition and decreased delusions, anxiety, disi nhibition and irritability in these patients compared to placebo-treated subjects (Feldman et al., 2001). Galantamine was approved as a selective and reversible AChEI. It also functions as an allosteric ligand at nicotinic ACh receptors in order to increase presynaptic release of ACh and postsynaptic neurotransmission within pyramidal neurons. Galantamine is al so metabolized in the liver by cytochrome P450 enzymes which results in a 5 hour half -life and therefore re quires two daily doses for treatment (Scarpini et al ., 2003). A clinical trial ut ilizing galantamine found it to decrease symptoms of anxiety, disinhibition, hallucinations and aber rant motor behaviors associated with mild to moderate AD, as we ll as improve cognition (Tariot et al., 2000).
31 Some therapies comparing the acetylcholine esterase inhibitors found differences in cognitive outcome. Jones et al. (2004) f ound a greater cognitive improvement with the use of donepezil versus galantamine whereas Wilcock et al. (2003) had opposite effects resulting in a significant c ognitive effect of galantamine over donepezil. Despite differences from other studies, six months of treatment by the Cochrane Dementia Group with donepezil (Birks et al., 2002a), galantamine (Birks et al., 2002b) or rivastigmine (Olin et al., 2002) resulted in similar effects for all three drugs. Also, a similar study by Wilkinson et al. (2002) found no cogn itive difference between donepezil and rivastigmine as treatments for patients with mild to moderate AD. All three acetylcholine esterase inhibitors show a st abilization of cognition through 6 months of treatment and as long as 36 months of treatment. At 6 m onths, donepezil (Tariot et al., 2001) and galantamine (Raskind et al., 2000; Patterson et al., 2004) were both found to maintain the improved cognitive performance in comparison to placebo-treated subjects. Following 12 months of treatment with galantamine (R askind et al., 2000; Kurz et al., 2003) or rivastigmine (Doraiswamy et al., 2002) cognitive performance was maintained as compared to improved cognition after 6 months of treatment. Additional studies show maintenance of cognitive improvement through 18 months of treatment with donepezil (Matthews et al., 2000), 24 months of treatme nt with rivastigmine (Grossberg et al., 2004) and even 36 months of treatment w ith galantamine (Raskind et al., 2004). Additional treatments incl ude the use of statins and non steroidal antiinflammatory drugs (NSAIDs). Statin use ha s been associated with a 39% (Zamrini et al., 2004) 70% (Jick et al., 2000) decrease in the risk of developing AD. The mechanism by which statins decrease AD risk might be through a shift from and -
32 secretases to -secretase to decrease A production (DeKosky, 2003). In vitro studies show that -secretase functions best in a high chol esterol environment so that a decrease in cholesterol leads to a decr ease in BACE function. Statin therapy has been shown to lower cholesterol in cultu red cells by blocking 3-hydr oxy-3-methylglutaryl-CoA reductase (HMG-CoA reductase) and thereby inhibiting A production (Simons et al., 1998; Fassbender et al., 2001). Guinea pigs fe d a diet supplemented with simvastatin showed a strong decrease of cerebral A 40 and A 42 levels from CSF and brain (Fassbender et al., 2001). AD transgenic mice fed a high chol esterol diet had increased -amyloid deposition within the brains (Refol o et al., 2000). Add itional mechanisms of statins may be that they remove A from the brain or have a direct anti-inflammatory effect on AD through undetermined pathways (DeKosky, 2003). Because statins do not cross the blood-brain barrier their effects must be from a peripheral pathway that is not yet determined. NSAIDs have been proposed to inhibit COX and therefore decrease the synthesis of prostaglandins, specifically PGE2 and PGI2 which are pro-inflammatory (Aisen, 2002a). Prostaglandins may potentiate glutamatergic transmission through NMDA receptors by inhibiting the reuptake of glutam ate by astrocytes which leads to an excess of glutamate in the extracellu lar spaces (in t Veld et al ., 2002). In AD there is an increase of extracellular glutamate which leads to an increased activation of NMDA receptors and therefore an increase in intracel lular calcium levels. The increased calcium could lead to neurotoxicity and neuronal cell death (Scarpini et al., 2003). The use of NSAIDs could inhibit this pa thway. However, numerous studies have suggested that NSAIDs exert their protective effect indepe ndent of COX inhibiti on, by acting to inhibit
33 -secretase or its substrate. In vitro the addition of ibuprofe n, indomethacin or sulindac decreased the release of A 42 within cultured cells that ov er express APP (Sascha et al., 2001). Along this line, a study involving cultur ed cells treated with either ibuprofen, indomethacin or sulindac sulphide found reduced A 42 peptide independent of COX activity, most likely due to selective activity from -secretase to alter A production (Weggen et al., 2001). Additional work by We ggen and colleagues (2003) and Eriksen et al. (2003) also support the fi ndings that NSAID activity can decrease levels of A 42 within cultured cell lines and AD transgen ic mouse models independent of COX inhibition. Weggen et al (2001) found a significant reduction of A 42 within the brains of APP transgenic mice with no change in A 40 following treatment with ibuprofen. The authors suggest that this difference in A production is due a selective shift of -secretase activity. Similarly, administration of ibuprof en to APP transgenic mice resulted in a lower brain A 42 levels (Sascha et al., 2001). NSAI Ds could also inhibit inflammation by suppressing the expression of specific proi nflammatory genes or by directly blocking the induction of interl eukins 1 (IL-1), IL-1 and possibly IL-6 (in t Veld et al., 2002). A six month oral supplementation of ibuprof en to AD transgenic mice resulted in a significant decrease in A brain burden and microglial response (Lim et al., 2000). In addition, the NSAID NCX2216 has been shown to decrease A plaques in APP/PS1 mice more than ibuprofen or celecoxi b (Jantzen et al., 2002). Treatment with some NSAIDs resulted in no improvement in cognition in humans. A study by Van Gool et al. (2001) involving early AD subjects showed no cognitive or behavioral ch anges using the NSAID hydroxychl oroquine versus the placebo
34 group. An additional study administering nimesulide to subjects with probable AD resulted in no significant effect on measures of cognition, activitie s of daily living or behavioral aspects compared to the placebo su bjects (Aisen et al., 2002b). Aisen et al. (2003) also found no benefit to naproxen treatment with mild to moderate AD patients. A clinical trial for mild to moderate AD patients using indomethacin showed improved cognition with the drug compared to a decline of cognition in the pl acebo-treated group. However, indomethacin use had a low toleran ce illustrated by adverse reactions to the drug and thus far there are no additional studies utilizing it (R ogers et al., 1993). Prednisone treatment to AD s ubjects also resulted in negative effects and decline in behavior with no change in cognition, after a 4-week period compar ed to placebo-treated subjects (Aisen et al., 2000). In addition to anti-inflammatory drugs, antio xidants have been used as therapeutic treatment for AD. The aggregation of A during AD progression can directly and indirectly induce oxidative st ress. Regarding indirect i nduction of oxidati ve stress by A the interaction of A with the receptor for advanced glycation end products (RAGE) in neuronal cell lines induces an increase of lipid peroxidati on (Yan et al., 1998; GasicMilenkovic et al., 2003). Al so, the interaction of A with endothelial cells or blood vessels results in an increase in superoxide radicals leading to oxi dative stress in the vasculature (Thomas et al., 1996). An additiona l mechanism for the indirect formation of oxidative stress by A is through macrophages. Work by Klegeris et al. (1994) and El Khoury et al. (1996) both demonstr ate the association between A and macrophages to increase reactive oxygen species (ROS) such as nitric oxide. Re garding more direct actions, A 25-35 has been shown to directly ge nerate ROS in the presence of O2 (Hensley
35 et al., 1994). In additi on, the aggregation of A 1-40 can lead to generati on of free radicals that induce neurodegeneration through neurona l insults, such as lipoperoxidation or damage to intracellular protei ns (Harris et al., 1995). Ginkgo biloba has been defined as contai ning antioxidant and anti-inflammatory properties as well as be ing a memory enhancer. In a clin ical trial involving patients with mild to severe AD there was a stabilization of cognitive and social function with 6 months to 1 year of Ginkgo biloba extract tr eatment (Le Bars et al., 1997). Mild to severe AD subjects showed improvements in cognition and daily behaviors following treatment from 3 months (Maurer et al., 1997) to 1 year (Le Bars et al., 2002) with high or low doses of the extract, EGb 761. In cont rast to the above studies, mild to moderate AD patients supplemented for 24 weeks with high or low doses of EGb 761 did not show improved cognitive function (van Dongen et al., 2000; van Dongen et al., 2003). Other antioxidants, vitamins E and C, exert neur onal protection against oxidative damage and cell death. A Rotterdam study found a 34 43% decreased AD risk with supplements of vitamins C and E, respectively (Enge lhart et al., 2002b). Vitamin E, -tocopherol, has been associated with a decreased risk of AD in patients that were not carrying the ApoE allele (Morris et al., 2002). -Tocopherol acts as an an tioxidant by inhibiting lipid peroxidation within membranes as a free radi cal oxygen scavenger. It is the primary antioxidant candidate for AD and exerts signi ficant beneficial e ffects on functional decline for patients with moderate to severe AD (Sano et al., 1997). Aside from drugs and supplements, nonpharmacologic AD treatments include multiple types of behavioral modification. Fe w studies that try to improve cognitive and non-cognitive behaviors have been performe d on AD patients. In 1989, Quayhagen &
36 Quayhagen administered a cognitive stimula tion protocol including memory exercises, communication skills and problem -solving activities to AD patie nts. This resulted in maintenance of cognitive function across the treatment period, whereas the group that did not receive the treatment declined in cogniti on. More recently Sobel et al. (2001) tested the effect of the game of Bingo on subject s with AD versus AD patients that underwent daily physical activity. The AD subjects that played Bingo had improved performance of short-term memory, word retrieval and word recognition tasks. Th ese alternatives can result in an increase of self-care skil ls as well as appropr iate ambulation and socialization. Additional types of be havior modification include sensory and environmental intervention, stru ctured activities and behavi or therapy (Cohen-Mansfield, 2001). VIII. Animal Models Animal models have been used to aid in determining the functions of genes/proteins that were isolated from in vitro studies or to study the phenotypic results from gene changes, such as relocation of developmental transcription factors. In addition, transgenic animals from an overexpr ession or mutation in single or multiple genes are used to model human diseases a nd can therefore facilit ate in determining possible preventions or treatments for such diseases. Most Alzheimer disease studies use mice as transgenic models. Even with th e longer gestation and developmental period compared to invertebrates ( Drosophila & C. elegans ) commonly used as research models, physiologically mice are more closely related to humans and therefore provide a better foundation for a research model. Rats cannot generally be used to model human
37 disease and its effect on learni ng because of a lack of gene ration of developmental stem cell lines needed to generate rats lacking/over expressing a specific gene (Piccioto & Wickman, 1998). Due to this, mice are the most common and advantageous model used to study human neurodegenerative diseas es such as Alzheimers disease. The generation of an AD transgenic anim al requires the selection of the specific transgene cDNA to be transcribed downstream of a certain promoter sequence. The transgene DNA includes 1) the specific cDNA se quence that is to be over expressed, 2) the appropriate promoter specific for brain ce lls for neurodegenerative diseases that will create new mRNA and 3) a polyadenylation signal site that w ill stabilize the novel mRNA. The promoter is primarily importa nt for determining the level of gene expression and the site and temporal pattern of expression. The most common protocol for creation of transgenic animals is via pronuclear injection into a single-cell an imal embryo. The fertilized eggs are removed from the female and injected with the transgenic cDNA. Multiple embryos (10-20) are then implanted into a pseudopregnant female fost er mouse. A pseudopregnant female is generated by placing her, during a normal cyc ling period, with a vasectomized male. The transgene is either incorporat ed into the mouse genome or is degraded by endonucleases within the cell. If the transgene is incor porated early in devel opment, all cells would inherit the transgene and be passed on to offspring. At weaning, 3 weeks of age, offspring are typically genotyped by polymeras e chain reaction (PCR) for the presence of the transgene of interest. If the incorporation of the tran sgene is successful, roughly half of the offspring should inherit at least one copy of the gene and can pass it on to future generations. Alzheimers disease research is performed with multiple lines of transgenic
38 mouse models as described in detail below in order to provide the most diversity for researching preventions and treatments. PDAPP Transgenic Mice. The PDAPP model is generated from a V717F mutation of the APP gene, also called the London mutation. Multiple studies have been done to characterize the pathology and behavior of th e PDAPP model at various tim e points. Beginning at 3-4 months, PDAPP mice display an increase in synaptic density within cingulate, frontoparietal and parietal cor tices (Dodart et al., 2000b). At this early age, there was also a substantial decrease in the size of the dorsal hippocampu s as compared to the wild type mice. Mature A deposits (plaques with dense core s) were first detected at 3-4 months within CA1 of the hippocampus, the me dial part of the ci ngulate cortex and the corpus callosum level to the dorsal hippocampus. By 6-7 and 10-12 months, PDAPP mice display mature A deposits with dense cores a nd significantly decreased dorsal hippocampus. Both age ranges displayed increas ed synaptic density within cingulate and parietal cortex, and decreased synaptic density in the dentate gyrus and septum. The 1012 month mice also showed decreased syna ptic density within region CA1 and the frontoparietal cortex (Dodart et al., 2000b). One factor that plays a role in the amount of pathology present in this model is the number of transgenes present (heterozygous compared to homozygous). Between 3-4 months, all of the homozygous PDAPP mice exhibited some extent of mature A deposits, whereas only half the heteroz ygous PDAPP mice revealed mature deposits (Dodart et al., 2000b). The 6-7 month homozygous mice had more A deposits than the
39 10-12 month heterozygous mice. At the la tter age, there was 3-4 times more A deposits in homozygous than heterozygous PDAPP with the largest difference evident within the cortical regions (Doda rt et al., 2000b). In opposition to the above study by Dodart et al. (2000b), Games et al. (1995) measured plaque density and A deposits and found no adverse pathology in heterozygous mice at 4-6 mont hs of age. Only at 69 months of age were A deposits detected within hippocampus, corpus callosum and cerebral cortex. By 9 months of age there was a decreased synaptic and dendritic density within the dentate gyrus. The study also reported that A plaques from 9 month old heterozygous PDAPP mice were surrounded with GFAP positive reac tive astrocytes and distorted neurites. In addition, the neocortex from these mice showed diffuse activation of microglia (Games et al., 1995). Masliah et al. (1996) comp ared some of the AD-like pathology from 8-12 month old heterozygous PDAPP mice to tissue sa mples from AD brains. This study found similarities in amyloid depos ition, dystrophic neurites and glial cell reactivity between the PDAPP mice and AD brains. The dense co re structure of the PDAPP mouse plaques from hippocampus resembled the deposits found within frontal cortex of the AD brain. Similarities in dystrophic neurites and the dysfunction of their syna ptic junctions were also seen. The synaptic junctions exhibited damage evident by enlargement of the nerve terminals and fewer synaptic vesicles. The majority of this damage was seen within plaques, not in the adjacent area surro unding the deposits. PDAPP mice also displayed glial cell reactivity (strong GFAP staining), similar to AD, associated with neuritic plaques and the neuropil of the neocortex a nd hippocampus. However, the transgenic
40 mice showed a stronger reactivity of astr oglial cells and less microglial reaction compared to AD brains (Masliah et al., 1996) Overall, this study shows that the 8-12 month PDAPP mouse exhibits marked A pathologic similarities to the AD brain. In aged PDAPP mice, 16-17 months of ag e, Chen et al. (2000) found increased plaque density within multiple layers of the en torhinal cortex, the dentate gyrus, striatum and the moleculare lecunosum of CA1. At 18-21 months diffuse plaques were located within the hippocampus and cingulat e, insular and entorhinal cort ices (Chen et al., 2000). Dodart et al. (1999) beha viorally tested heterozygous and homozygous PDAPP mice at 3, 6 and 9 months of age compared to age-matched non-transgenic littermates. Over all ages tested, PDAPP mice had increas ed activity and decreased rearings in a 50minute open field task compared to wild type (WT) mice; however there were no differences between any of the groups in a simple motor task indicating similar locomotive behaviors between all groups. Young (3-4 month old) PDAPP mice were found to be impaired in reference memory erro rs in the radial arm maze compared to WT mice. Working and reference memory impa irment was correlated positively with increased synaptic density within cingulate and parietal cortices (Doda rt et al., 2000b). In addition, reference memory impairment was posi tively associated with increased synaptic density in frontoparietal cort ex and regions CA3 and dentat e gyrus of the hippocampus. Over all 8 days of radial arm maze testi ng, the PDAPP mice did not show any learning effect, while the WT mice did show significan t learning over all test days (Dodart et al., 1999). At this young age the PDAPP mice were not impaired in object recognition; however they had significantly fewer presses th an wild type control mice in an operant memory bar task indicating memory impairment Adult (6 months of age) and aged (9
41 months old) PDAPP mice also exhibited impa irments in working and reference memory errors in the 8-arm radial maze task. As was the case for young PDAPP mice, there was also no learning effect evident compared to WT mice over the 8-day task for both 6 and 9 month ages. Adult and aged PDAPP mice we re also impaired in object recognition showing a lower time spent exploring the novel object compared to the familiar object. Only at the 6 month age, there was a signi ficant negative correla tion between synaptic density in CA1 and the object reco gnition index (Dodart et al., 2000b) Dodart et al. (2000a) also found an impa irment of object recognition in 10-12 month old PDAPP and PDAPP+ apolipoprot ein E Knock-Out (PDAPP/ApoE-KO) mice compared to age-matched WT and ApoE-KO litt ermates. This study also found a lack of habituation on day-1 of a two-day open field task for PDAPP and PDAPP/ApoE-KO mice; however WT and ApoE-KO mice did elic it habituation. The si gnificance regarding the lack of habituation is su spect due to the fact that the activity of the PDAPP and PDAPP/ApoE-KO mice at the beginning of the trial equaled that of the WT and ApoEKO mice at the end of the trial (i.e., after their significant habituation). In contrast to Dodart, Chen et al. (2 000) found normal object recognition in 6-9, 13-15 and 18-21 month old PDA PP mice compared to WT age-matched control mice. In a modified working memory version of the water maze task where mice were required to reach a criterion prior to changing locations of a submerged platform, 6-9 and 13-15 month old PDAPP mice exhibited a higher latency for each new platform location compared to wild type control mice indicati ng impairment of spatial acquisition memory. From 6 through 21 months, there was a signi ficant negative correlation between the learning capacity in the modified water maze ta sk and plaque burden (Chen et al., 2000).
42 Huitron-Resendiz et al. (2002) also found impairment of spatial memory; however it was evident only in aged PDAPP mice. Young (3-5 months of age) and aged (20-26 months old) mice were tested in a 12day circular platform maze consisting of two trials per day. Young PDAPP mice performed similarly to WT mice initially and exhibited a significant increase in errors to lo cate the escape on the last block of testing only. In contrast, the aged PDAPP mice were significantly impaired to locate the escape on all trials as measured by errors made into non-escape holes. These results indicate a significant impairment of spatial memory that becomes slightly apparent at 3-5 months and prominent by 20-26 months of age in PDAPP mice. Most recently, Nilsson et al. (2004) demons trated the importance of inflammatory proteins, ApoE and ACT, within the PDAPP transgenic mouse lines. All four lines, PDAPP/ApoE/ACT, PDAPP/ApoE, PDAP P/ApoEKO/ACT and PDAPP/ApoEKO had similar amounts of soluble and membrane bound A 1-40 and A 1-42 in hippocampus and cerebral cortex at 2 months of age. At th is age, all lines exhi bited no sensorimotor, learning or memory deficits compared to non-transgenic mice. However, by 18 months of age, differences in pat hology and cognition were evident between the different lines. PDAPP/ApoEKO/ACT mice had increased levels of A deposition within hippocampus and cerebral cortex compared to PDAPP/A poEKO mice. Also, PDAPP/ApoE/ACT mice had increased deposition of compact A plaques in cerebral cortex and hippocampus due to an increase in plaque density compared to mice lacking ACT. In addition, both lines of ApoEKO mice showed no congophilic A deposition in cerebral cortex or hippocampus, indicating a primary role for ApoE to promote formation of compact plaques. Similar to the younger transgenic mice, none of the 18 month old PDAPP mice
43 showed any sensorimotor deficits compared to non-transgenic mice. However, all PDAPP/ApoE mice had increased latency (acqui sition deficit) in the Morris water maze task compared to PDAPP/ApoEKO and non-tran sgenic mice. In addition, PDAPP/ApoE mice exhibited overall working and reference me mory deficits in the radial arm water maze task compared to PDAPP/ApoEKO and non-transgenics. More specifically, PDAPP/ApoEKO/ACT showed deficits in working memory compared to PDAPP/ApoEKO and non-transgen ic mice, indicating that ACT promotes the formation of amyloid, as described earli er, and cognitive decline in addition. This study also positively correlated reference memory from the radial arm water maze task to immunoreactive A deposits and mature compact deposits in the hippocampus. Lastly, PDAPP/ApoEKO mice that showed diffuse A in hippocampus had impaired working memory compared to PDAPP/ApoEKO that showed no diffuse A indicating the association between diffuse A and working memory deficits. Overall, this study showed the importance of the PDAPP transgen ic model interaction with ApoE and ACT genes (Nilsson et al., 2004). Using the same four PDAPP lines mentioned above, Leighty et al. (2003) found significant a ssociations, between cognitive impairment in the Morris water maze, platform recognition task and the radial arm water maze task to A deposition in the hippocampus and cerebral cort ex in transgenic mice at 15-16 months of age. More specifically, working memory and recognition impairme nts were correlated with A deposition in hippocampus and cerebral cortex. This study strongly supported the association between memory/ learning impairments to A deposition within the PDAPP transgenic mouse (Leighty et al., 2003).
44 APPsw Transgenic Mice. In addition to the PDAPP mutation, a commonly over expressed FAD double mutation at residues 670 and 671 of the APP ge ne has been termed the Swedish model, APPsw. Multiple APPsw models have been used with different backgrounds, providing a variety of results. APPsw model with an FVB/N background is hi ghly vulnerable to neuronal cell death within the hippocampus (Moh ajeri et al., 2004). Necrosis, apoptosis and astrogliosis within the hippocampus and cortex are also found within this mouse strain (Moechars et al., 1996) This mouse model also exhibits decreased glucose utilization and astrog liosis within the cereb ellum (Hsiao et al ., 1995). There were, however, no differences in levels of A 40 or A 42 compared to other FVB mixed backgrounds that do not exhibit the previously mentioned CNS disorders (Carlson et al., 1997). A predominant negative feature of mice with the FVB/N background is the prevalence of retinal degenerati on that occurs around post-natal day 5 for transgenic mice and post-natal day 9 for non-transgenic mice (Vinores et al., 2003) ; therefore cognitive evaluation of this strain becomes difficu lt. With this in mind, FVB/N mice are spontaneously hyperactive and aggressive ; in addition, these mice exhibit poor performance in spatial and non-spatial tasks (Voikar et al., 2001; Mineur & Crusio, 2002). Additional behavioral studies suggest that the prevalence of premature death, neophobia, aggression and seizures are depende nt on the level of APP expression (Hsiao et al., 1995; Carlson et al., 1997). Neophobia ha s been associated with thigmotaxis, agitation and tremulousness as early as one month of age (Carlson et al., 1997). A second model of the APPsw mouse is th e APP23 model, generated by insertion of the mutated gene into an XhoI site of an expression casse tte containing the murine
45 Thy-1.2 gene (Sturchler-Pierrat et al., 1997). At 6 months of age, this line shows rare A deposits, but at 24 months there is a substantial deposition of A plaques, with dense cores, including an increase in size and number of plaques associated with an increase of age. Deposits occupied regions of the th alamus, olfactory nucleus, caudate putamen, neocortex and hippocampus; these regions also contained a massive glial response. A deposits were associated with a decrease in the number of adjacent cell bodies in hippocampal pyramidal cells and were also surrounded by dystrophic neurites (SturchlerPierrat et al., 1997). Robust plaque deposits were present even at 14-18 months of age throughout the neocortex and hippocampus of heterozygous and homozygous APP23 mice (Calhoun et al., 1998). The plaque load within region CA1 of the hippocampus inversely correlated with the number of pyramidal neurons within this area constituting a significant neuronal loss compared to non-transg enic control mice. There was, however no significant neuron loss within the neocortex despite the high load of plaques present. Amyloid was also detected within vessels of the meningeal, neocortical and thalamic vasculature at the same age of 14-18 months (Calhoun et al., 1998). Spatial memory deficits exhibited by th e Morris water maze task preceded the formation of plaque deposition within APP 23 heterozygous mice (Van Dam et al., 2003). At 3 and 6 months of age, these mice showed significant learning and memory deficits in Morris water maze acquisition and retention; however significant formation of plaques occurred first at 6 months (Van Dam et al., 2003). In agreement with Van Dam, 3, 18 and 25 month old APP23 mice were found to be impaired in the Morris water maze, as exhibited by increased latency and distance swum to the hidden platform (Kelly et al ., 2003). Also at all ages, APP23 mice had
46 fewer annulus crossings in the memory retent ion trial compared to the non-transgenic mice; however, only at 3 and 18 months did the APP23 mice have a lesser amount of time spent in the former platform-containing quadrant versus age-matched non-transgenic mice in this memory retention trial. B ecause all mice performed similarly to agematched non-transgenic mice in the visible platform version of the water maze, the authors suggested that the spatial impairment of the APP23 mice was not due to sensory, motor or motivational alterations. As opposed to the above findings, Lalonde et al. (2002) found impairment of acquisition in the Morris water maze of 16 month old APP23 mice; however, there was no impairment in the probe trial (memory retention). These mice also exhibited hypoactivity in contrast to the hyperactivity found in mice ranging from 6 weeks to 6 months of age (Van Dam et al., 2003). An additional model of the APPsw m ouse was created with a C57BL/6 X C57BL/6/SJL background, the T g2576 mouse model. Multiple studies have focused on analyzing various markers to charact erize the Tg2576 mouse such as A plaque deposition, and inflammatory, synaptic and oxi dative markers as indicators of an AD-like pathology within a mouse model. Hsiao et al. (1996) first classified the Tg2576 APP mouse by analyzing brain A 1-40 and A 1-42 levels as well as seni le and diffuse plaques within 2-8 month and 11-13 mont h old transgenic mice. As early as 2 months of age, Tg2576 APP mice showed increased neurochemical levels of A 1-40, with a 5-fold increase by 11 to 13 months of age, and A 1-42 with a 14-fold increase by 11 to 13 months. In addition to brain levels of A amyloid deposits were s een in the older (11-13 month) transgenic mice within the frontal, temporal and entorh inal cortices, in addition to
47 the hippocampus and cerebellum. Associated with these A deposits were immunoreactive astrocytes and dystrophic neurites (Hsiao et al., 1996). In agreement with Hsiao et al. (1996), Pratico et al. (2001) showed that the Tg2576 APP mice exhibit A 40 and A 42 at low levels in the br ain from 4-8 months with high levels seen at 12 months, progressing in to 18 months of ag e within the cerebral cortex and hippocampus. Scattered A deposits within cerebral cortex and hippocampus were not evident until 12 m onths of age with more abundant deposits seen in the neocortex and hippocampus at 18 m onths (Pratico et al., 2001). In conjunction with Pratico, Benzin g et al. (1999) found moderate A deposits within 12 month old Tg2576 mice with consid erable deposition occurring by 18 months of age. However, the A deposition was predominately found within temporal and cingulate cortices and to a le sser extent within frontopariet al cortex and hippocampus. This study also examined the presence of astrocytes and microglial cells and their association with plaque depos ition. Astrocytes were main ly localized near fibrillar deposits typically on the periphery of the plaq ue; microglial positive activity was detected around A deposits at 12 and 18 months. In conjunction with the absence of A deposits at 3 months of age, there was also no detec tion of astrocyte or mi croglial activity. The presence of the inflammatory markers, CD -45 and MAC-1 were positively associated with microglial activation and fibrillar A deposits at 12 and 18 months, whereas IL-6 immunoreactivity was associated with as trocyte activation near fibrillar A deposits also at 12 and 18 months within the Tg2576 mouse (Benzing et al., 1999).
48 In contrast to the above studies, Mehlhorn et al. (2000) did not detect any deposition of fibrillar A plaques within cortex of Tg25765 mice through 13 months of age. At 14 months, a low burden of A deposits was detected with in cerebral cortex and hippocampus with the presence of reactive astr ocytes detected around the plaques. This study also detected IL-1 immunoreactive cells around A plaques that co-localized with the astrocyte activation at 14 m onths (Mehlhorn et al., 2000). Irizarry et al. (1997) detected A deposition in Tg2576 mice at 16 months of age within the cingulate cort ex, molecular layer of the dentat e gyrus, entorhinal cortex and region CA1 of the hippocampus. At the subcorti cal levels, rare deposits were evident in the internal capsule and the basal ganglia at th is late age. In conjunction with the above references, positive astrocyte activation was asso ciated with neuritic plaques. This study also looked at neuronal pat hology in reference to the progression of AD. They found that dystrophic neurites were f ound surrounding and within A deposits, however there was no difference in the actual number of neur ons within region CA1 compared to nontransgenics, despite the amount of A deposits within that area. This study also showed no difference in synaptophysin staining (a glyc oprotein located in neuronal synapses and vesicles) of the outer, middle or inner molecular layers of the dentate gyrus between Tg2576 and non-transgenic animals (Irizarry et al., 1997). Although there were no differences seen in synaptophysin staining at 14 months by Irizarry et al. (1997), King a nd Arendash (2002b) noted str ong staining within interior cortex and hippocampus of 19 month old T g2576 APP mice compared to non-transgenic mice, but not at any of the younger ages. The strongest staining occurred in the plaque periphery, with minimal staining present w ithin the core of plaques. Tg2576 mice
49 exhibited a linear plateau of synaptophysin staining from 3 to 19 months, whereas the non-transgenics have a decreasing trend up to 19 months creating the significant difference between the genotypes. Thus, syna ptophysin was higher within the neocortex of Tg2576 mice at 19 months compared to non -transgenic mice. As well, the outer molecular layer and polymorphic layer of th e dentate gyrus had increased synaptophysin staining in the 19 month APP mice versus non-transgenics (King & Arendash, 2002b). In addition to the pathologic markers that are exhibited in the Tg2576 APP model, changes occur in oxidative stress markers as well. Pratico et al (2001) isolated and quantified the presence of is oprostanes (sensitive marker s of lipid peroxidation and oxidative stress) within urine, plasma and brain tissue of Tg2576 APP mice from 4 to 18 months of age. Human AD s ubjects exhibit increased isopro stane (IsoP) levels within urine, plasma and CSF, therefore this oxida tive marker could be used as a useful peripheral indicator for AD. After 6 months of age, APP mice show significant elevations of IsoP levels within the urin e with a maximum plateau occurring after 12 months of age compared to non-transgenic co ntrol mice. Plasma levels of IsoPs are significantly increased after 8 months in APP mice versus non-transgenic animals. As a general reference, there was no difference f ound within the non-tran sgenics at any age within urine or plasma IsoP levels. Prior to 8 months there was no difference in brain levels of IsoPs, specifically cerebral cortex and hippocampus At 8 months and older, cerebral cortex and hippocam pus exhibit increased levels of IsoPs compared to nontransgenics with no difference at any age within the cerebe llum (Pratico et al., 2001). In agreement with Pratico, Smith et al (1998) found increased oxidative stress markers in 13-25 month old Tg2576 APP mice. This study specifically analyzed
50 hemeoxygenase-1 (HO-1) and HNE (lipid per oxidation marker). Both HO-1 and HNE were found to be associated with the peri -plaque regions and neuronal cell bodies that were distant from A deposition. Their presence at th e distant locations indicates a global increase of oxidative stress. However, the detec tion of reactive oxygen species (ROS) was only found to be pr esent at the sites of A deposition (Smith et al., 1998). In a similar age of mice, Pappolla et al. (1998) de tected an increase of superoxide dismutase (SOD) by 21-25 months of age. The presence of SOD overlaps with the evidence of dystrophic neurites and A plaque deposition. The aut hors suggest that the overlap between SOD detection and A deposition supports the idea that A is neurotoxic and that toxicity is mediated by free radicals (Pappolla et al., 1998). Aside from the plethora of pathologic ev idence used to characterize this mouse model, behavioral analyses of learning and me mory are also present to further support the similarities between the progres sion of AD in humans and in mouse models. At 3 months of age, the Tg2576 mouse was found not to be impaired in sensorimotor tasks, Y-maze alternation or entries, Morri s water maze or visible platform (Holcomb et al., 1999). However by 6 and 9 months of age, they e xhibit a significant decrease in Y-maze alternation compared to non-transgenic mi ce indicating a deficit in basic memory function. This study found no impairment in visible platform or Morris water maze acquisition or retention even at 9 months of age. Wher eas Holcomb et al. (1999) found no impairment in visible platform at 9 months, King and Arendash (2002a) found a significant increase in latency at 9, 14 and 19 months for the Tg2576 mouse compared to non-transgenic mice.
51 In agreement with Holcomb, King and Ar endash (2002a) found no impairment at 3 months age in Morris water maze, visible platform and Y-maze entries, as well as circular platform. They did, however, find im pairment in Y-maze alternation, an increase of open field activity and poor balance beam perfo rmance at this early age. This study behaviorally tested mice from 3-19 months of age which resulted in overall impairment in open field activity, balance beam performan ce, string agility and Y-maze alternation in comparison to age-matched non-transgen ic mice (King & Arendash, 2002a). In contrast to King and Arendash, Hs iao et al. (1996) did not observe an impairment of Y-maze alternations at 3 months, but did see impairment at 10 months of age within Tg2576 mice. Also at 10 mont hs of age the mice exhibited acquisition deficits, evident by increased latencies in the Morris water maze as compared to nontransgenic control mice. Mi ce between 9 to 15 months ha d a significant impairment of memory retention also within the Morris wate r maze. Overall, the authors suggest an association between the appearance of A 40 and A 42 neurochemically within the brains of 2 month old APP mice that progressively increased through 13 months to the progression of learning deficits that were observed in thes e ages (Hsiao et al., 1996). Westerman et al. (2002) al so found significant impairme nt of memory retention within the probe trial of the Morris wate r maze in groups of mice ranging from 6 to 25 months of age, however no impairment was evid ent in very young mice (4-5 months old). The presence of insoluble A was detected in all mice over 10 months old; therefore the authors suggested an association between the presence of A insol and the occurrence of memory impairments within this m ouse model (Westerm an et al., 2002).
52 APP/PS1 Transgenic Mice. The last model of Alzheimers disease reviewed here is a double transgenic mouse involving the APPsw mutation and a misse nse mutation of the presenilin 1 gene. In general, APP/PS1 mice exhibit an acceler ated amyloid deposition compared to the single transgenic APP m ouse. Some studies even suggest that deposition begins to occur as early as 3 months of age. Holcomb et al (1998) detected the presence of a modest amount of compact A deposition between 12-16 weeks of age. By 24-32 weeks of age the deposition was increased and the plaques became surrounded by reactive astrocytes (Holcomb et al., 1998). In agreement with Ho lcomb, Takeuchi et al (2000) detected A deposition in the double transgenic model at 3 months of age. Specifically, plaques were found in cingulate, superior fr ontal and parietal neocortices with a lesser amount detected within the hippocampus. From this age, th e deposition progressed to include larger compact plaques scattered throughout the cort ex and small diffuse plaques displayed by 6 months of age. By 9 to 12 months, larger dense plaques and sma ll diffuse plaques had increased in number and were found within th e neuropil of the entire neocortex. Despite the increased number of plaques found by 12 mo nths, Takeuchi et al. (2000) did not see any changes in the number of neurons pr esent in region CA1 or in synaptophysin immunoreactivity in the molecula r layer of the dentate gyru s and CA1 compared to agematched non-transgenic and single APP transgenic mice. In contrast to the above studies, Gor don et al. (2002) did not detect any A deposition at 3 months of age in APP/PS1 mi ce; the first deposits were identified at 6 months, mainly within frontal and entorhinal cortices a nd hippocampus. These deposits were also associated with r eactive astrocytes and dystrophic ne urites. Diffuse deposits in
53 striatum, thalamus and brainstem were defined in 12 month old APP/PS1 mice. Progression of deposition density and number of diffuse plaques occurred at 15 months; however no deposits were found in granul ar and pyramidal cell layers of the hippocampus or the corpus callosum. At th is late age, GFAP staining for reactive astrocytes was greatly increased throughout the brain, mainly c oncentrated in the striatum and cerebral cortex, compared to their first detection from 6 month old mice (Gordon et al., 2002). Although Borchelt et al. (1997) also detected substantial A deposition within cortex and hippocampus in 12 month old A PP/PS1 mice, no deposition was found in any other region in contrast to Go rdon et al. (2002) that found pl aques in striatum, thalamus and brainstem. Borchelt et al. (1997) also reported a progression of deposition from 9 to 12 months with increased number of plaque s within hippocampus and occipital and frontal cortices. In conjunction with the previous studies, many of the deposits were associated with reactive astrocytes (Bor chelt et al., 1997). Overall, the APP/PS1 transgenic model develops modest A deposition associated with reactive astrocytes around 3-6 months of age with a progression in density and number within cortex and hippocampus thereafter. Numerous behavioral studies have been performed similar to the previously discussed models to characterize the developm ent of cognitive impairment. At 3 months of age, APP/PS1 mice were found to exhibit no deficits in sensorimotor tasks, but did become impaired in Y-maze alternation and exhibit increased activ ity in Y-maze entries by 6 to 9 months of compared to age-ma tched non-transgenic mice (Holcomb et al., 1999). However, even at 9 months there we re no sensorimotor deficits or memory
54 deficits revealed by visible platform or water maze acquisition or retention trials (Holcomb et al., 1999). By contrast Arendash et al. (2001b) found no difference of APP/PS1 mice from non-transgenics at 5-7 months in Y-maze alte rnation; however, they did have increased activity as measured by Y-maze entries. The same study also found impairment in a balance beam task at this age with no im pairments in any additional tasks including sensorimotor and learning or memory tasks (Arendash et al., 2001b). At 15-17 months of age, double transgenic mice exhibited increased activity in open field and Y-maze entries versus non-transgenics with additional impa irments in balance beam, string agility, Morris water maze acquisition and radial arm water maze (RAWM) working memory (Arendash et al., 2001b). With in this study correlation anal yses revealed a significant positive relation between working memory errors in the radial arm water maze task and Congo red staining in the frontal cortex (Gordon et al., 2001). There was also a significant negative correlation between total A load in the frontal cortex and hippocampus and radial arm water maze acquisi tion error reduction between trial 1 and trial 4 (Gordon et al., 2001). Of the three AD transgenic models pr esented, the APP/PS1 mouse has certain advantages, including co nsiderably earlier A deposition and a comprehensive behavioral evaluation at multiple time points. It is, therefore, ideally suited for studies of this dissertation, which involve the testing of vaccination and dietary therapeutics against behavioral impairment and AD neur opathology provided by that model.
55 IX. AD Vaccination Current treatment options for AD include cholinesterase inhibitors, NSAIDs, statins and a variety of antioxidants as previo usly reviewed. However, these therapies are largely aimed at the symptomatic basis of th e disease instead of the underlying pathology involved. Utilization of immunotherapy as a possible treatment of AD involves inducing an immune response to fight misfolded proteins or aggregates of proteins that accumulate in the pathogenesis of AD. Therefore, im munotherapeutic appro aches toward AD hold valid opportunities to treat th is neurodegenerative disease through a humoral or cellular immune-mediated response. In Vitro Studies. In 1996, Solomon et al. first experimented with monoclonal an tibodies (mAbs) to test their effectiveness against A peptide aggregation, fibrilliza tion and toxicity in vitro. They found that mAbs against A 1-28 were effective at converting fibrillar A to an amorphous state. In 1997 Solomon et al. then showed that the mAb 6C6 against A 1-16 significantly solubilized fibrillar A in comparison to control cells. Starting in 1998 Frenkel et al. began to determine the sp ecific epitope for antibody binding which was responsible for comple tely inhibiting A fibril formation. They determined that the EFRH epitope corresponding to A 3-6 was most effective at solubilizing fibrillar A These results support the studies previous ly performed by Solomon et al. (1996, 1997), who determined that mAbs against A 1-28 or A 1-16 were more successful than mAbs against A 8-17 or A 13-28 for disaggregating A fibrils. Following these studies were A
56 immunotherapy experiments in multiple transgenic models to determine the efficacy of previously determined in vitro results in an in vivo environment. Active A Immunotherapy. Schenk et al. (1999) fi rst explored active A immunization by administering the A 1-42 peptide to 6 week old PDAPP mice once monthly until 13 months of age. The vaccination sequence resulted in al most complete prevention of A deposition. Immunization of 11 month PDAPP mice until 15 a nd 18 months resulted in arrest of total A deposition within the cortex to the level present at 11 months. Immunization resulted in the generation of blood anti-A antibodies and activation of microglia found near plaques; therefore one possible mechanism for A clearance, before or after plaque formation, could be due to anti-A antibodies triggering microglial cells to clear A through an Fc receptor-mediated phagocytosi s. Schenk et al. (1999) concluded that active immunization could be used as both a pr evention and therapy to block or retard the pathological development within an AD-lik e pathology. Disappointi ngly, there were no behavioral measures performed on these mice to determine if the improved pathology could also retard behavioral impairments. In 2001, Dickey et al. vaccinated 5-month old APP+PS1 mice with the A 1-42 peptide monthly for 7 months. They found that more than three vaccinations are necessary to in duce a 50% maximal antibody titer that progressively increased after the 6th booster vaccination. A competition binding inhibition assay was used to determine binding efficacy and specificity of the sera from the immunized mice. A 1-16 was the only truncated A peptide from the sera that
57 inhibited binding of antibodies to A 1-42. Also, the A 1-42 was more effective that A 1-40, indicating the importance of the 2 amino acid difference between the peptides in antibody binding. Because A 1-42 forms fibrils more readily than A 1-40, this conformational change could significantly contribute to th e preferential recognition by the immune system. Lastly, T-cell proliferation was increased in spleen cells of APP+PS1 vaccination mice (Dickey et al., 2001). Recently, Koller et al. (2004) vaccinated APP mice with A 1-42 cross-linked with an Hsp70 homolog of Escherichia coli DnaK, beginning at 6 weeks of age (prior to onset of plaque deposition), with 3 inoculations ending at 12 weeks of age. In opposition to Dickey et al. (2001), measur able antibody titers against A 42 were detected after only one inoculation with the DnaK -A 42 vaccination. At 6-9 months, brain levels of A were determined for human and murine A in brain homogenate. The DnaK -A 42 immunized mice showed a surprising increase of extractable A indicating an increase of A aggregates as a result of the vaccination. This change occurred without any significant plaque deposition. By 12-14 months of age, both DnaK -A 42 immunized and control APP mice showed low levels of plaque bur den in the neocortex with no difference between the two groups. Also at th is age, the extr actable total A and A 42 were increased in the brains of the DnaK -A 42 immunized mice compared to the nonimmunized APP mice. Both immunized a nd non-immunized APP transgenics exhibited A deposits in the walls of cerebral blood vesse ls. There was a trend toward a slightly higher A deposition in the brains of immunized mice, as evidenced by their having A deposits within smaller cerebral blood vessels in addition to A deposits in large cerebral
58 blood vessels. The authors propose a po ssible link between the increased A aggregates resulting from vaccination and the non-significant increase in A within small cerebral vessels. The authors propose that this link may be a direct result of the vaccination, however there is no significant evid ence to support this association. Monsonego et al. (2001) immunized A PP Tg2576 transgenic mice with A 1-40 at 5 weeks of age, with blood samples colle cted ten days later for antibody and T-cell analysis. Immunized APP mice had lower levels of anti-A antibodies compared to immunized wild-type littermates; the APP control vaccinated mice did not produce any detectable levels of anti-A antibodies. The A 1-40 APP immunized mice also had decreased T-cell pr oliferation, INF, and IL-2 secretion compared to non-transgenic littermates. In opposition to Dickey et al. (2001), Monsonego et al (2001) did not show an increase in T-cell proliferation. The diffe rence could largely be due to the number of inoculations given, type of v accination, and/or the difference in transgenicity used in the animal models. The sera from the APP immunized mice were used to positively stain neuritic plaques in brains of mature APP mice; sera fr om non-immunized mice did not positively stain any neuritic plaques. The au thors suggest that the APP immunized mice exhibit a hyporesponsive im mune response to the A 1-40 vaccination, as evident by reduced T-cell and anti-A antibody production. They rela te this impaired immune response of the APP mice to humans that have chronic elevations of brain and peripheral A resulting in their also exhibiting an impaired immune response to A vaccinations. Aside from the previous studies that used direct injections of the A 42 peptide, Weiner et al. (2000) performed nasal ad ministration of the less amyloidogenic A 40
59 peptide to 4.5 month old PDAPP mice once weekly for 7 months. Mucosal administration has the ability to induce an anti-inflammatory immune response which is antigen-specific within lymphoid tissue. Th is response then acted systemically to significantly reduce A deposition within cortical and hippocampal tissues. This occurred in conjunction with a decrease in mi croglia and astrocyte activation within these areas (Weiner et al., 2000). Mo re recently, Lemere et al. (20 03) showed that 5 week old APP+PS1 mice immunized intranas ally with a cocktail of A peptides (3 parts A 1-40, 1 part A 1-42) twice weekly for 8 weeks exhibi ted a 75% reduction in cerebral A plaque number and 58% decrease in neurochemical levels of A 1-42 in the brain. There was also a large increase in serum A antibody levels compared to untreated control transgenics. The authors concluded that b ecause most of the serum A was bound to antibodies, that the antibodies attached to serum A to aid in clearance. McLaurin et al. (2002) vaccinated TgCRND8 mice with protofibrillar aggregates of A 1-42 or control vaccinations. This resulted in recognition of A 42 monomers, tetramers, hexamers and large oligomers by A 42-immune sera. The A 42-immune sera did not activate the Th1 helper cell pro-inflammatory cascade, but instead the Th2 helper cell response was activated, which pr omotes B cells to make anti-A antibodies. Specifically, they found that A 4-10 peptide was highly recognized by the anti-A 42 IgGs. A fibril formation and toxicity were inhibited by anti-A 4-10 without activation of surrounding microglia. Overall, McLaurin et al. (2002) concluded that using smallmolecules could decrease some of the detrim ental problems seen in immunotherapy such as a pro-inflammatory response. This study had similar results compared with previous
60 work by Frenkel et al. (1998), who fi rst reported that the epitope of A 1-42 is the A 4-10 segment; Frenkel et al. proposed the segment specifically corresponded to A 3-6. Additional active immunization animal stud ies have been done to prevent against AD progression. These st udies administer A active immunization prior to A deposition in transgenic models and behavi orally assess changes that could occur in conjunction with A pathology. Vaccination of 6 w eek old TgCRND8 mice with A 42 for 3 months resulted in a 50% de crease in the number and size of A -positive dense core plaques within the hippocampus and cort ex (Janus et al., 2000). At various age points from 11 to 23 weeks, Morris water maze testing was done; improved performance of the A 42 vaccination group in comparison to c ontrol immunized mice was observed at several, but not all, test points (Janus et al., 2000). Morgan et al. (2000) began administration of A 1-42 peptide to APP+PS1 transgenic (Tg+) mice at 5-7 months of age prior to A deposition and behavioral deficits. The inoculations were given monthly until 15.5 months of age. High antibody tit ers were found in immunized Tg+ and nontransgenic (Tg-) mice; control vaccinate d and untreated mice did not produce any detectable antibodies to A 42, indicating that transgenic mice do no spontaneously produce anti-A antibodies. A modest, non-significant, decrease in A burden in the frontal cortex was produced in the brains of Tg+ immunized mice; however, there was a significant reduction in Congo red-positive compact plaques within the frontal cortex of Tg+ immunized mice. Trends toward reduced A burden and Congo red staining were found in the hippocampus of vaccinated mice. All mice were behaviorally pretested before vaccination began, ~5-7 months of age, and similarly behaviorally tested at ~16
61 months of age after 8 months of A inoculations. At the 16 m onth test point, vaccinated Tg+ mice performed similar to Tgmice in radial arm wa ter maze working memory, and significantly better than contro l Tg+ mice (Morgan et al., 2000). Despite 8 months of vaccinations, Tg+ mice remained impaired in balance beam compared to Tgmice (Arendash et al., 2001a). In order to relate A deposition to behavioral performance, correlation analyses were performed. Thes e analyses showed significant correlations between A 1-40 deposition in hippocampal region CA1 and radial arm water maze working memory impairment for all Tg+ mice grouped together. Additionally, A 1-40 deposition and Congo red staining in frontal cort ex correlated with radial arm water maze working memory impairment for both Tg+ groups combined. The authors suggest that the behavioral protection offered by an A vaccine selectively and primarily preserves hippocampal-associated working memory function (Arendash et al., 2001a). Additionally, the authors propose that the si gnificant prevention of memory deficits exhibited in the radial arm water maze task, coupled to the slight ly reduced but still substantial A load, could be due to the ability of A antibodies to neutralize soluble, non-deposited A (Morgan et al., 2000). This type of A has been implicated in memory loss due to synapse loss in de ntate gyrus. A second proposed theory to account for the observed results is through a clearance of deposited A by activated microglia; however the lack of significant difference in A deposition between the vaccinated and control transgenic groups does not support this theory (Morgan et al., 2000). Recently, Sigurdsson et al. (2004) immuni zed 6 to 8-month old Tg2576 APP mice with monthly inoculations of nontoxic A 1-30 or K6A 1-30 until 19 to 21 months of
62 age. K6A 1-30 binds specifically to re sidues 1-11 and 22-28 of A 42 without adopting a -sheet conformation and is therefor e less toxic. Immunization with A 1-30 produced no difference in amyloid plaque burden or soluble A levels. Likewise, vaccination with K6A 1-30 to the APP mice did not produce a si gnificant change in overall amyloid burden; however, there were reductions in small and medium-sized plaques in the vaccinated group compared to control-immuni zed APP mice. There was also no change in soluble A levels between the K6A 1-30 mice and the control group. In addition to the pathologic markers, Sigurdsson et al. (2004) showed that A PP mice vaccinated with K6A 1-30 had improved cognitive performance, as shown by a decrease in errors in the radial arm maze task compared to cont rol vaccinated APP mice. Both the K6A 1-30 vaccinated and control wild-type mice exhibite d a learning effect, as shown by a decrease in errors from day 1 to day 9; the contro l vaccinated APP mice did not show any learning effect. Overall, because the K6A 1-30 vaccinated mice showed an improvement in learning with minimal changes in amyloid burden, the authors suggest that this type of immunization clears A oligomers which are linked to behavioral performance. In comparison to vaccination studies for prevention against AD, few studies have focused on the use of immunization as a treatme nt option for AD. Wilcock et al. (2001) immunized 3 groups of A PP+PS1 mice with the A 1-42 peptide beginning at 7.5, 13 and 14.5 months of age for 9, 3 and 5 months, re spectively. Overall, all groups exhibited decreases in Congo Red staining in hippocampus, but not in frontal cortex. In another study, Sigurdsson et al. (2001) vaccinat ed 11-13 month old Tg2576 mice with a nonamyloidogenic, non-toxic A homologous peptide for 7 months until the age of 18-20
63 months. This resulted in decreased amyloi d burden in cortex and hippocampus as well as decreases in Congo Red staining and soluble A 42 deposition in comparison to control vaccinated mice. In opposition to Wilcock et al. (2001), a decrease in Congo red staining was observed in frontal cortex. The disc repancy between Wilcock and Sigurdsson is most likely due to the difference in vaccination protocol used between the two groups, non-toxic A homologous peptide versus A 1-42 peptide. In 2003, Frenkel et al. injected th e EFRH epitope, corresponding to A 3-6, into 16-month old PDAPP mice monthl y for 4.5 months. In agreement with Dickey et al. (2001), antibody titers reached high levels after 6 injections. They demonstrated that monoclonal antibodies that disaggregate A also bind to the EF RH-phage; conversely, monoclonal antibodies that do not disaggregate A do not bind to the EFRH-phage. Incubation of prepared A fibrils with anti-EFRH sera re sulted in disaggregation of the fibrils into an amorphous state. Likewi se, plaques from the hippocampus of PDAPP mice and AD human subjects were positively stained with the anti-EFRH anti-sera. Analysis of 21-month old PDAPP vaccinate d mice showed a significant reduction in amyloid burden compared to the control-vacc inated PDAPP mice. The authors suggest that the EFRH-phage induced an auto-imm une response by PDAPP mice to produce antiEFRH that binds and disaggregates A fibrils to reduce the number of A plaques in the brain (Frenkel et al., 2003). Austin et al. (2003) performe d biweekly injections of A 1-42 for 6 weeks in cognitively impaired 16 month ol d APP+PS1 mice. Mice were pre-tested in radial arm water maze and platform recognition tasks pr ior to the first vaccination. All Tg+ mice
64 exhibited impaired working me mory retention within both tasks. Within the platform recognition task, overall both groups of Tg+ mi ce had increased latency compared to Tgmice, but by the end of the task (day 4) they exhibited similar performance to Tgmice. Following the vaccination period, aged transgenic mice exhibited working and reference memory impairment in the radial arm wate r maze task that was not corrected with A immunization. However, both groups of Tg+ still showed similar performance to Tgmice in platform recognition by having decreased latencies by day 4 in this task. By 20 months of age after the final behavioral assessment, anti-A antibody levels were detected in all Tg+ mice that were vacci nated; in opposition, no antibody levels were detected in control vaccinated mice. The au thors concluded that a longer vaccination immunotherapy might provide a better cognitiv e advantage to aged APP+PS1 mice. Passive A Immunotherapy. Passive immunization studies involvi ng injection of antibodies against A 1-42 have been performed in A -depositing transgenic mouse mo dels. Pfeifer et al. (2002) suggest that the binding of an antibody to soluble A has the potential to cause a local inflammatory reaction and can destabilize w eakened vessel walls. Transgenic APP23 mice (which develop amyloid angiopathy) were immunized with a monoclonal antibody against A that recognized A 3-6. This resulted in a significant increase in hemorrhage severity mainly within amyloid-rich vessels. The authors suggest th at the weakening of the vessels could be due to an A -induced increase in vascular permeability, resulting in the loss of smooth muscle cells. In additi on, they suggest that an increase of antibody
65 binding to soluble A in the blood could lead to increase d risk of cerebral hemorrhage. However, no other passive immunization studies use mouse models that develop cerebral amyloid angiopathy and therefor e do not show an increased risk for cerebral hemorrhage to support these findings. Also, the immuni zation resulted in a 23% decrease of A load in neocortex, mainly due to a reduction of diffuse deposits (Pfeifer et al., 2002). Bard et al. (2000) however, showed that PDAPP mice immunized weekly with monoclonal antibodies, 10D5 or 21F12, beginning at 8 to 10 months of age for 6 months had more than an 80% decrease in plaque burden within frontal cortex with the 10D5 antibody. The 10D5 monoclonal antibody is specific to A residues 3-6, whereas 21F12 is specific to A 42 residues 34-42. The same group also immunized 11.5 to 12 months old PDAPP mice with monoclonal antibodi es, 3D6 and 16C11 for 6 months. A significant decrease in frontal cortex plaque burden was seen only with 3D6 and in the absence of T-cell proliferation within splenocytes. 3D6 is specific to A residues 1-5, whereas 16C11 is specific to A residues 33-42. The authors c oncluded that this type of passive immunization against A is sufficient to induce decreases in amyloid deposition without T-cell mediated cellula r immunity. In agreement w ith Schenk et al. (1999), these authors suggested that the mechanism for A clearance involves antibodies entering the CNS and binding to Fc receptors on microglial cells to trigger phagocytosis of deposited A peptide (Bard et al., 2000). DeMattos et al. (2001) vaccinated 4-mont h old PDAPP mice for 5 months with the m266 monoclonal antibody, which is specific to A 13-28. M266 vaccinated mice had marked reductions in A deposition in the cortex and dorsal hippocampus compared to
66 control vaccinated PDAPP mice. Despite the reduction of A deposition, m266 did not bind to A deposits within the brains of PDAPP mice. In addition, A levels in the plasma were increased 1000-fold in vaccinate d transgenics compared to control mice. Since, m266 does not cross the blood brain ba rrier, the authors suggested that m266 acted as an A sink to bind and sequester A from the CNS to significantly increase plasma A to aid in clearance. Specifically, the aut hors suggested that m266 altered the plasma and brain A equilibrium to favor A in the plasma and theref ore reduce deposition in the brain (DeMattos et al., 2001). An additional passive immunization trea tment study involved a single ICV injection of a monoclonal antibody AMY33 (specific to A 1-28), into the third ventricle of 10-month old Tg2576 APP mice (Chauhan & Si egel, 2003). The authors claim that AMY33 immunization reduced the size of co mpact plaques, at one-month after the injection, compared to control vaccinated APP mice; however, the figures do not support this statement and no quantific ation of plaques was performed. There seems to be no difference in staining between the treatment a nd control groups. The study also claimed that AMY33 immunization decreased th e number of immunoreactive microglia surrounding compact cerebral plaques. Overall, AMY33 does not seem to be effective at reducing AD-related pathology such as A burden or microglial response. A similar short-term passive vaccina tion study topically applied the anti-A antibody 10D5 by craniotomy to the cortex of 19-23 month old PDAPP mice (Lombardo et al., 2003). The antibodies were only eff ective to the immediat e cortical area where they are applied; remote areas from the appli cation site were not a ffected by the antibody. Amyloid burden was decreased si gnificantly only in the immedi ate area of application at
67 4 and 32 days after topical treatment with 10D 5. Also in the immediate area from 4 to 32 days post-treatment, neurites from 10D5 immunized PDAPP mice appeared normal similar in fact to those in non-vaccinated Tg mice; neurites in a remote area of PDAPP mouse brains (not exposed to 10D5) had a gr eater curvature ratio. The authors concluded that the curvature neurite abnormalities exhib ited in the remote area of PDAPP brains are linked to A deposition and that these neurites undergo rapid restoration to a normal morphology following A plaque clearance (L ombardo et al., 2003). Kotilinek et al. (2002) found that intrap eritoneal administration of BAM-10 (a monoclonal antibody that recognizes A residues 1-12) for one week resulted in no change in neurochemical A 40 or A 42 levels within brains of 9-11 month old Tg2576 mice versus Tg+ controls. However, the BAM-10 group did perform significantly better in Morris water maze acquisition and probe tr ial retention compared to a pretreatment performance within the same group. The authors propose that toxic A assemblies or oligomers are responsible for cognitive impa irment seen in Tg+ mice; therefore, neutralization of these A species could lead to cognitive improvement. This is the proposed mechanism of action of BAM-10 si nce resulting improvement in memory/ learning of Tg+ mice occurred wi thout alterations in brain A levels. In contrast to the preventi on study performed by DeMatto s et al. (2001), Dodart et al. (2002) found that a 6 week treatment with the antibody m266 in 24 month old PDAPP mice resulted in no change in A burdens in hippocampus or cortex, but the vaccinated mice given high doses of m266 did perform sign ificantly better in an object recognition task than Tg+ mice given PBS injections. Since m266 does not cross the blood-brain
68 barrier, the mechanism is thought to be thr ough the peripheral clea rance or sequestration of soluble A in order to efflux A from the brain. This theory stems from the detection of an A /m266 complex in the plasma and CSF at high doses of m266. The authors suggested that the high doses are necessary for m266 to enter the central compartment and draw soluble A from the brain into the pl asma (Dodart et al., 2002). More recently, Wilcock et al. (2004) vaccinated 19-month old Tg2576 APP mice weekly with the anti-A antibody 2286 (specific to A 28-40) for 1, 2, or 3 months. Serum A antibodies were detected at high le vels after 1 month of vaccination and associated with high circulating serum A levels. However, by 2 and 3 months of vaccination, serum A decreased, but remained elevated compared to control vaccinated APP mice. Both compact and diffuse A deposition were reduced in the cortex by ~60% and hippocampus by ~55% of APP mice vaccina ted for 2 and 3 months compared to control vaccinated APP mice. Also, transient microglial activation was seen in vaccinated APP mice at 2 months of treatment compared to control APP mice. By 3 months of treatment, microglial activation retu rned back to levels found at 1 month of treatment. At 22 months of age, a behavior al assessment was done to evaluate the effects of the vaccination. All groups were tested in the Y-maze task, which indicates general memory function and activity. APP vaccina ted mice had improved performance shown by a higher percent alternation compared to A PP control vaccinated mice. Also, the APP control mice, in addition to APP mice vaccina ted for 1 month, had significantly increased arm entries versus non-transgenic mice. However, following 2 or 3 months of vaccination treatment, APP mice exhibited si milar activity levels to that of the non-
69 transgenic mice. The aut hors suggested that anti-A antibodies are able to enter the brain and activate Fc receptors on micr oglial cells to stimulate phago cytosis, as first described by Schenk et al. (1999). A Immunotherapy in Humans. Following the evidence that both activ e and passive immunization of mouse models decrease A deposits within brain regions a nd reduce cognitive deficits, human trials began with immunization of A 1-42 (AN-1792) by Elan and Wyeth-Ayerst Laboratories. After a few m onths during Phase II trials, ab out 5% of the participants receiving the vaccine developed severe infla mmation in the spinal cord and brain, so the trials were halted. However in a Zurich cohort there was no significant correlation between the production of antibodies and the incidence of aseptic meningoencephalitis (Hock et al., 2003). This tr ial included 30 participants w ith mild to moderate AD; 24 patients received the vaccination of AN1792 and developed antibodies. After all the trials were halted, antibody production and reactivity, due to the vaccination, were analyzed. They discovered that human sera produced antibodies that reacted with amyloid plaques and vascular amyloid on br ain sections of APP+PS1 transgenic mice and -amyloid from human brains (Hock et al., 2002). Following a one-year follow-up, 19 of 28 patients that deve loped antibodies showed a stabilization of cognitive performance, as measured by the MMSE, co mpared to subjects that did not produce antibodies and exhibited cogniti ve decline. The patients th at developed antibodies also had preserved hippocampal function as shown th rough the Visual Paired Associates Test versus patients that did not develop anti bodies. Lastly, due to unchanged CSF and
70 plasma levels of A 40 and A 42, the antibodies produced from the vaccine did not sequester serum A in order to alter cognition (Hock et al., 2003). Despite the above beneficial results from human A 1-42 vaccination, the serious inflammatory complications from the vaccine do not allow for additional hum an trials with this approach. Passive immunization or vaccinations with modified forms of A have been recommended to circumvent some of the problems associated with active immuniza tion, such as stroke, encephalitis and sterile meningitis (McGeer & McGeer, 2002). Proposed mechanisms that could account for the unexpected inflammato ry reactions seen in a few of the patients in the clinical trials include : 1) autoimmune disorder due to antibody reactivity with host proteins and 2) autodestruc tion due to the host cells be ing damaged by the membrane attack complex (MAC). Under normal c onditions antibodies ar e generated against foreign material; however spontaneous genera tion against host prot eins can lead to disorders known as autoimmune diseases. Autodestruction occurs when MAC, which normally protects host cells by attacking bacter ial and viral pathogens, is over activated as occurs during AD; this results in more complement activation and destruction of neurons leading to an increase of dama ge to host cells (McGeer & McGeer, 2002). IX. Omega-3 Fatty Acids General Background. Omega-3 fatty acids are found mainly en riched within fish and fish oil supplements. The specific fatty acids at high concentrations within these foods include docosahexaenoic acid (DHA) and eicosapentaeno ic acid (EPA). Omega-3 fatty acids can
71 also be found in hemp oil, flax oil and green leaves (Haag, 2003). Retrospective studies indicate that a high intake of n-3 fatty acids is associated wi th a reduced risk of dementia, specifically AD, prevention of cardiac a rrhythmias, through reduced thrombosis, decreased inflammation including arthritis and decreased weight loss to reduce the risk of death associated with cancer and AD (K almijn et al., 1997a; Kalmijn et al., 1997b; Horrocks and Yeo, 1999; Knittweis, 1999). Omega-6 fatty acids are found mainly within meats and vegetable oils such as sunflower oil, evening primrose oil, corn oil and safflower oil (Haag, 2003). Arachidonic acid, an omega-6 fatty acid, can specifically be found in meats, eggs, shrimp and prawns (Das, 2003). Essential fatty acids (EFAs) in clude linoleic acid (LA) and -linolenic acid (ALA). These fatty acids can not be produced de novo and must be provided by dietary intake. All of the polyunsaturated fatty ac ids (PUFAs) are synthesized from these two EFAs; LA gives rise to all of the omega-6 fa tty acids while ALA synthesis results in all of the omega-3 fatty acids. Fatty acids are named to identify the structure of the molecule so that DHA is denoted as 22:6n-3. The first number, 22, refers to the number of carbon atoms in the hydrocarbon backbone the 6 represents the number of double bonds and the 3 gives the position of the last double bond. N is equal to the number of total carbon atoms so that n-3 represents the number of carbon atoms from the last double bond to the terminal methyl group. As indicated from animal studies, the primary supplier of brain DHA levels is from the plasma. Plasma DHA is either obtai ned from dietary intake or from precursors synthesized into DHA by the liver. Add itional studies indicat e that DHA can be synthesized directly from the brain and it s concentration can be regulated by n-3
72 precursors or the availability of preforme d DHA. Specifically, DHA synthesis occurs within astrocytes, which provide DHA in high concentrations to neurons and blood-brain barrier endothelium cells (Williard, et al., 2001). Because mammals do not possess the enzyme omega-3 desaturase they can not convert omega-6 fatty acids into omega-3 fatty acids. The synthesis of fatty acids within the liver consists of desaturation and el ongation reactions within the endoplasmic reticulum and peroxisomes as illustrated below in Figure 2. Initially, ALA receives an additional double bond to change from an 18:3n-3 fat into an 18:4n-3 fatty acid by -6desaturase. Elongation of the carbon chain occurs to produce 20:4n-3 which is then altered by -5 desaturase to yield EPA (20:5n-3). EPA is then elongated by two separate reactions to produce 24:5n-3. -6 desaturase then adds another double bond to produce 24:6n-3, which undergoes -oxidation to yield DHA (22: 6n-3) (Ferdinandusse, 2001; Haag, 2003). Unfortunately only 5% of the dietary ALA is ultimately converted into DHA (Sarsilmaz et al., 2003). Initially it wa s proposed that all of the fatty acid conversion to DHA were performed within the endoplasmic reticulum (ER), however recent research has proven than the final -oxidation step to yield DHA occurs within the peroxisomes. The potential enzymes responsib le for this conversion have been identified as either a straight-chain acyl-CoA oxidase (SCOX) or a D-bifunctional protein (DPB) (Ferdinandusse, 2001). Omega-6 fatty acid c onversion from LA (18:2n-6) undergoes the same path using the same enzymes to ultim ately yield 20:4n-6 (arachidonic acid, AA). Additional products are also produced at several stages in the pathway by both omega-3 and omega-6 fatty acids. Di-homolinoleic (DGLA) acid, an n-6, can give rise to precursors for 1-series eicosanoids. Eicosanoids encompass prostaglandins and
73 thromboxanes. Arachidonic acid (AA), also an n-6, can form precursors to 2-series eicosanoids, while EPA, an n-3 fatty acid, can produce precursors for 3-series eicosanoids (Fig. 2). The 2-cl ass is pro-inflammatory whereas the 3-class of eicosanoids is anti-inflammatory in function (Das, 2003).
74 -linoleic acid (ALA) 18:3n-3 18:4n-3 20:4n-3 Eicosapentanoic acid (EPA) 20:5n-3 24:5n-3 Docosahexaenoic acid (DHA) 22:6n-3 Linolenic acid (LA) 18:2n-6 -linolenic acid (GLA) 18:3n-6 20:3n-6 Arachidonic acid (AA) 20:4n-6 Fig. 2. Fatty acid synthesis within liver from dietary intake of -linolenic acid (ALA; 18:3n-3) and linoleic acid (LA; 18:2n-6) to ultimately give rise to docosahexaenoic acid (DHA; 22:6n-3) and arach idonic acid (AA; 20:4n-6) Dietary Omega-6 Omega-3 -6-desaturase elongase -5-desaturase elongase 2-Series Eicosanoids 3-Series Eicosanoids SCOX or DPB ( -oxidation )
75 Dietary Manipulation of n-3 and n-6 Fatty Acids. Most of the impacts from fatty acids occur following their incorporation into cellular membranes. DHA has the potential to influence membrane fluidity, hormone production, enzyme activities and the formation of lipid peroxidation products (Stillwell & Wassall, 2003). While affinity for DHA inco rporation into membranes is highest in retinal cells (Alessandri, 2003), it is also present in synaptosomes and sperm cell membranes (Stillwell & Wassall, 2003), wher eas the hippocampus and cerebral cortex show the lowest affinity for DHA membrane incorporation (Alessandri, 2003). Heart, skeletal muscle, liver and kidney membranes also contain levels of DHA within their phospholipids (Turner et al, 2003a). Particul ar phospholipids (PL) classes incorporate DHA at distinct rates and affinities. Stillw ell and Wassall (2003) de termine that within plasma membrane and mitochondria, DHA firs t integrates into phos photidylethanolamine (PE) followed by incorporation into phosphotid ylcholine (PC) and the remainder of the PL classes. Within these two PL classes, cerebral cortical cell membranes from hamsters fed an omega-3 deficient diet cont ained significant decreases in DHA and docosapentaenoic acid (DPA) levels as well as total n-3 fatty acids. Supplementation of these cells with DHA resulted in increases of DHA and EPA in a dose-dependent manner in addition to a decrease of AA and total n-6 fatty acids (Champeil-Potokar et al., 2004). In another study, synaptic plasma membranes from ApoE-deficient mice contained an increased amount of DHA within PC and phospho tidylserine (PS) classes possibly due to an overcompensation of DHA transport due to the decreased transport of other PUFAs from the ApoE deficiency (Igbovboa, 2002).
76 Animal studies show that aged rat (2124 month) hippocampus (Favreliere et al., 2003) and whole brains (Bacelo-Coblijn et al ., 2003) have significantly less DHA in PE, phosphotidylserine (PS) and pl asmenylethanolamine (PmE) phospholipids classes versus young rat (2 month) brains. Within some phospholipid classes, such as phosphotidylinositol (PI), sphingomyelin and PS, the levels of AA, C16 (palmitic acids), total n-6 and n-3 fatty acids have also been shown to be lower in aged rat (27 month) whole brains in comparison to adult rat (7 month) brains; there were no differences shown specifically in hippocampus (Ulmann et al., 2001). Old rats supplemented with fish oil for one month had restored levels of DHA within ethanolamine phosphoglycerides of whole brains compared to the levels in young control-fed rats (Barcelo-Coblijn et al., 2003) In addition, 3 months of DHA supplementation to adult rats (18 month) resulted in significant in creases of AA, DHA and total n-6 PUFAs in hippocampal phospholipid classes, PE and PmE, in comparison to unsupplemented control fed rats (Favrelier e et al., 2003). Young mice ha ve even benefited from supplementation with fish oil. Puskas et al., (2004) supplemented 4 month old mice for 2 months with cholesterol alone, cholesterol pl us fish oil, or sta ndard chow. Mice that were supplemented with only cholesterol ha d a reduced amount of DHA in whole brain, whereas mice fed both cholesterol plus fish oil had increased DHA in whole brain and retina as compared to mice fed the standard chow. In addition to measuring fatty acid levels in the brain and reti na, this study also analyzed gene expression, specifically peroxisome proliferators-activ ated receptors (PPARs), st erol-regulatory element binding proteins (SREBPs), fatty acid bi nding proteins (FABPs) and inflammatory proteins, all of which will be discussed in the following sections.
77 Many animal studies focus on young animal s and involve n-3 deficiency or supplementation for several generations. Two si milar studies by Ikemoto et al. illustrate that rats fed an n-3 deficient (safflower oil as main fat source) diet for two generations supplemented with either DHA (2001) or EP A (2000) exhibit a re stored level of DHA within the whole brain compared to contro l mice fed an n-3 sufficient diet. The n-3 deficiency led to increased levels of n6 PUFAs, AA and 22:5n-6 within whole brain compared to supplementation with either DHA or EPA (Ikemoto et al., 2000; Ikemoto et al., 2001). Gamoh et al. (2001) de monstrated that rats fed a fi sh oil-deficient diet for 2 generations and subsequently supplemented with DHA showed a sign ificant increase in DHA content and DHA/AA ratio within the cere bral cortex, with no difference between fatty acid levels within the hippocampus. Si milarly, Moriguchi et al. (2001) fed rats either an n-3 adequate or n-3 deficient diet for 2 generations. The F2 generation rats on the n-3 deficient diet were weaned onto the n3 adequate diet at 7 weeks of age and then groups were sacrificed immediately and ev ery 2 weeks thereafter for 8 weeks. Specifically, this study measured the recovery rate of serum, retina, liver and brain of fatty acid levels after 2 genera tions on an n-3 deficient diet, as compared to being raised on an n-3 adequate diet. For the serum, recovery of n-3 fatty acids to levels seen in the n3 adequate rats were accomplished after 2 w eeks of switching diets. The liver, was the most rapidly recovering of all the tissue m easured in that liver n-3 levels were comparable to those of the n-3 adequate gr oup after only 1 week. In contrast, the brain had a slower recovery rate, such that even after 8 weeks of being fed an n-3 adequate diet, levels of DHA were stil l not comparable to the mice originally fed the n-3 adequate diet. With a faster recovery than the brain, the retina had full recovery of DHA in the rats
78 that had been on the n-3 adequate diet for 8 weeks The authors analyzed the recovery rate for each of the tissues in order to normalize the data They found that the serum actually had the slowest rec overy rate per tissue weight, which was most likely a reflection of the low content of DHA in the serum compared to the other tissues. The brain and retina had higher rec overy rates than the plasma w ith respect to tissue weight. Because of their higher recove ry rates yet longer recovery time within the brain and retina; this indicates an inabil ity in uptake or delivery of fatty acids to rapidly replenish these areas of the nervous system (Moriguchi et al., 2001). Even without generations of n-3 deficien cy, many animal studies show significant changes in fatty acid composition within plas ma or brain lipids following an n-3 deficient diet. In general, r odents fed an n-3 defici ent diet exhibit decreas es in total n-3 PUFAs including DHA and EPA as well as increases in total n-6 PUFAs such as AA; diets enriched in n-3 fatty acids show the opposite results (Jensen et al., 1996; Minami et al., 1997; Suzuki et al., 1998; Wainwright et al ., 1999; Lim & Suzuki, 2000; Petursdottir et al., 2002; Aid et al., 2003; Du et al, 2003; Ch ampeil-Potokar et al., 2004). To date, there is minimal work being performed using AD transgenic mice to study effects of n-3 deficiency or supplementation. Hashimoto et al. (2002) did, however, experiment with rats infused with the A 1-40 peptide into the left ventricle to gain similar results that could be seen from an AD transgenic model. Ra ts infused through osmotic minipumps with the A 1-40 peptide for 3 weeks and supplemented with DHA for 12 weeks prior to A infusion and continued for 3 weeks after infu sion showed a decreased level of AA lipid content within blood plasma, the cort ex, and the hippocampus compared to unsupplemented A -infused rats. There was also a higher level of DHA lipid content
79 detected in blood plasma, the cortex and the hippocampus of the A -infused group with DHA supplementation. The authors suggest that the decrease in AA for the DHA supplemented group could be due to a physical replacement of AA within blood plasma, the cortex, and the hippocampus by DHA or through an unexplained mechanism to decrease AA synthesis by incr eased DHA synthesis. The study also found significant positive correlations between cortical and plasma DHA levels as well as between hippocampal and plasma DHA levels. These two associations suggest a high degree of integration of DHA into the cortex and hippo campus from the plasma. Rats that were infused with the A 1-40 peptide not supplemented with DHA exhibited decreased plasma DHA versus vehicle rats (Hashimoto et al., 2002). This result is sim ilar to that seen in patients with Alzheimers disease where plas ma levels of DHA are decreased compared to age-matched non-demented control subjects (Conquer et al., 2000; Tully et al., 2003). A more recent study by Hashimoto et al. ( 2005) used rats fed a fish-oil deficient diet for 3 generations after which they were infused with either A 1-40 or vehicle. Rats from each group were administered either DHA or vehicle and fatty acid levels in the plasma, cerebral cortex and hippocampus were determined. The results were similar to those in the authors previ ous work, such that there was an increase of DHA and a reduction in AA content in all three regions analyzed for groups administered DHA irrespective of A infusion. In addition to fatty acid levels, the authors also measured oxidative stress in the cerebral cortex a nd hippocampus. TBARS and ROS levels were increased in the A -infused group compared to the rema ining three groups. These results suggest that infusion of A 1-40 is sufficient to produce oxi dative damage to specific regions of the brain, so that administrati on of DHA reduced the levels of oxidative
80 damage to those of the vehicle group. Th e authors suggest that the reduction in AA content in the brain played a role in the ab ility for DHA to correct the oxidative damage generated by the infusion of A 1-40 (Hashimoto et al., 2005). Recently, Calon et al. (2004) fed 17 m onth old non-transgenic (Tg-) and Tg2576 APP mice for ~3 months a control diet, n-3 de ficient safflower oil di et and n-3 deficient supplemented with DHA. Analysis of the front al cortex of the APP mice showed that the n-3 deficient diet resulted in a decreased amount of DHA cont ent compared to the control and DHA-supplemented diets. Also, the DHAsupplemented APP frontal cortex showed a decrease in the amount of AA compared to the n-3 deficient mice. Within the Tgmice, the DHA-supplemented group had an increased amount of DHA content and decreased AA content in the frontal cortex versus the n-3 deficient Tgmice. Aside from frontal cortex fatty acid conten t, the study by Calon et al. (2004) also measured dendritic spine pathology. Fractin is a caspase-cleaved fragment of actin that is labeled within dendrites of ta ngle-bearing neurons. The fract in/actin ratio was measured in the cortex of aged (22 month) APP mice in comparison to age-matched non-transgenic mice. The APP mice had an increased rati o in the cortex however, the level of synaptophysin staining did not differ between the groups and the level of drebrin (dendritic spine actin-regulat ing protein), was decreased in APP mice. In comparison, temporal cortex from AD brains showed d ecreased content of dr ebrin and synaptophysin in comparison to control brains. To test th e hypothesis that DHA dire ctly plays a role in regulating caspase activity in the synapses of dendrites, Calon et al. (2004) measured the fractin/actin ratio in the cortex of APP and Tgmice and found that DHA supplementation reversed the increased ratio of the control and n-3 deficient APP mice.
81 In order to determine if the caspase activ ity was induced by preor post-synaptic changes, drebrin (postsynaptic marker) was compared to SNAP-25 (presynaptic marker). The n-3 deficient APP mice exhibited a signifi cant decrease of drebrin in the membrane fraction with an increase within the cytosoli c fraction, leading the authors to conclude that drebrin is not a caspase substrate. The cy tosolic increase of dreb rin is consistent with its retention by intact F-actin filaments. The APP control and DHA-supplemented groups appeared to have opposite effects of drebri n versus the n-3 defi cient diet group. The postsynaptic marker, SNAP-25 did not differ between any of the groups indicating that caspase activity is not altered postsyn aptically by n-3 deficiency or DHA supplementation. In order to account for th e presynaptic change associated with DHA supplementation, neuronal loss was examined. Consequently, no differences in neuronal number were found in cortex or hippocampus between APP or Tgmice despite dietary intervention (Calon et al., 2004). Enrichme nt of mouse neuroblastoma cells with DHA showed a protection against apoptosis via phos photidylinositol 3-kinase (PI3-K) (Akbar & Kim, 2002). In order to determine this effect in an AD mouse model, Calon et al. (2004) measured the levels of a protein subunit of PI3-K protein, p85 The n-3 deficient APP mice showed a significant reduction on p85 in the cortex; DHA supplementation resulted in an increased amount of p85 compared to the n-3 deficient mice, but still significantly less that the controls. From this the authors suggest th at n-3 depletion in APP animals disables the PI3-K pathway to prevent caspase activation which leads to postsynaptic pathology similar to that seen in AD subjects. Calon et al. (2004) also investigated oxidative stress markers following n-3 depletion and DHA supplementation, as m easured by dinitrophenylhydrazin (DNPH)
82 derivatized carbonyls in the co rtex. The n-3 deficient gr oup exhibited an increased amount of carbonyls compared to the APP control and DHA-supplemented groups. The authors conclude that caspase-mediated actin cleavage and loss of drebrin in an APP mouse model exhibiting AD pathology is regula ted by n-3 PUFAs within the brain. Also, DHA, specifically, is involved in maintaini ng the PI3-K pathway to decrease apoptosis; however, there were no differences in neurona l number found between any of the diet groups. Despite the lack of a DHA effect on the post-synaptic marker, SNAP-25, Calon et al (2005) analyzed the eff ect of DHA on NMDA receptors within the same groups of APP transgenic mice Both n-3 deficient and DHA supplemented Tg+ mice had reduced levels of NR2A, NR2B receptors, as well as reduced calcium/calmodulin-dependent kinase II (CaMKII) within th e cortex compared to cont rol Tg+ mice. Although, the DHA supplemented Tg+ mice had slightly elevated levels of NMDA receptor subtype 2A (NR2A), NMDA receptor subtype 2B (NR2B) and CaMKII, there was no significant difference between the dietary groups. Howeve r, the authors state that the addition of DHA resulted in a protective effect with resp ect to the loss of NR2A, NR2B and CaMKII (Calon et al., 2005). The authors offer no cl ear explanation as to how the n-3 depletion and DHA addition affected NMDA receptors. However, evidence shows that an increased number of NMDA receptors can le ad to increased learning and memory in transgenic mice (Tang et al., 1999). Conversely, a reductio n of NMDA receptor subtype 1 (NMDAR1) in the hippocampus of mice led to impairment of spatial learning (Tsien et al., 1996). The previous studies provide s upport for the improved memory that resulted from DHA supplementation as will be discussed in a later section. However, Lesn et al.
83 (2005) showed that activation of NMDA receptors in cell culture led to an increase of calcium influx, thereby activating Ca MK and increasing production of A peptides. Although the present study examined in vitro conditions, the implication exists that an alteration of NMDA receptor activation in a system that is prone to A production will result in over-secretion of toxic A peptides, ultimately leading to cognitive impairment (Lesn et al., 2005). Lim et al. (2005) reported alterations of A within 17 month old Tg2576 APP mice that were fed a control diet, n-3 defi cient safflower oil diet and n-3 deficient supplemented with DHA for ~3 months. The DHA supplemented mice had reduced levels of insoluble A as compared to the n-3 deficien t group, but there was no difference with respect to control mice. There we re also no differences in soluble A levels between any of the three groups. Despite the lack of eff ect of soluble A the DHA supplemented group had an overall reduction in A plaque burden compared to the n-3 deficient mice. However, there was no comparison between the DHA supplemented group and the control group w ith respect to plaque burde n. In addition, there was no difference in soluble A insoluble A or A 40 between the DHA supplemented mice and the control APP mice, indicating minimal affects on A by DHA. Human Studies Involving Fatty Acids. There have been numerous human trials examining fatty acid composition within blood or brain tissue of normal elderly and elde rly subjects with deme ntia. Payet et al. (2004) performed a study using non-impaired el derly subjects to simply measure the
84 incorporation of fatty acids into eryt hrocyte membranes and plasma. DHA was supplemented for 9 months and resulted in a significant increase in total PUFAs within the blood mainly due to the increase of DH A and AA in the plasma and erythrocyte membranes. The observed increase of AA coul d have been the result of an increased amount of this fatty acid within the DHA s upplementation diet compared to the control groups diet. There was no follow-up to i ndicate changes in c ognition or everyday function (Payet et al., 2004). In a simila r study, Boston et al. ( 2004) administered EPA for 12 weeks to AD patients. This study assessed cognitive improvement, as well as erythrocyte membrane compositi on of fatty acids both before and after treatment. At baseline, erythrocyte membrane fatty acid levels did not differ between AD and nondemented control patients. In short, th e 12 week period of EPA administration was unsuccessful at altering any of the cognitive measures. However, treatment with EPA did result in significant increases in total n3 fatty acids, specif ically EPA and 22:5n-3 (DPA), compared to baseline le vels (Boston et al., 2004). In an earlier study, Tully et al. (2003) reported that low pl asma levels of total n-3 PUFAs (specifically DHA) were associated w ith an increased risk of developing AD. This is in opposition to the above study by Boston et al. (2004) who found no difference in blood fatty acid levels between AD and non-de mented subjects. The increased risk for AD with low plasma n-3s as reported by Tully et al. (2003) could be a ttributed to a lack of protection against cardiovascul ar disease, an increased pr oduction of pro-inflammatory cytokines, or a lack of nervous system home ostasis due to low plasma DHA and/or total n-3 PUFA content that would nor mally protect against these pat hologies. Analysis of the MMSE scores of AD versus non-impaired subjec ts revealed that blood serum levels of
85 DHA and total saturated fat were significant predictors of AD (Tully et al., 2003). Engelhart et al. (2002a) found an associati on between risk of developing AD and an increased intake of total fat, saturated fat, transaturated fat and cholesterol. However, there were no significant associations between intake of total PUFA, n-6, or n-3 and the risk of AD (Engelhart et al., 2002a). Prior to much of the above information regarding the association between AD risk and n-3 inta ke/membrane content, Yehuda et al. (1996) showed that subjects with existing AD that were supplemented with SR-3 (which is a mixture of ALA and LA to yield a ratio of 4 to 1 of n-6 to n-3 fatty acids) had improved behavior, including enhanced short-term a nd long-term memory, more cooperation, better mood, increased appetite and more organization skills. This paper suggests that many of the improved symptoms offer benefit to the patie nt as well as the caregiver. The authors do not propose a mechanism of action for th is compound, but suggest that changes in neuronal membranes due to SR-3 could lead to increased function of the neuronal system thereby yielding the above impr ovements (Yehuda et al., 1996). Fatty acids, specifically DHA, rapidly in corporate into numerous different cell types, primarily into the phospholipids la yers of their plasma membranes and mitochondria (Stillwell & Wassall, 2003). Most human AD studies involving analysis of fatty acids measure plasma levels or post-mo rtem brain tissue phospholipid levels. Early studies indicated that post-mort em brain tissue sections from subjects with AD or some other type of dementia had decreased le vels of n-3 PUFAs, including ALA, DHA and EPA; decreases in n-6 PUFAs including AA and LA were also observed. In 1998, Corrigan et al. reported lower levels of total n-3 PUFAs, sp ecifically decreased ALA in the parahippocampal phosphotidylcholine (PC) fractions from AD brains. The authors
86 suggest that the decrease in ALA is most likel y due to a deficiency within the PC fraction to properly incorporate ALA into the phos pholipid. Because no other PL fraction contained a significant decrease of ALA, th e difference could not be accounted for by a change in dietary intake or liver synthesis of FAs between AD and control subjects. This study also found decreases in n-6 PUFA s, including AA, mainly within the phosphotidylethanolamine (PE) fraction of phos pholipids of AD brains compared to agematched controls (Corrigan et al., 1998). Prasad et al (1998) also found a decreased level of AA in PE phospholipids, specifical ly within parahippocampal gyrus and the inferior parietal lobe of AD brain tissue compared to normal elderly brain tissue. DHA was only found to be significantly below norma l levels within the PE fraction of the parahippocampal gyrus and PC fraction with in the cerebellum. There were, however, decreased levels of total fa tty acids in phospholipid classe s within the parahippocampal gyrus and the inferior pariet al lobe. Also measured we re phospholipids within the superior and middle temporal gyri, where the only difference was a decrease of oleic acid in AD brain tissue in comparison to control brain tissue. This study also determined that there was a significant increase of senile pla que density within all three areas measured, parahippocampal gyrus, inferior parietal lobe and the superior and middle temporal gyri. The cerebellum contained only di ffuse plaques. One of th e possible links between the presence of senile plaques and the decrease s of various fatty acids could be due to changes in the biosynthesis or degradation of the membrane phospholipids. The overall conclusion from Prasad et al (1998) was that the changes in phospholipids content were due to the oxidative cascade present in AD brai n tissue. Previous studies have indicated that the A peptide is a prominent source of free radical generation. These free radicals
87 could be the source of oxidative damage to the phospholipids classes resulting in decreased levels of many of the fatty acids present in the membrane PLs. Additional mechanisms that could impact membrane PL composition are neuronal loss, changes in ion channel function or cellular biosynthesis mechanisms all due to the presence of A plaques or NFTs (Prasad et al., 1998). In contrast to both of the above studies, Skinner et al (1993) found an increase in DHA in white matter of the parietal cortex from AD patients, in addition to an overall reduction in n-6 fatty acids in white matter of the frontal and parahippocampal cortices. However, this study also reported increases in adrenic acid (22:4n-6) in grey matter from the parietal, frontal and parahi ppocampal cortices, resulting in an increase of total n-6 fatty acids in the grey matter of the frontal cortex. The authors s uggest that the changes in fatty acid composition between white and gr ey matter of the brain in AD patients could be due to an irregularity in the transportation of essential fa tty acids to the brain (Skinner et al., 1993). Recent evidence shows that not only brain tissue but plasma levels of fatty acids are different between AD subjects and norma l elderly control s ubjects. AD subjects exhibit decreased plasma levels of DHA, EPA and total n-3 PUFAs in comparison to normal elderly controls (Conquer et al., 2000; Tully et al., 2003). Tully et al. (2003) determined no significant difference in total n-6 PUFA content in plasma levels; Conquer et al. (2000), however, found that AD subjects had a significantly lower n-3 to n-6 fatty acid ratio within plasma with a significant increase in the le vel of total n-6 PUFAs. The authors suggested that the mechanism by which these fatty acid levels become lower in AD subjects is simply a decrease in the dietar y intake of these fatty acids (Conquer et al.,
88 2000). The authors also suggested that the lower levels could be due to an increase in the breakdown of DHA within the brain or a decr eased amount within th e plasma PC content which is the main source for plasma DHA. DHA content was not measured within the liver, therefore decreased liver synthesis of DHA could also account for the difference in plasma levels. Similar to Conquer, Heude et al. (2003) found a direct association between total n-6 fatty acids in erythrocyte membranes and risk of cognitive decline in a group of subjects during a 4 year study. Li kewise, this study also determined that subjects with increased levels of n-3 fatty acids had a reduced risk of cognitive decline. However, this study did not in clude any information regardi ng dietary intake, therefore the reasons for the changes in fatty acid leve ls is unknown and could be due to alterations in liver metabolism, intake or transportation of essential fatty acids throughout the body. Incorporating adequate amounts of omega-3 fatty acids into daily dietary intake is not plausible for the population as a whole. Metcalf et al. designed a study to survey healthy individuals for their n-3 consump tion by providing foods enriched with n-3 PUFAs (2003). This study showed that pe ople could consume adequate amounts of n-3 fatty acids in their diet to decrease plasma compositions of AA, LA and total n-6 fatty acids, as well as increase plas ma levels of EPA, DHA and total n-3 fatty acids (Metcalf et al., 2003). However, trying to incorporate everyday foods with n-3 fatty acids will probably never occur. Kang et al. (2004) provided an altern ative to enriching foods to obtain n-3 PUFAs. As previously mentione d, mammals can not convert n-6 fatty acids into n-3 fatty acids; however integrating a fat-1 gene from C. elegans into a mouse allows this conversion to occur. Transgenic mice carry ing this gene fed an n-3 deficient/n-6 rich diet had a higher percentage of n-3 fatty ac ids in tissues (muscle, heart, brain, liver,
89 kidney, lung and spleen) compar ed to non-transgenic mice fed the same diet. The n-6 to n-3 ratio within the transgenic mice was 1:1 whereas the non-transgenic ratio was 2050:1 (Kang et al., 2004). This implies that diet ary intake does not ha ve to be altered in order to increase incorporation of n-3 fatty acids into tissues and plasma membranes. Yancy et al. (2004) and Stern et al. (2004) both show that a lo w fat diet results in less weight loss than a low carbohydrate diet at 6 months after dietary intervention. The low carbohydrate diet allowed for unlimited amounts of animal foods which implies increased protein and fat intake. This diet resulted in significantly more adverse side effects versus the low fat diet. These side effects included increased headaches, muscle cramps, diarrhea, general weakness a nd rash (Yancy et al., 2004). The low fat group was encouraged to decrease their fat intake to less that 30% of th eir daily energy. This group over one year lost the same amount of we ight as the low carbohydrate group; neither group had any significant adverse side effects (Stern et al., 2004). Neurochemical Effects of n-3 Fatty Acids. Incorporation of fatty acids from fish oil into neuronal membranes has been shown to increase neuronal sensitivity to oxidation. Because of the high number of double bonds within DHA, it is more likely to be oxidized than other fatty acids, such as AA. Peroxidation of PUFAs can lead to damage to genes, membrane lipids and enzymatic proteins. Despite this, PUFA s upplementation has been shown to correct the effects of oxidative stress by decreasing fr ee radicals within the brain (Yehuda et al., 2002). Specifically, EPA can reduce reactiv e oxygen species (ROS) produced by lipid peroxidation by inhibiting th e phospholipase A2 enzyme and stabilizing membrane
90 structure. Lipid peroxidation products can contribute to neurodegeneration by inhibiting glutamate or glucose transpor t within the brain. Supplemen tation with essential fatty acids for 30 days can decrease production of TBARS (end product of lipid peroxidation), xanthine oxidase (source of ROS) and nitric oxide (gaseous free radical) as well as increase superoxide dismutase (enzyme antioxi dant) within the corpus striatum of adult rats (Sarsilmaz et al., 2003). In addition, enhanced activati on of catalase and glutathione peroxidase, enzymes that decrease ROS, we re evident in macrophages from ApoE-KO mice fed a diet supplemented with fish oil. These macrophages also had reduced levels of ROS and superoxide anion af ter treatment with fish oil back to normal levels (Wang et al., 2004). Increased production of lipid pe roxides occurs in cerebellum and some cortical regions with age and in certain neurodegenerative disease, such as AD. Specifically in AD, increased lipid peroxi dation markers are detected within CSF, urine, plasma and the brain including all cortical lobes, hipp ocampus and cerebellum (Montine et al., 2004). The free radical induced per oxidation of AA yields a prostaglandin F2-like compound named F2-isoprostane (F2-IsoP). Peroxidation of EPA produces F3-IsoPs and per oxidation of DHA forms F4-IsoPs also known as F4neuroprostanes (F4-NP) because of their presence sp ecifically within neurons (Roberts II et al., 1998; Nourooz-Zadeh et al., 1999). In some regions, brains of AD patients undergo significantly more lipid peroxidation than normal aging controls (Montine et al., 2004). Using an indicator that could refl ect the amount of DHA oxidation within the brain could be used as a marker for detecti ng AD. Multiple studies have investigating the detection of DHA peroxidation within AD br ains post-mortem and found significantly more F4-NPs in occipital and temporal lobes in addition to the cerebral cortex (Nourooz-
91 Zadeh et al., 1999; Reich et al., 2001). This provides a sensitive indicator for oxidative damage to neurons because of the high concentration of DHA within the neurons. The peroxidation of DHA could be due to the out come of the AD pathology or simply a predisposition to increase vulnerability to developing AD. However, following brain ischemia-reperfusion, DHA is released from the phospholipids membrane via PLA2 and induces neuroprotection via s uppression of pro-inflammatory cytokines, anti-apoptotic genes and formation of a novel docosanoid, NPD1 (Lukiw et al., 2005). NPD1 (10, 17Sdocosatriene) is a bioactive DHA-derived lipid mediator. Lu kiw et al. (2005) found that in vitro production of A was attenuated with addition of DHA to the medium. The addition of DHA also produced an increase in NPD1 in these culture d cells, suggesting that the attenuation of A peptide release could be partia lly due to the appearance of NPD1. The authors also determined a dose-dependent increase of NPD1 by sAPP in conjunction with DHA, suggesting a positiv e feedback regulation between sAPP and the DHA/NPD1 pathway to protect cells agains t neuronal damage. La stly, Lukiw et al. (2005) found that NPD1 protected cultured cells from A 42-induced apoptosis by upregulating anti-apoptotic genes, Bcl-xl, Bcl-2 and Bfl-1(A1). Some research indicates that fish o il supplementation have anti-inflammatory actions, including decreases of some pro-inflammatory cytokines (IL-1 IL-2, IL-6, TNF) and increases of anti-inflammatory cytokines (IL-10, TGF). Few studies have focused on the effects of DHA and/or EPA specifically on inflammatory measures. Tomabe et al. (2000) showed that dietary DHA, but not EPA, supplementation resulted in a decrease of CD4-positive T lymphocytes to reduce ear swelling of mice with contact
92 hypersensitivity. In addition, DHA decreased the expression of pro-inflammatory IFN, IL-6, IL-1 and IL-2 mRNA within the ears (Tomobe et al., 2000). In contrast to the lack of effect on inflammatory cytokines by EPA by Tomobe, Komatsu et al. (2003) found that a high dos e of EPA inhibited nitric oxide (NO) production by LPS-activated peritoneal macr ophages from mice. However, EPA did not have any effect on the inducible NO syntha se (iNOS) protein. Treatment with DHA resulted in a significant i nhibition of NO production and i NOS expression by murine peritoneal macrophages dose-dependently. Th e authors showed that the DHA treatment suppressed activation of the transcri ption factor nucl ear factor (NF) B most likely through inhibition of intracellular peroxides induced by IFNand LPS, which lead to the suppression of NO production and iNOS expression. Within the macrophages, DHA treatment resulted in an up-regulation of in tracellular glutathione (GSH); lowering GSH levels reversed the effects indu ced by DHA on NO production and NFB. Therefore, the authors suggested that DHA inhibite d NO production via suppression of NFB and mediated by an up-regulation of GSH (Komatsu et al., 2003). Peritoneal macrophages from mice fed a control diet supplemented with menhaden fish oil for 15 weeks had decreased production of IL-1 and TNFin addition to a reduction of mRNA for IL-1 and TNF(Renier et al., 1993). Billiar et al (1988) found an anti-inflammatory response exhibited by decreased IL-1 production by Kupffer cells of rats fed a fish oil supplemented diet for 6 weeks. N-3 PUFA supplementation, mainly DHA and EPA, in older women (51-68 years) for 3 months resu lted in decreased production of all proinflammatory cytokines measur ed in the blood, including IL-1 TNFand IL-6
93 (Meydani et al., 1991). Speci fically, IL-2 expression was reduced as a result of a decrease in helper-T cells, w ith fish oil supplementation leading to an anti-inflammatory result. In addition, there was an increase in the number of suppressor T cells detected that aid in the anti-inflammatory re action by secreting IL-10 and/or TGF(Meydani et al., 1991). In some contrast to the above studies, Pu skas et al. (2004) reported a combination of pro-inflammatory and anti-inflammatory eff ects within mice enriched with fish oil and cholesterol. Puskas et al. (2004) fed mice a chol esterol-rich diet or a cholesterol-rich plus fish oil enriched diet for 2 months beginning at 4 months of age. Within whole brain of mice fed the cholesterol-rich diet with fi sh oil, the pro-inflammatory cytokines (IL-6 and TNF) were down-regulated in addition to an anti-inflammatory cytokine, IL-10. Whereas within the retina, for the same gr oup, an opposite result was seen, so that there was an up-regulation of TNF. The authors suggested that the eye was more sensitive to the high fat content of the diet and ther efore and increased pr o-inflammatory gene expression in which the addition of fish o il was not able to correct. However, the addition of the fish oil led to an overall anti -inflammatory result within the brains of mice fed a diet enriched with choleste rol plus fish oil, so the authors sugges tion is unclear. In addition, the positive effects expected by the fish oil could not be seen without including a group that was only administered fish oil wi thout the addition of excess cholesterol. Without this group, it is difficult to draw c onclusions without chol esterol confounding the results. Additional studies find that n-3 supplement ation results in net pro-inflammatory reactions. Lipopolysaccharide-stimulated macrophages from mice given fish oil
94 exhibited an increase in TNFsecretion as well as a decrease of IL-10 secretion (Petursdottir et al., 2002). Liver phospholipid content showed a significant increase of n3 PUFA content and a decrease of n-6 PUFA content in mice fed the fish oil supplement. The authors propose that the increased TNFsecretion was mediated by a decrease in prostaglandin production. As discussed prev iously, AA and EPA can form prostaglandin intermediates; AA results in PGE2 whereas EPA synthesis can lead to PGE3 production. EPA and AA both compete for prostaglandin sy nthesis so that incr eased amounts of n-3 fatty acids (e.g. EPA) displace production of PGE2 by production of PGE3. However, PGE3 is not as mitogenic or inflammatory as PGE2 (Bagga et al., 2003). An additional study reported an invers e relation between PGE2 production and TNF production in LPSstimulated macrophages from mice fed a diet with a high n-3 to n-6 ratio (Watanabe et al., 1993). The authors suggested that PGE2 was acting as a negative feedback effector on TNFproduction; decreased TNFproduction from n-3 enriched macrophages concomitantly with increased production of PGE2. However, no established mechanism was suggested to explain the inverse relationship. It is therefore difficult to determine the mechanism responsible for this pro-inflammato ry response from fish oils. In support of the findings by Petursdottir, Wallace et al. (2003) found that fish oil supplementation to healthy adults (18-39 years of age) resulted in a significant increase of IL-6 expression in blood mononuclear cells. There were no othe r changes evident in any of the other cytokines (TNF, IL-1 IL-2, IFN) or inflammatory markers (Tand B-cell lymphocytes). The authors concluded that the dosage was not high enough to induce changes in additional cytokine producti on (Wallace et al., 2003). Overall, the mechanisms that PUFAs use to induce changes in the inflammatory cascade are
95 unknown, but it appears that dosages and durati on of fish oil enrichment highly affect cytokine production positively and negatively. DHA has also been shown to have effects on the cholinergic system. Increased aging results in decreased ACh levels in hippocampus, striatum (Ikegami et al., 1992), cingulate cortex (Baxter et al ., 1999) and pyramidal neurons of the cerebral cortex (Casu et al., 2002); in addition ther e is an age-relate d decrease of DHA incorporation into membrane phospholipids within the whole br ain, and specifically the hippocampus (Bacelo-Coblijn et al., 2003 l; Favreliere et al., 2003). The mechanism whereby DHA affects the cholinergic system is not we ll understood. A study by Jones et al. (1997) showed that an intravenous injection of a cholinergic agonist, arecoline, resulted in increased incorporation of DHA into memb rane phospholipids and microsomal fractions within the rat brain af ter intravenous infusion of DHA. There was no change in response to an injection of saline or in combination with palmitic acid. The possible pathway involves the activation of phospholipase A2 by arecoline to indu ce the release of DHA from the phospholipid PE, the main storage site of DHA (Jones et al., 1997). This suggests that the reduction in ACh levels leads to decreased DHA incorporation into neuronal membranes within the brain during aging. Multiple studies illustrate the effect of DHA supplementation on the agedependent dysfunction of the cholinergic sy stem. PET scanning of the somatosensory cortex of aged monkeys fed a diet enriched with DHA showed a si gnificant increase in their regional cerebral blood flow (rCBF) compared to contro l fed monkeys (Tsukada et al., 2000). Previous work has demonstrated that the cholinergic system induces an increase in rCBF so that the addition of a cholinergic antagonist re sulted in decreased
96 rCBF. Overall, the authors propose that the ch ange in the rCBF is due to the ability of DHA to modulate neuronal activation via th e cholinergic system; specifically they speculate that DHA acts directly on cholinergic receptors to facilitate signal transduction and increase rCBF (Tsukada et al., 2000). Additional rodent studi es also show that dietary supplementation of DHA for 3 months to aged (18 mo nth) rats reverses an agedependent cholinergic dysfunction so that ba sal ACh levels are similar to those of young rats (Favreliere et al., 2003) In stroke-prone spontan eously hypertensive rats, a significant increase in basal ACh levels in hippocampus and cortex is seen with the addition of DHA to the diet. Under contro l conditions these mice exhibit a dysfunction of the cholinergic system; DHA supplementati on reversed that dysfunction (Minami et al., 1997). However, Aid et al (2003) found that rats fed an n-3 deficient diet also increased basal ACh within the hippocampus co mpared to control fed rats. The authors suggested that this increase could be due to an increased release or decreased catabolism of ACh within the synaptic cleft due to changes in neuronal membrane composition which affects ion channels and therefore ACh levels (Aid et al., 2003). A diet rich in DHA results in increased KCL-induced re lease of ACh from the right ventral hippocampus in aged rats compared to aged rats on a control diet (Fav reliere et al., 2003). The authors suggest that this re lease could be due to an incr ease of synaptic transmission or a decrease of membrane rigidity due to DHA incorporation into the plasma membrane. Similar to these results, an n-3 deficient diet results in decreased KCL-induced ACh release in hippocampus compared to c ontrol-fed rats (Aid et al., 2003).
97 Effects of n-3 Fatty Acids on 2nd Messengers. N-3 and n-6 fatty acids have been show n to impact intrac ellular signaling, including second messenger systems. Briefly, within the visual transduction pathway, G protein (Gt) is activated by coupling to me tarhodopsin (MII), resulting in closure of cGMP-gated channels within the rod outer segment (ROS) a nd induction of the neuronal response to light. Specifical ly, rhodopsin is pho toactivated into M II by absorption of a photon of light and subsequently coupled to Gt This coupling activ ates a cGMP-specific phosphodiesterase (PDE); hydrolysis of cGMP leads to closing of the cGMP-gated channels, changes in the transmembrane potentia l, and initiation of the neuronal response to light (Salem et al., 2001). MII formation has been determined to be a function of phospholipid acyl chain unsaturation; thus DHA-enriched phospholipids from ROS membranes of rats raised on an n-3 adequate diet result in an increase of MII formation (Mitchell et al., 2003). Also, the rate of coupling of MII to Gt is increased in DHAenriched bilayers compared to less unsatura ted phospholipids. Conve rsely, the activity of PDE from ROS membranes of rats raised on an n-3 deficient diet was decreased (Mitchell et al., 2003). Also, the inclusion of cholesterol in ROS membrane bilayers results in a decrease of MII formation a nd PDE activation, in addition to a slower coupling of MII to Gt. The addition of c holesterol in conjunction with DHA, however, results in increases of MII formation and PD E activation as well as a faster coupling of MII to Gt (Litman et al., 2001). The aut hors conclude that DHA promotes optimal functioning of the G-protein c oupled signaling pathway in the retina and suggest that a deficiency of n-3 fatty acids, specifically DHA, could lead to a decreased efficiency in related neurotransmitter functioning a nd the visual signaling pathway.
98 Cyclic AMP and GMP activation have also been implicated as being affected by DHA and/or EPA enrichment. Cultured neonatal rat ventricular myocytes were grown in a DHAor EPA-enriched medium (Picq et al., 1996). Cultures from both fatty acids resulted in increased basal levels of cAMP and more prominently cGMP. Also, stimulation of cGMP was increased in n3 enriched myocytes (Picq et al., 1996). Likewise, cultured rat myocardial cells have been studied to determine the impact of DHA or EPA incubation on adrenoceptor functi on (Grynberg et al., 1995). Incubation with either fatty acid resulted in similar n6 to n-3 ratios within the cultured cells. However, the DHA-enriched cells showed a significantly higher stimulation of betaadrenergic receptors with no change in th e actual number of receptors. Also, DHA enrichment resulted in a decreased affi nity of the beta-receptor for the ligand, dihydroalprenolol, and a decrease in beta -adrenergic induced cAMP production. To account for the enhanced stimulation of beta -adrenergic receptors and decrease of cAMP production, the authors incubated the myocar dial cells with a permeant analogue of cAMP in conjunction with DHA; this result ed in a positive chronotropic response. Overall, they concluded that DHA enrichment of rat myocardial cells results in a positive effect on beta-adrenergic transduction via an increase of cAMP efficiency (Grynberg et al., 1995). In addition to the affect of DHA and/or EPA on cAMP and cGMP, activation of the cAMP pathway affects DHA release fr om rat brain astrocytes (Strokin et al., 2003). Because astrocytes are a source of DHA synthesis for the CNS, the study investigated the role of sp ecific second messengers on DHA rele ase for use in the CNS. Initially, addition of ATP to astrocytes stimulated release of DHA mediated by a Ca2+independent phospholipase A. Additional ne urotransmitters which cause comparable
99 release of DHA were bradykinin, glutamate a nd thrombin. Adenylyl cyclase, protein kinase A (PKA) and cAMP all caused an incr eased release of DHA, which is blocked by inhibitors of adenylyl cyclas e and PKA (Strokin et al., 2003). Despite the data to support the relationship between n-3 fatty acids and second messenger systems, such as cAMP and cGMP pathways, no strong evidence is s hown to illustrate a mechanism causally linking fatty acids to these intracellular systems. In addition to inducing effects on cAMP, DHA has also been shown to affect Ca2+ and Na+ channels. PUFAs have previously be en shown to inactiv ate voltage-gated sodium channels so that cardiac myocytes rema in in a hyperpolarized state, similar to the effect seen by local anesthetics and anticonvu lsant drugs (Xiao et al., 1995). Similarly, Vreudgenhil et al. (1996) showed that inc ubation of rat CA1 pyramidal neurons with either DHA or EPA resulted in a decrease in neuronal excitability via inactivation of voltage-gated sodium and calcium. Monounsatur ated fatty acids, saturated fatty acids and LA were also incubated with the CA1 ne uronal cells, but did not result in any change in neuronal excitability or s odium or calcium currents (Vre udgenhil et al., 1996). This implies that some specific mechanism is u tilized by DHA and EPA to alter voltage-gated sodium and calcium channels. Bonin and Kh an (2000) illustrate that DHA induces a mobilization of calcium from intracellular stores in the endoplasmic reticulum pool by the opening of Ca2+ release-activated Ca2+ (CRAC) channels. Human (Jurkat) T-cells were cultured and incubated with eith er AA, EPA or DHA. DHA induced a dosedependent calcium release intracellularly result ing in a spike with eventual return to baseline concentration; minimal response was seen with EPA and AA. DHA was also shown to cause an increase in intracellular calcium con centration, exhibited by refilling
100 of the ER pools. This effect was independe nt of IP3, an intrac ellular messenger that facilitates the opening of calcium channels on the ER to refill the intracellular pools. However, the effect of DHA was found to be a result of increased influx of extracellular calcium through CRAC channels. Overall, DHA has the ability to in fluence intracellular calcium concentrations via CRAC channels and can therefore influence additional intracellular systems. Gene expression can also altered by the addi tion of a fish oil-rich diet. Fish oil supplementation to two-year old male Wistar rats for 1 month resulted in a significant increase in the expression of transthyretin (TTR) within the hippocampus (Puskas et al., 2003). Barcelo-Coblijn et al (2003) also found an up-regul ation of TTR in whole rat brains after supplementation with a fish oilenriched diet. TTR is a thyroid hormone transport protein that is secreted by hepato cytes into the serum, and by the choroids plexus into the cerebral spin al fluid (CSF). Importantly, previous studies have shown that TTR inhibits aggregation of A within the CSF (Schwarz man & Goldgaber, 1996) and binds to insoluble A to prevent polymerization into plaques (Redondo et al., 2000). AD transgenic mouse models that exhibit A plaques do not exhibit other neuropathologies observed in AD human subj ects, such as neur ofibrillary tangle formation or the neuronal loss characteristic in AD. However, APP transgenic mice express high levels of -secretase cleaved APP (sAPP ) and TTR. However, infusion of an antibody, into the midscapular region, ag ainst TTR into the ag ed (18 month) APP mouse resulted in increased A accumulation, tau phosphorylation, neuronal loss and apoptosis all within region CA1 (Stein et al., 2004). The authors concluded that TTR expression is protective in the APP transgenic mouse against some of the
101 neuropathologies present in AD human subjects, mainly A deposition, neuronal loss and abnormal tau phosphorylation. Likewise, human AD subjects express lower levels of TTR compared to age-matched non-demented control subjects (Stein et al., 2004). Therefore, fish oil supplementation could induce an increase of TTR expression and in turn provide protective benefits agai nst AD neuropathological progression. In addition to alterations in TTR expre ssion, changes in peroxisome proliferatoractivator proteins (PPARs), sterol-regulatory element-bi nding proteins (SREBPs), and fatty acid binding proteins (FABPs) have been observed with fish oil supplementation. Puskas et al. (2004) fed 4 month old mice for 2 months standard chow, a cholesterolrich diet or a combined diet of high choleste rol plus fish oil. There was an up-regulation of PPARand PPARwithin the whole br ain of the mice fed the combined diet compared to standard-fed mice. However, there was also an up-regulation of PPARin the brain of mice fed the cholesterol-rich diet as well. PPARs play an important role in regulating lipid and glucose metabolism and are activated by a diversity of ligands, specifically including PUFAs and fatty acid metabolites. PPARhas been shown to increase fatty acid catabolism, thereby having an overall lipid lowering effect (Schmitz & Langmann, 2005). This effect wa s most likely in response to the increased intake of cholesterol despite fish oil supplementation. However, PPARwas up-regulated within the combined diet group only. PPARregulates genes that cont rol cell proliferation and differentiation and is highly concentrated in adipocytes (Schmitz & Langmann, 2005). The authors offer no direct explanation fo r the changes in PPAR gene expression between the diets (Puskas et al., 2004). There was also an up-regulation of retinoid X receptor(RXR) in the whole brain of the mice fed the combined diet of cholesterol
102 and fish oil versus the standard diet. RXRis increased during neuronal maturation, implicating an important role during development. In addition, RXRis also implicated in working memory function related to the front al and perirhinal cortices (Wietrizych et al., 2005). However, Puskas et al. (2004) did not behaviorally eval uate these animals, thereby not determining if the increase in RXRis sufficient to provide cognitive benefit. SREBPs are essential in maintaining fa tty acid and choleste rol homeostasis. Within the different isoforms, the main role for SREBP-1a is to regulate lipogenesis and cholesterol synthesis proteins while SREBP-2 regulates ge nes involved in cholesterol metabolism. In Puskas et al (2004), mice fe d the cholesterol-rich diet had increased expression of both SREBP-1a and SREBP-2 with in whole brain and retina as compared to the standard fed mice. In contrast, the a ddition of the fish oil to the cholesterol diet returned these levels back to those found fr om the standard-fed mice. Therefore, the addition of the fish oil attenuated the affect induced by the high cholesterol within both brain and eye of these mice. Lastly, Puskas et al. (2004) measured di fferent forms of FABP within the brain and eye. Of the four types of FABPs measur ed (brain, epidermal, liver, and heart), the cholesterol-rich diet did not al ter any of these proteins. However, addition of fish oil to the cholesterol-rich diet induced an incr eased expression of both epidermal FABP and heart FABP in both the whole brain and re tina. Because FABPs regulate fatty acid content in various tissues (Verekamp & Zi mmerman, 2001), the authors suggested that the FABPs compensated for the addition of th e fish oil which caused an accumulation of fatty acids in the retina and the brain (Puskas et al., 2004).
103 Effects of n-3 Fatty Acids on LTP. Long-term potentiation (LTP) has been associated with learning and memory functions. Increased aging is also associated with a decline of learning and memory and therefore also with an impairment of LTP. The mechanism involved in the agedependent impairment of LTP is not know however the possible mechanism could involve a decrease in the rele ase of glutamate which normally assists in maintaining LTP within the perforant path of CA1 of the hippocampus (McG ahon et al., 1999; Martin et al., 2002). As previously mentioned, increase d aging is associated with a reduction of PUFA concentration within the neuronal me mbranes which leads to a decrease of membrane fluidity. This fact or could contribute to the decr eased release of glutamate and impaired LTP seen in aged individuals. Supplementation of PUFAs including AA+GLA, EPA (Martin et al., 2002) and DHA (McGahon et al., 1999) to aged ro dents results in a reversal of the age-related impairment of LTP. Specifically, the effect by DHA is proposed to involve increasing membrane flui dity to affect neurotransmitter release (McGahon et al., 1999). Within the dentat e gyrus (DG), DHA induced an excitatory effect on the excitatory post -synaptic potentials (EPSPs), but did not affect LTP within this region (Itokazu et al., 2000). The aut hors suggested that th e effect of DHA on the EPSPs is mediated through potassium cha nnels. DHA has previously been shown to block potassium channels which is abolished by the presence of zinc (Poling et al., 1995). Because zinc is more abundant within DG, the effect of DHA on potassium channels to cause a stimulatory effect on EPSPs should ha ve been inhibited in DG. However, the authors saw a stimulation of EPSPs with DHA. Thus, their explanation of results is flawed and is contradictory to the role of pot assium channels as ha ving inhibitory effects
104 on neurons. DHA, in combination with zinc in the DG, should have resulted in an overall inhibitory effect on EPSPs, not excitatory, if th eir explanation was correct. In contrast to the DG, within region CA1 the addition of DHA induced a dose-dependent inhibitory effect on EPSPs (Itokazu et al., 2000). Prev ious work has shown that DHA promotes the opening of NMDA receptors via a DHA-binding s ite on the receptor; however addition of an NMDA antagonist does not affect the EPSP in DG or CA1 resulting from supplementation with DHA (Itokazu et al., 2000). Young et al. (2000) also showed that DHA can act independent of the post-synaptic NMDA receptor. The authors suggest that DHA blocks sodium channels to stabilize presynaptic membranes at the resting membrane potential in order to decrease glutamate release which would lead to an inhibition of EPSPs (Young et al., 2000). Because of maintena nce of the resting membrane potential, Ca2+ would not be released from intracellular stores to bind to glutamate autoreceptors, thereby decreasing a release of glutamate from the presynaptic cell. Overall, n-3s, specifically DHA, have been shown to have an inhibitory affect on EPSPs primarily through closures of sodium and potassium channels. Behavioral Effects of n-3 Fatty Acids in Animals. Numerous animal studies show associat ions between n-3 or n-6 intake and behavioral performance, as seen in Table 1 below. In reference to general locomotor activity, Carrie et al. (2000) showed that sardine oil supplementa tion for 2 generations resulted in an increase of activity in F2 young (7-11weeks) mice, where adult (9-10 months) and aged (17-19 months) mice showed no difference with respect to palm oil fed animals. In contrast, Chalon et al. (1998) also found that fish o il supplementation for 2
105 generations decreased activity in young (2 month) rats. Within the same study, there were no differences in entries or time in open arm for the elevated plus maze task with young rats either fed a diet enriched with fish oil or control diet (Chalon et al., 1998). Rats that underwent surgery that occlude s the common carotid arteries (2VO) had impaired spatial learning ab ilities. De Wilde et al. ( 2002) showed that n-3 PUFA administration from 3 weeks to 4 month of age to 2VO rats did not result in any differences seen in the elevated plus maze task at 5 months of age in comparison to vehicle control rats. Changes in learning and memory performan ce are also associated with n-3 dietary intake. Ikemoto et al. (2001) weaned 1-month old rat pups onto either a diet supplemented with DHA or safflower oil fr om dams fed a safflower oil rich, n-3 deficient, chow. Between 11 and 18 weeks of age the rats were tested in a brightness discrimination task. Rats that received DHA supplementation performed similarly to rats not supplemented with DHA (Ikemoto et al., 2001). This shows that the cognitive impairment measured by a brightness discrimi nation task caused by an n-3 deficient diet is not reversible. Impairment present in stroke-prone spontan eously hypertensive (SHRSP) rats was also reversed for a passive avoidance test (Minami et al., 1997). In that study, SHRSP rats were given either a c ontrol diet with 0, 1 or 5% DHA beginning at 1 months of age for 3 months. SHRSP rats fed the DHA had increased latency (better memory) in passive avoidance testing compar ed to SHRSP rats fed 0% DHA (Minami et al., 1997). In another study, Carri e et al. (2000) showed that young rats (2-3 months) fed a diet enriched with sardine oil for 2 genera tions showed an increase in active avoidance performance compared to young control rats onl y on day 1 of the 5-day task, indicating a
106 lack of an overall effect of the sardine oil enrichment. There were also no differences in active avoidance performance between sardin e oil supplemented and control diet groups composed of adult and aged ra ts (Carrie et al., 2000). Cons istent with Carrie et al. (2000), de Wilde et al. (2002) found that ad ult (5 month) 2VO ra ts given dietary DHA and EPA since 1 months of age showed no di fference in active avoidance compared to non-supplemented and vehicle control rats. In another study, Hash imoto et al. (2002) administered a shuttle avoidance task, similar to active avoidance, to adult (9-month) rats that had been fed a DHA rich or control diet for 3 months followed by either sham surgery or infusion of the A 1-40 peptide into the left ventri cle. The avoidance task was given 1, 2 and 3 weeks after the surgery. The A -infused mice were impaired in this task compared to the vehicle infused mice. Notably, DHA administration significantly improved avoidance learning for both the sham and A -infused groups compared to both groups with no DHA supplementation (Hashimoto et al., 2002).
107 Table 1. Summary of behavioral studies invo lving n-3 deficiency or supplementation to rodents. Study Animal Model Dietary Supplementation Duration of Supplementation Task Results BarceloCoblijn et al., 2003 2M and 24M rats Fish oil or control 1 M Morris water maze No effect on acquisition or retention Calon et al., 2004 Tg2576 APP and NT mice n-3 deficient supplemented with DHA 17M to 22M Visible platform No effect Morris water maze Tg+ DHA improved acquisition; no effect on retention Carrie et al., 2000 Mice Sardine oil or Palm oil 2 generations Young (2-3M) Mature (9-10M) Old (17-19M) Open Field Young sardine increased activity Active Avoidan ce No effect Morris water maze Mature sardine increased time in former platform quadrant above chance Chalon et al., 1998 Wistar rats Fish oil or control 2 generations 2M Open Field Fish oil decreased ambulation Elevated plus maze No effect
108 Table 1 (Continued). Study Animal Model Dietary Supplementation Duration of Supplementation Task Results De Wilde et al., 2002 Bilateral carotid occlusio n in 4M rats n-3 PUFAs or control 1M to 7M Elevated plus maze at 5M No effect Active avoidan ce at 5M No effect Morris water maze at 7M No effect on acquisition or retention Gamoh et al., 2001 Wistar rats Fish oil deficient for 3 generations DHA supplement from 24M to 26M 8-arm radial maze at 25M No effect Hashim oto et al., 2005 A 1-40 infused rats Fish oil-deficient for 3 generations DHA supplement from 8-12M 8-arm radial maze at 10M DHA had increased reference & working memory Hashim oto et al., 2002 A 1-40 infused rats Fish oil-deficient 3 generations DHA supplement from 6M to 9M Shuttle avoidan ce at 9M DHA had increased avoidance learning Ikemoto et al., 2001 Donryu rats Safflower oil or Perilla oil (high n-3) 2 generations 1M to 3M supplemented safflower diet with DHA Brightne ss discrimi nation No effect Jensen et al., 1996 Wistar rats Seal oil; fish oil; vegetable oil; control for 4 generations 3M Morris water maze No effect on acquisition or retention Lim & Suzuki, 1999 Young mice (1M); Aged mice (14M) DHA or palm oil 5 months supplementation Maze learning -3 trials DHA decreased errors & latency for selected trials
109 Table 1 (Continued). Study Animal Model Dietary Supplementation Duration of Supplementation Task Results Lim & Suzuki, 2000 Mice DHA or palm oil Beginning at 3M for , , 1 or 3M Maze learning -3 trials DHA decreased errors & latency for selected trials Minami et al., 1997 SHRSP rats 0, 1, 5%DHA 1 M to 5M Passive avoidan ce SHRSP+DH A had increased avoidance learning Sugimot o et al., 2002 Mice DHA or control 9M to 11M 8-arm radial maze DHA increased working memory Suzuki et al., 1998 mice Sardine oil or palm oil Beginning at M for 12M Maze learning -3 trials Sardine oil decreased errors/latenc y for initial trials Wainwri ght et al., 1999 Artificia lly reared rats 0-2.5% DHA/AA combinations Postnatal day 5 to 1M Morris water maze No effect on acquisition or retention Abbreviations: M, month; NT, non-transgen ic mouse; Tg+, APP transgenic mouse; DHA, docosahexaenoic acid; PUFAs, polyunsatur ated fatty acids; SHRSP, stroke-prone spontaneously hypertensive rats; AA, arachidonic acid.
110 Aside from avoidance tasks, spatial maze learning is a good indicator of both reference and working memory. Sugimoto et al. (2002) found a decrease in working memory errors in the eight-arm radial maze for adult (9 months) mice supplemented with a DHA-rich diet for 2 months compared to unsupplemented control mice; there was, however, no difference in reference memory errors. Gamoh et al. (2001) showed no overall improvement in performance in the sa me task with aged rats administered DHA from 24 months to 26 months of age after tw o generations on a fish oil-deficient diet. Eight-arm radial maze learning, assessed by 3 trials per day for 3 separate days, beginning around 25 months of age for 5 week s showed no difference in working or reference memory with DHA supplementation compared to rats without this supplement (Gamoh et al., 2001). More recently, Hashimoto et al. (2005) fed a fish oil-deficient diet to rats for 3 generations. At 5 months of age, the F3 generation rats were either infused through osmotic minipumps with A 1-40 or vehicle alone for 5 weeks. Subsequently, rats from each group were administered either oral DHA or vehicle for 7 weeks and then tested in the eight-arm radial maze for 5 w eeks. Overall, DHA administration improved reference and working memory in both vehicle and A -infused groups (Hashimoto et al., 2005). Utilizing a different task involving m aze learning, Suzuki et al. (1998) supplemented 3-week old mice for 12 months with sardine oil. Maze learning was assessed by 3 trials administered 4 days apar t using a simple maze construction with one entry/exit and many blind alleys. Suzuki et al. (1998) found decrease d latency and errors for trials 1 and 2; by trial 3 the palm oil fe d mice exhibited a similar performance. A similar study by Lim and Suzuki (1999) fed DH A-enriched or palm oil-enriched diets to
111 3-month old and 14-month old mice for 5 months, followed by 3 trials of maze learning with the same protocol used by Suzuki et al. (1998). The young mice supplemented with DHA showed decreased latency and errors only on trial 2, whereas the aged mice supplemented with DHA showed decreased late ncy and errors on trials 2 and 3 (Lim & Suzuki, 1999). Similar to the above maze st udies, young mice (3 months old) were fed DHA enriched diet or palm oil enriched diet for 1 week, 2 weeks, 1 month, or 3 months then tested for the same maze learning as previously discussed (Lim & Suzuki, 2000). The mice fed the DHA enriched diet for 1 and 3 months had decreased errors and latency only in trial 3 compared to palm oil enri ched mice (Lim & Suzuki, 2000). The above results, performed with DHA administration for different durations, suggest that n-3 enrichment can improve cognitive performance at any age, although the data are far from conclusive. A major task used to evaluate spatial learning/ memory among rodents is the Morris water maze. This task provides measures for reference learning during acquisition and memory retention during the probe trial. Jensen et al. (1996) showed that rats fed fish, seal, or vege table oil-rich diets for 4 ge nerations had a trend toward improved performance in the Morris water m aze versus rats fed a control diet; however there were no differences between the groups in cognitive performance. The fish oil group had an insignificant decrease in acquis ition latency at 3 months of age during the 9-day task with no overall difference within any of the groups (Jensen et al., 1996). In conjunction with the lack of effect from fish oil supplementation by Jensen et al. (1996), Barcelo-Coblijn et al. (2003) found that 3-month old rats that had been fed a diet enriched with fish oil for 1 month showed no ove rall differences in Morris water maze
112 performance (acquisition or retention) be tween the enriched ve rsus control groups. Within the same study, there was also no difference in Morris maze acquisition or retention performance in 24 month old rats fe d a fish oil supplement for 1 month versus control fed rats (Barcelo-Coblijn et al., 2003). In another study, artificially reared rats were fed milk supplements with 0 to 2.5% DHA and/or AA in combination from postnatal day 5 to postnatal day 18 (Wainwright et al., 1999). After postnatal day 18, rats were weaned onto diet s that contained identical amounts of DHA and AA compared to the milk s ubstitutes. At 1 months of age, rats supplemented with DHA and/or AA showed simi lar latencies to controls during Morris water maze acquisition and rete ntion testing; thus, no e ffect of DHA and/or AA was observed (Wainwright et al., 1999). In addition to open field and active avoidance, Carrie et al. (2000) used the Morris wate r maze to evaluate cognition in young (2-3 months), mature (9-10 months) or old ( 17-19 months) mice administered fish oil supplementation (sardine oil) or a palm oil en riched diet over tw o generations. This study found no difference in acquisition be tween any of the groups; however, only mature mice fed the sardine oil diet show ed significant memory retention during the probe trial, as indicated by their spending mo re time than chance in the former platformcontaining quadrant (Car rie et al., 2000). Clearly, Morris water maze acquisition and retention cognitive performance does not seem be strongly affected by generations of fatty acid-supplemented diets in nontransgenic mice. Likewise, 17-month old Tg2576 APP transgenic and non-transgenic mice fed an n-3 deficient diet with or without DHA enrichment for ~3 months showed no differences in latency in visible platform testing (Calon et al., 2004). Within the first
113 3 blocks of visible platform testing, th e non-transgenic mice supplemented with DHA showed improved performance, as exhibited by a decreased latency compared to the nontransgenic mice on a low DHA diet. During Morris water maze testing, in the same study, APP mice fed the n-3 deficient diet show ed impaired spatial learning only over the last two blocks of testing, days 7-9 and 10-12, compared to APP mice supplemented with DHA and non-transgenic mice with or without DHA enrichment. Although there was no overall effect of DHA supplementation on spat ial learning, the aut hors conclude that DHA supplementation to the n-3 deficient APP mice corrected the learning impairment caused by n-3 deficiency. However, the probe trial measuring memory retention showed a lack of retention by all groups except the non-transgenic group supplemented with DHA that exhibited a partial quadrant preferen ce. This group spent significantly more time in the quadrant that formerly contained the platform versus the opposite quadrant. No other group showed this preference, indicating that DHA supplementation could not correct the spatial memory deficits by the APP mice fed the n-3 defi cient diet (Calon et al., 2004). Overall, n-3 supplementation appe ared to impact avoidance learning more than spatial memory tasks such as the Morris water maze, possibly due to a difference in cognitive domains used between the tasks. In conclusion, omega-3 fatty acids have been shown to have some beneficial affects on membrane lipid composition, inflam mation, the cholinergic system and LTP. Also described were improvements of cognition in some human trials and animal studies. However, there are some drawbacks to the pr ior behavioral studies involving n-3 and n-6 fatty acids in human and animal studies. To date, there have been no human studies done that have used DHA and EPA in combination to analyze effects on AD as treatment or
114 prevention. Yehuda et al. (1996) did test the behavioral eff ects of LA and ALA together in AD patients over a short one-month treatme nt period; however, only about 15% of ALA is converted to DHA for use in the body. No other study has followed a treatmentbased approach using fatty acids or fish oil to reverse cognitive decline due to Alzheimers disease. However, essentially all of the protection-based human studies have been epidemiologic, retrospective stud ies which are often inaccurate and unable to control for many other factors that could affect cognition later in life. Additionally, there are no longitudinal studies that have administered or retrospe ctively looked at n-3 fatty acids or fish oil supplements to determine any beneficial behavioral affect. Aside from human studies, there are no animal studies involving an AD transgenic mouse model which utilized DHA and EPA as prevention therapy. Calon et al. (2004) used APP transgenic mice on an n-3 deficient diet s upplemented with only DHA as treatment for AD. This approach was also used many tim es in non-transgenic rodent studies, as previously discussed, and is not a realistic approach to human dietary intervention. A more realistic approach would be a supplemen tation with an appropria te ratio of n-6 and n-3 fatty acids (composed primarily of DH A and EPA) that could be duplicated for human dietary prevention of AD. The majority of control diet s that were used in animal studies were n-3 deficient, having only AL A as the primary n-3 fatty acid, and had no DHA or EPA. Also, the experimental diets us ed for comparison to the control diets do not incorporate both DHA and EPA specifically; instead they used an ethyl ester form of DHA or a complete fish oil supplement which ad ds additional n-3 fatty acids. Neither of these diet compositions accurately displays a sufficient human dietary intervention that would incorporate amounts of n3 fatty acids present in foods. In addition, most of the
115 animal studies used only one behavioral ta sk, either active avoidance or Morris water maze, to evaluate the behavioral affects of n3 administration. Such behavioral analyses do not provide a complete assessment of cognitive changes that could occur from a dietary intervention study. In stead, a more complete behavi oral analysis should include tasks that test multiple sensorim otor and cognitive domains. XI. Statement of Purpose As detailed in the Background section, two potential therapeutics against AD are A -based immunotherapy and n-3 fatty acid admi nistration. This proposal will involve studies that directly test th e ability of these two therap ies to protect against AD-like behavioral impairment and pathology in a transgenic mouse model for AD the APP+PS1 mouse. Regarding A immunotherapy, to date there have been no studies that show lifelong or longitudinal beha vioral benefits of A vaccinations. Also, no studies have evaluated the same vaccinated transgenic animals in an extensive behavioral battery that incorporates multiple sensorimotor and cogni tive domains at several time points during the vaccination period. Therefore, this pr oposal involves long-ter m prevention-based A immunotherapy to APP+PS1 mice that will: longitudinally evaluate cognitive func tion in at two separate time points during aging in an extensive behavioral battery. determine extent of A pathology at the latter time point and correlate with behavior.
116 Regarding n-3 fatty acid administration, there have been no clinical studies performed to determine the protective or treatment effects of DHA+EPA supplementation against AD. Also, there have been no anim al studies that have used a comprehensive behavioral batter to assess th e effects of n-3 fatty acids, specifically DHA and EPA, on behavior in non-transgenic or AD transgenic animal models Therefore, this proposal will administer to APP+PS1 mice a preventionbased high n-3 fatty acid supplement, rich in DHA and EPA. This diet is modeled to mi mic a twice-weekly dietary intake of fish for humans. In addition, a high n-6 fatty acid diet, rich in LA, will be given to other APP+PS1 mice. This n-6 diet is modeled to mimic a typical (bad) Western diet. To compare the dietary effects of a high n-3 fatty ac id diet versus a high n-6 fatty acid diet in APP+PS1 mice and non-transgenic mice, this proposal will: evaluate cognitive function in an extensive behavioral battery at 4 to 5 months into dietary treatment. examine fatty acid levels in multiple tissues including brain, liver and plasma to note differential incorporat ion rates into membranes, which may potentially affect in tracellular signaling. analyze markers that could impact AD progression and be affected by n-3/ n-6 dietary manipulation, such as A pathology, oxidative damage and inflammatory proteins, whic h correlation to behavior.
117 Materials and Methods I. Life-long Vaccination Study Animals. Double transgenic APP + PS1 mice we re derived from a cross of the mutant APPK670N,M671L transgenic line with the mutant PS1 transgenic line 5.1. APP mice had a C57B6/SJL X C57B6 background and PS1 -5.1 mice had a Swiss Webster/B6D2F1 X B6D2F1 background, providing progeny with a mixed background. Non-transgenic littermates were used as cont rols, with all mice being from the eighth or ninth generation following the initial cross. After weaning, mice were genotyped and group housed until the start of vaccinations at two months of age. All mice were maintained on a 12 hour light-dark cycle, with free acce ss to rodent chow and wate r. Behavioral testing was performed in the light period and in the same room where animals are being housed. General Protocol. The general protocol for this st udy is temporally depicted in Fig. 3. Beginning at 2 months of age, APP+ PS1 mice were injected monthly with the A 1-42 peptide (n=8) or phosphate-buffered saline (PBS) with adjuvant (n=6). Control non-transgenic littermates (n=9) received mont hly injections of PBS with adjuvant. All mice were tested in a 6-week behavioral ba ttery at 4-6 months of age (2-4 months into immunotherapy), then again in the same test battery at 15-16 months of age (1314 months into immunotherapy). The test battery consisted of three sensorimotorbased, one anxiety-based, and five cognitive-based tasks, that were performed in the
118 following order: open field activity, balance b eam, string agility, Y-maze, elevated plus maze, Morris water maze, circular platform, platform recognition, and radial arm water maze.
119 Fig. 3. General protocol time line for life-long vaccination study. 2 M Begin Monthly Injections 4-6 M 15-16 M 6-Week Behavioral Testing 6-Week Behavioral Testing
120 Following completion of the test battery at 16 months of age, all animals were euthanitized, their brain re moved and processed for A histopathology. Using DNA extracted from tail samples, PCR genotyping of animals for mutant APP and PS1 genes was performed at weaning, and again (for c onfirmation) on the day of euthanitization. Procedures for all of the above were revi ewed and approved by the USF Institutional Animal Care and Use Committee (IACUC). Vaccination protocol. The human A 42 peptide (Bachem) was first suspended in pyrogen-free Type I water at 2.2 mg ml-1 and then mixed with 10X PBS in order to yield 1X PBS. This solution was then incubated ove rnight at 37C, then mixed with Freunds complete adjuvant at 1:1; 100 g of the A vaccination mix was injected subcutaneously into APP+PS1 mice. Each boost va ccination at monthly intervals was prepared fresh with incomplete Freunds. Injections for control APP+PS1 and non-transgenic mice consisted of PBS plus adjuvant prepared in the same way. The experimenter that performed these injections had no role in the beha vioral testing of mice. Histopathology and Image Analysis. Following completion of behavioral testing, mice were overdosed with pentobarb ital (100 mg/kg i.p. ) and perfused transcardially with 25ml of normal saline (0.9 %). Brains were removed and immersion fixed in a fresh solution of 4% paraformal dehyde (pH 7.4) for 24 hours; they were then cryoprotected in a series of sucrose solutions, frozen and sectioned in the horizontal plane at 25 m using a microtome. Brain sections were stored at 4C in Dulbeccos PBS for
121 A immunocytochemistry (total A = both A 1-40 & A 1-42) and Congo red (A 1-40) histologic staining for compact depositi on as described in a latter section. II. Omega-3 Fatty Acid StudySurvival Analysis Animals. All APP/PS1 and PS1 transgenic mice were obtained from a cross between heterozygous male mice carrying the mutant APPK670N,M671L gene and homozygous female mice with the mutant PS1 transgenic line 6.2. All of the offspring contained a mixed background of 37.5% C57, 25% B6, 12.5% SJL and 25% Swiss Webster. The non-transgenic and APP single transgenic mice were obtained from a cross of the F1 PS1 female mice with P (parental generation) heterozygous APP male mice to obtain APP+PS1, APP, PS1 and non-transgen ic (Tg-) offspring with a 56.25% C57, 12.5% B6, 18.75% SJL and 12.5% Swiss Webster mixed background. The mice were genotyped and singly housed after weaning. General Protocol. Thirty-eight F1 APP+PS1 and twenty-seven F2 Tgmice were started on either a low omega-3, omega-3deficient diet (omega-6 only), or standard rodent chow at an average of 6 weeks of age (range = 4-8 weeks) (Fig. 4). The low omega-3 diet contained 13 A PP/PS1 and 9 Tgmice; there were 15 APP/PS1 mice and 9 Tgmice in the omega-6 only diet group and th e standard diet incl uded 10 APP/PS1 and 9 Tgmice. There was a 100% mortality rate for both Tg+ and Tgmice on the omega-3 deficient diet, with most Tg+ mice dying betw een 20-25 days after diet initiation and all Tgmice dying within 2-8 days of diet initiat ion. Because of the longer survival of the Tg+ mice on the omega-3 deficient diet, Study 1 has been designed to investigate the
122 potential role of APPsw and/or PS1 transgenes in protecting against early mortality that otherwise occurred in Tgmice. Thus, 7 APPsw and 7 PS1 transgenic mice will be started on the omega-3 deficient diet at 2 months of age, with daily monitoring for mortality. To determine the effects of age on the mortality of APPsw+PS1 mice resulting from the omega-3 deficient diet, 2-3 APPs w+PS1 mice will be started on the omega-3 deficient diet at 2, 3 and 4 months of age. Daily monitoring for mortality will occur. Figure 2 indicates a timeline of when all Tg+ and Tgmice were or will be started on the omega-3 deficient diet. Table 1 gives all of the major fatty acids present within the three diets. None of these animals will be behavi orally tested; they will only be monitored for health problems and rate of mortality once they are started on their respective diets.
123 Fig. 4. General protocol time line for omega-3 fatty acid study-Survival Analysis. Weeks 4 6 8 10 12 14 16 Began 6w APP/PS1 & NT on diet Began 4M APP/PS1 on diet Began 3M APP/PS1 on diet Began 2M APP/PS1 on diet Began PS1 & APP on diet
124 Fatty Acid Diets. Both experimental diets were obt ained from Purina Mills Test Diets (Lafayette, IN) and lipid profiles were performed by Barrow-Agee Laboratories to confirm the fatty acid content for each diet The 2018 Teklad Global 18% Protein Rodent Diet standard diet used wa s obtained from Harlan. The low omega-3 fatty acid diet included 8.55% fat with the majority of fat fr om oleic acid and palmitic acid (Table 2). The n-6/n-3 ratio for this diet was 6.3 to 1. The omega-3-defici ent diet included 6.72% fat with the majority of fat also from oleic acid and palmitic acid (Table 2). This diet contained an n-6/n-3 ratio which could not be determined because it contained no detectable amounts of any n-3 fatty acids. There were no de tectable amounts of DHA or EPA in any of the three diets presented in Table 1. There were similar amounts of n-6 fatty acids present in both diets. The major fa tty acids present in each diet are presented in Table 2.
125 Table 2. Percent fatty acid composition of total fat for three diets. Standard Diet Low Omega-3 Diet Omega-3Deficient Diet % Total Fat 5.00 8.55 6.72 Total Saturated Fats 9.59 42.07 30.23 Total Monounsaturated Fats 12.84 32.30 28.46 Total Polyunsaturated Fats 34.11 8.42 4.64 Saturated Fats Butyric C4:0 0.00 0.00 0.00 Caproic C6:0 0.00 0.00 0.00 Caprylic C8:0 0.00 0.34 1.16 Capric C10:0 0.00 0.00 0.00 Lauric C12:0 0.26 0.19 0.00 Myristic C14:0 0.06 8.19 1.99 Pentadecanoic C15:0 0.00 0.69 0.22 Palmitic C16:0 7.64 25.40 19.03 Magaric C17:0 0.00 0.67 0.00 Stearic C18:0 1.50 5.80 6.61 Arachidic C20:0 0.01 0.48 0.73 Behenic C22:0 0.03 0.316 0.50 Lignoceric C24:0 0.00 0.00 0.00 Monounsaturated Fats Palmitoleic C16:1n-7 0.07 6.99 1.39 Heptadecenoic C17:1n-8 0.00 0.39 0.00 Oleic C18:1n-9c 12.59 19.61 24.77 Gadoleic C20:1n-9 0.17 0.00 0.00 Erucic C22:1n-9 0.01 0.00 0.00 Polyunsaturated Fats Linoleic C18:2n-6c 31.35 7.27 4.64 -Linolenic C18:3n-3c 2.76 0.00 0.00 -Linoleic C18:3n-6c 0.00 0.00 0.00 Octadecatetraenoic C18:4n-3 0.00 0.00 0.00 Eicosadienoic C20:2n-6 0.00 0.00 0.00 HomoLinolenic C20:3n-6c 0.00 0.00 0.00 Arachidonic C20:4n-6c 0.00 0.00 0.00 Eicosapentaenoic C20:5n-3c 0.00 0.00 0.00 Docosapentaenoic C22:5n-3 0.00 1.15 0.00 Docosahexaenoic C22:6n-3c 0.00 0.00 0.00 Total n-3 Fats 2.76 1.15 0.00 Total n-6 Fats 31.35 7.27 4.64 n-6/n-3 ratio 11.4 to 1 6.3 to 1
126 II. Omega-3 Fatty Acid StudyBehavioral Pathologic and Neurochemical Assessments Animals. All mice were obtained from a second generation cross between heterozygous APPK670N,M671L and heterozygous PS1transgenic line 6.2. The backgrounds of all offspring were a mix of C57/B6/SJL/S wiss Webster. The mice were genotyped and singly housed after weaning and a confirmato ry genotyping was performed 1 month prior to behavioral testing. All mice were main tained on a 14 hour light and 10 hour dark cycle for the duration of the study; all behavioral testing was performed during the light cycle. General Protocol. A total of 16 double transgen ic APP+PS1 mice, 7 APP mice and 24 non-transgenic (NT) littermates were randomly divided into one of three diet groups. Beginning at 2 months of age, NT a nd transgenic (Tg+) mice were divided into three separate diet groups: high omega-3, hi gh omega-6, and standard diets (Fig. 5). Because of the oxidative properties of omega-3 fatty acids the diets were stored at -20C in 4-7 day supply aliquots. Mice fed the hi gh omega-3 diet consisted of 7 APP+PS1 mice and 2 APP mice and 8 NT mice; the standard diet fed mice included 9 APP+PS1 and 1 APP mouse and 9 NT mice; the high omega-6 fatty acid diet group consisted of 1 APP+PS1 mouse and 4 APP mice and 7 NT. At 6 months of age, the NT groups were behaviorally tested in a 6-week battery th at included three sensorimotor-based, one anxiety-based, and five cognitive-based tasks, performed in the following order: open field activity, balance beam, string agility, Y-maze, elevated plus maze, Morris water maze, circular platform, platform recognition, and radial arm water maze. At 7.5 months of age, Tg+ mice were examined in the same 6-week battery of tasks (Fig. 5). NT and
127 Tg+ mice could not be tested concurrently be cause of the large total number of animals involved (e.g., 47 mice). Immediately follo wing behavioral testing all mice were euthanitized; brains were then removed for A histological and fatty acid analyses.
128 2 M Begin Tgand Tg+ on Diets 6 M 7.5 M Behavioral Testing of Tg+ Sacrifice TgSacrifice Tg+ 9 M Behavioral Testing of TgFig. 5 General protocol time line for omega-3 fatty acid study II-Behavioral, Pathologic and Neurochemical Assessments
129 Fatty Acid Diets. Fatty acid diets were obtained fr om Purina Mills Test Diets (Lafayette, IN). Lipid prof iles were performed by Barrow-Agee Laboratories for each diet prior to administration to the mice to c onfirm their fatty acid content. Each of the experimental diets contained 10% fat. The high omega-3 diet included 6% safflower oil and 4% menhaden fish oil with an n-6:n3 ratio of 3.8 to 1; the high omega-6 diet included 9.5% safflower oil and only 0.5% menhade n fish oil resulting in an n-6:n-3 ratio of 32.8 to 1. The standard diet was the 2018 Teklad Global 18% Pr otein Rodent Diet received from Harlan, which was normally gi ven to all mice within our mating colonies who were not on a special diet. As indicated in Table 3, this Harlan diet had only half the fat as the other two diets. The exact lipid composition for the major fatty acids present in each diet is indicated in Table 3.
130 Table 3. Percent fatty acid composition of total fat for three diets. Standard Diet High Omega-3 Diet High Omega-6 Diet % Total Fat 5.00 10.44 9.99 Total Saturated Fats 9.59 20.04 10.79 Total Monounsaturated Fats 12.84 19.73 14.59 Total Polyunsaturated Fats 34.11 64.22 74.27 Saturated Fats Butyric C4:0 0.00 0.00 0.00 Caproic C6:0 0.00 0.00 0.00 Caprylic C8:0 0.00 0.00 0.00 Capric C10:0 0.00 0.00 0.00 Lauric C12:0 0.26 0.06 0.00 Myristic C14:0 0.06 3.37 0.58 Pentadecanoic C15:0 0.00 0.28 0.06 Palmitic C16:0 7.64 11.88 7.54 Magaric C17:0 0.00 0.29 0.07 Stearic C18:0 1.50 2.84 2.16 Arachidic C20:0 0.010 0.28 0.33 Behenic C22:0 0.03 0.00 0.00 Lignoceric C24:0 0.00 0.09 0.11 Monounsaturated Fats Palmitoleic C16:1n-7 0.07 4.03 0.60 Heptadecenoic C17:1n-8 0.00 0.29 0.08 Oleic C18:1n-9c 12.59 11.94 13.03 Gadoleic C20:1n-9 0.17 0.094 0.24 Erucic C22:1n-9 0.01 0.07 0.00 Polyunsaturated Fats Linoleic C18:2n-6c 31.35 49.99 71.99 -Linolenic C18:3n-3c 2.76 0.70 0.13 -Linoleic C18:3n-6c 0.00 0.26 0.00 Octadecatetraenoic C18:4n-3 0.00 1.39 0.22 Eicosadienoic C20:2n-6 0.00 0.22 0.04 HomoLinolenic C20:3n-6c 0.00 0.07 0.00 Arachidonic C20:4n-6c 0.00 0.28 0.04 Eicosapentaenoic C20:5n-3c 0.00 4.74 0.85 Docosapentaenoic C22:5n-3 0.00 0.87 0.11 Docosahexaenoic C22:6n-3c 0.00 5.71 0.89 Total n-3 Fats 2.76 13.40 2.20 Total n-6 Fats 31.35 50.82 72.07 n-6/n-3 ratio 11.4 to 1 3.8 to 1 32.8 to 1
131 Brain Collection and Dissection. Immediately following be havioral testing mice were euthanitized with an overdose of pent obarbital (100mg/kg), bl ood and liver sections were collected and the brain were removed a nd bisected sagittally. Blood (0.2 ml) was drawn directly from the hear t prior to saline perfusion (100 ml). The blood mixed with 0.5M EDTA was immediately centrifuged to sepa rate plasma from red blood cells. Both portions were stored and fro zen at -80C for lipid analys is. After saline perfusion a similar portion from each animals liver was re moved and stored in a 1.5 ml tube at -80C for lipid analyses. For the brain, the le ft half was stored overnight in a 4% paraformaldehyde solution for immunohistochemist ry. The left half was then transferred to a graded series of sucrose solutions at 4C ending at 30% sucrose for storage until sectioning with a slid ing microtome into 25 m coronal sections for A immunocytochemistry and histol ogy as described in a latter section. The right half of each brain was momentarily placed in a cold sa line solution, and then dissected into brain stem, cerebellum, posterior cortex, anterior cortex, striatum and hippocampus; each brain region was transferred to a separate pre-labe led 1.5 ml tube and stored at -80C for neurochemical analysis of protein carbonyls, lipid per oxidation and cytokines as described below. Extraction of Brain Protein for Sa ndwich Enzyme-Linked Immunosorbent Assay (ELISA). A 5% sucrose homogenate (wet weight of tissue/ volume) from frozen mouse hippocampus was prepared and extracted as described by Schmidt et al. (2004). Each hippocampal tissue was individually weighed, and then combined with tissue homogenization buffer (THB) at 2 ml per 100 mg of tissue. In addition to the THB, a
132 protease inhibitor cocktail was also added at 20% of the we t tissue weight to prevent against degradation of the protei ns within the tissue samples. On ice, each of the samples were combined with the calculated amounts of THB + inhibitors and fully homogenized. After homogenization, the samples were frozen and stored at -80 C until diethylamine (DEA) and formic acid (FA) extractions. DEA extraction was used to separate soluble A from the tissue homogenate. First, 100 l of 5% homogenate was mixed with a 0.4% DEA solution (diluted in 100mM NaCl) over ice. This mixture was then tran sferred into a thick-wa lled polycarbonate tube and centrifuged at 100,000 x g for 1 hour at 4C. Followi ng centrifugation, 170 l of supernatant was removed and added to a microcentrifuge tube containing 17 l of 0.5M Tris Base, pH 6.8 (1 l per 10 l of supernatant) and vortexed briefly. These samples were then frozen on dry ice and stored at -8 0C for later analysis by the ELISA kits. FA extraction was used to isolate insoluble A from tissue homogenate. For this extraction 100 l of THB was added to the pellet from the initial polycarbonate tube to return the mixture to the original volume. Following the addition of 220 l of 95% FA, pellets were sonicated for 1 minute on ice. This mixture was then centrifuged at 100,000 x g for 1 hour at 4C. After centrifugation, 52.5 l of the intermediate phase of the FA extracted mixture was added to 1 ml of FA neutralization solution. Samples were vortexed, immediately frozen on dr y ice, and then stored at -8 0C for later analysis by the ELISA kits.
133 A Brain Deposition. Sandwich ELISA kits for A 1-40 and A 1-42 analyses were purchased from Signet Laboratories. Briefly, the standard curve for A 1-40 was generated with the maximum value at 2000 ng/ml and the minimum value at 0 ng/ml. Each well of the 96-well plate was loaded in duplicate with either standa rd curve samples or diluted tissue samples. For the formic acid (FA) ex tracted tissue, the homogenate was diluted with wash/sample diluent (provided) to a ra tio of 1:500 of homogenate to diluent; for the diethylamine (DEA) extracted tissue, the homogenate was diluted with wash/sample diluent to a ratio of 1:400. Once loaded into the plate, the standard curve and tissue samples incubated overnight at 4C. Follo wing this incubation, the plate was washed, diluted primary antibody was added, and the plate was incubated for 2 hours at room temperature. The plate was washed again, and then incubated for 2 hours with secondary antibody-HRP comple x. Following another wash, o -Phenylenediamine dihydrochloride (OPD) substrate was added, and incubated in a dark room for 45 minutes. Stop solution (3N H2SO4) was added to each well and the optical density was read at 490 nm. All of the above st eps were repeated to measure for A 1-42, except that the tissue samples were diluted to 1:100 for the DEA extracted samples, while the same 1:500 dilution was used for the FA extracted sa mples. The standard curve was generated so that the lowest point on the curve was used to correct all of the tissue samples to account for the background interference. Th e optical densities were then used to calculate wet brain protein concen trations (pmol/g) for either A 1-40 or A 1-42. Fatty Acid Extraction, Transesterifica tion and Fast Gas Chromatography. The extraction was performed on frontal cortex according to the method of Folch et al.
134 (1957). Briefly, individual samples were th awed and homogenized in 20 volumes of chloroform-methanol (2:1, v/v). The homoge nate was partitioned with an aqueous salt solution and a small aliquot of the lower phase (total lipid extract) was used for the fatty acid analysis. Transesterification was performed accord ing to Lepage and Roy (1986). Briefly, an internal standard solution (providing 10 g of 23:0 methyl ester) was combined with an aliquot of the lipid extract and a mixture of methanol-hexane (4:1; v/v) into a borosilicate glass tube. Samples were vorte xed and placed on ice; subse quently acetyl chloride was added. The tubes were then capped, placed under nitrogen, and transf erred to a heating block at 100C for 10 minutes, the tubes ag ain vortexed and the caps retightened. Following an additional 50 minutes on the he ating block, samples were placed on ice, uncapped and neutralized by addition of K2CO3. The samples were then centrifuged to remove emulsion and to separate the mixt ure into two phases. The upper phase was collected and evaporated under a stream of dry nitrogen to a volume of 60 l. This solution was transferred to a gas chromat ography (GC) vial for fast GC analysis. A fused silica capillary column (DB-FFAP) of 0.1 mm ID x 15 m length with a film thickness of 0.1 m was used for the fast GC analysis. The following temperature program was used: initial, 150C with a 0.25 minute hold; ramp: 35C/minute to 200C, 8C/minute to 225C with a 3.2 minute hold, a nd then 80C/minute to 245C with a 2.75 minute hold. The following instrumental c onditions were used: carrier gas was H2 at a flow rate of 56 cm/s and a constant head pr essure of 344.7 kPa; FID set at 250C; air and nitrogen make-up gas flow rate s at 450 ml/min and 10 ml/min, respectively; split ratio of 50:1; sampling frequency of 50 Hz; autosampler injections of 2 l volume. Individual
135 fatty acids were identified by comparison with retention times of internal standards and calculated as mean percentage of total frontal cortex fatty acids. Cytokines. Transgenic AD mice produce pro-infl ammatory cytokines that might play a role in the AD-like pathology exhibited by these mice. Therefore, to elucidate the role of fatty acid dietary in corporation on inflammation, proinflammatory cytokines (IL1 IL-1 IL-2, IL-6, IL-12p70, TNF, IFN and GM-CSF) and anti-inflammatory cytokines (IL-4, IL-10) were determined us ing the RayBiotech Custom Mouse Cytokine Antibody Array within plasma. Briefly, memb ranes containing the antibodies to the previously mentioned 10 cytokines were blocked against unspecific binding and subsequently incubated for one hour with 1:10 diluted plasma samples from each APP/PS1 and NT mouse. One NT mouse was deleted due to limited space requirements for the kit. The same NT mouse was also deleted from RAWM analysis due to nonperformance. The membranes were then incubated with 1x secondary biotinylated antibodies for one hour, followed by incuba tion with HRP-conjugated streptavidin overnight at 4C. Between each of the in cubations, the membranes were washed with provided washing buffer solutions. Detecti on of cytokine expression was determined with a two-minute incubation of provided dete ction buffers and using Fujifilm AR x-ray film that was allowed to develop for 7 seconds. A Kodak DC290 digital camera was used to take back-lit photographs of the AR film. The mean intensity of each signal was determined by densitometry and quantified by th e Kodak ID Imaging Analysis Software. Each of the signals were standardized to a zero to one scale based on minimum and maximum mean intensity readings for each cytokine. This was necessary due to the
136 naturally occurring variability (100x-1000x) in expression levels among the various cytokines. Relative cytokine expression levels among the four groups were compared using the standardized m ean signal intensities. IV. Behavioral Testing Procedures Open Field. Open field measured explorator y behavior and general activity. Mice were individually placed into an open black box 81 x 81 cm with 28.5 cm high walls. This area was divided by white lines into 16 squares measuring 20 x 20 cm. Lines crossed were counted for each mouse over a five-minute period. Balance Beam. This task measured balance an d general motor function. The mice were placed on a 1.1 cm wide beam 50.8 cm long suspended 46 cm above a padded surface by two identical columns. Attached at each end of the beam was a 14 x 10.2 cm escape platform. Mice were placed on the beam in a perpendicular orientation and were monitored for a maximum of 60 seconds. The time spent by each mouse on the beam before falling or reaching one of the platforms was recorded for each of 3 successive trials. If a mouse reached one of the escap e platforms, a time of 60 seconds was assigned for that trial. The average of all three trials was calculated and recorded. String Agility. To assess forepaw grip capacity and agility, mice were placed in the center of a taut cotton st ring suspended 33 cm above a padded surface between the same two columns as in the balance beam tas k. Mice were allowed to grip the string with only their forepaws and then released for a maximum of 60 seconds. A rating system, ranging between 0 and 5, was employed to asse ss string agility for a single 60-second trial. A string agility score was given to each mouse based on the following scale: 0
137 was awarded for mice that were unable to hang for even a short period of time, 1 if the mouse hung by both forepaws for 60 seconds without escape, 2 was given if the mouse hung with both forepaws plus one hind limb, 3 if the mouse hung with all four paws around the string with no escape, 4 if the mo use had all four paws plus tail around the string, and 5 was given if the mouse escaped to one of the vertical support columns. Y-maze. To measure general activity and basic memory function, mice were allowed 5 minutes to explore a black Y-maze wi th 3 arms. Each arm measured 21 x 4 cm with 40 cm high walls. Mice were placed in th e center of the maze facing the center area and allowed to explore for 5 minutes, with th e number and sequence of arm choices being recorded. General activity was measured as the total number of arm entries, while basic mnemonic function was measured as a percent spontaneous alternati on (the ratio of arm choices differing from the prev ious two choices divided by th e total number of entries). For example, the sequence of arm en tries (2,3,1,3,2,1,2,3) has six alternation opportunities (total entries minus two) a nd the percent altern ation would be 67%. Elevated Plus Maze. To measure anxiety/emotionality, mice were placed in the center of an elevated plus maze 82 cm above the floor. The maze consisted of two opposite open and two opposite closed arms, each 30 x 5 cm; 15 cm high black aluminum walls surrounded the closed arms. The mice were placed in the 5 x 5 cm maze center, facing a closed arm, and were allowe d to explore for 5 minut es. The total number of closed arm entries, open arm entries a nd total time (seconds) spent in the open arms was recorded. Morris Water Maze. To measure reference learning (acquisition) and memory retention, mice were placed in a 100 cm pool th at was divided into 4 equal quadrants by
138 black lines drawn on the floor of the pool. A transparent 9 cm platform was placed 1.5 cm below the surface of the wate r in the center of quadrant two. An assortment of visual cues surrounded the pool. For each of four successive one-minute trials per day, mice were started from a different quadrant; the sa me quadrant start pattern was used across 10 days of acquisition. Latency to find the plat form (maximum of 60 seconds) was recorded for each trial and the four daily trials were averaged for statistical analysis. Once a mouse found the platform, it was allowed to stay on it for 30 seconds; if a mouse did not find the platform, it was gently guided to the platform and given the 30 second stay. Animals that did not find the platform were given a latency of 60 seconds. On the day following acquisition testing (day 11), a memo ry retention (probe) trial was done. For this 60-second trial, the platform was remove d and the mouse was started in the quadrant opposite the platform-containing qua drant. Percent time spent in each quadrant, annulus crossings, and average swim speed was determ ined from videotape recordings of this probe trial. Circular Platform. As a test of spatial (refere nce) learning and memory, mice were placed in a 69 cm circular platform with 16 equally-spaced holes on the periphery of the platform. Underneath only one of the 16 holes was a box filled with bedding to allow the mouse escape from aversive stimu li. The aversive stimuli included two 150watt flood lamps hung 76 cm above the platform and one high-speed fan 15 cm above the platform. Each mouse was given one 5-minute trial per day to explor e the area. Between each animals daily trial, the box position was ch anged to a different one of three possible escape holes and the platform was cleaned with a dilute vinegar solution to control for olfactory cues. The box position was changed between mice so that different mice had
139 different escape holes, but a ny given mouse maintained th e same escape hole across all days of testing. Prior to day 1 of test ing, two consecutive shaping trials were performed wherein mice were placed in the ce nter of the platform and gently guided to their escape location. Fo r the single trial administered on each of 8 test days, mice were placed in the center of the platform faci ng away from their escape hole and given 5 minutes to explore. Escape latency was measured (maximum of 300 seconds), as was the total number of errors (e.g., the number of head pokes into non-escape holes). Platform Recognition. This task measured the ability to search for and identify/recognize a variably-placed visible platform. Although this task requires good vision, it is a cognitive-based task because: 1) it necessitates animals to change from the Morris mazes spatial strategy to a recognition/identificati on strategy, and 2) it requires animals to ignore the spatial cues present ar ound the pool, which was the same pool used in earlier Morris water maze testing. The platform recognition task employed a 9-cm circular platform raised 0.8 cm above the su rface of the water, with a prominent 10 x 40 cm black ensign attached. Mice were started from the same location in the pool for four 60-second maximum trials per day for 4 days. For each trial, the platform location was changed to a different one of the four qua drants. Mice were allowed a maximum of 60 seconds to search/identify and ascend the platform, with a 30second stay if they located the platform. Mice that did not find the platform within 60 s econds were gently guided to the platform by the experimenter and allowe d to stay for 30 seconds. For statistical analysis, escape latencies for all f our daily trials were averaged. Radial Arm Water Maze. Working (short-term) memory was evaluated in the radial arm water maze (RAWM) task, using the same pool that was involved in both
140 Morris water maze and platform recognition testi ng. This task also used the same clear platform and visual cues as in Morris m aze testing. For RAWM testing, however, an aluminum insert was placed in the pool to create 6 radially distributed swim arms 30.5 cm in length and 19 cm wide emanating from a central circular swim area 40 cm in diameter. The insert extended 5 cm above th e surface of the water. The last of four consecutive acquisition trials (T4) and a 30-minute delayed retention trial (T5) were indices of working memory. On any given da y of testing, the submerged clear platform was placed at the end of one of the six swim arms. The platform location was changed daily to a different arm in a semi-random patte rn. For both studies (I and IIB), mice were tested for 9-12 days. On each day, different st art arms for each of the 5 daily trials were selected from the remaining 5 swim arms in a semi-random sequence that involved all 5 arms. For any given trial, the mouse was pl aced into that trials start arm facing the center swim area and given 60 seconds to find th e platform with a 30 second stay. Each time the mouse entered a non-platform containi ng arm it was gently pulled back into the start arm and an error was recorded. An erro r was also recorded if the mouse failed to enter any arm within 20 seconds (in which case it was returned to that trials start arm) or if a mouse entered the platform-containing ar m but did not find the platform. If the mouse did not find the platform within a 60-second trial, it was guided by the experimenter to the platform, allowed to st ay for 30 seconds, and was assigned a latency of 60 seconds. An error was assigned to any an imal that, for any one minute trial, did not find the goal arm and refused to make at leas t 3 choices on their own during that trial. This number was calculated by averaging errors for all animals that did not locate the
141 platform for Block 1 (day 1 through day 3) on trial 1. Both the number of errors (incorrect arm choices) and escape latenc y were recorded for each daily trial. V. Brain A Deposition Determinations A Immunostaining. The total A primary antiserum was raised against A 40 in rabbits and recognized A 40 and A 42 via ELISA assays. Antiserum selective for both A 1-40 and A 1-42 was purchased from Quality C ontrolled Biochemical (Hopkinton, MA). Confirmation of selectivity was pr ovided by preabsorption experiments, which blocked all punctuate staining. Brain sections were incubate d with the primary antibody overnight at 4C. The sections were then incubated in biotinylat ed secondary antibody for two hours followed by streptavidin-peroxidase The peroxidase reaction consisted of a solution of 1.4mM diaminobenxidine and 0.03% hydrogen peroxide in PBS for five minutes. Congo Red Staining. Congo red staining for compact A deposition was performed on slide-mounted sections that were dried for at least 12 hours and then rehydrated for approximately 30 seconds prio r to staining. The hydrated slides were incubated in a freshly prepared alkaline alc oholic saturated chloride solution consisting of 2.5mM NaOH in 80% reagent alcohol (95% ethanol and 5% isopropanol) for 20 minutes. The sections were then incubated in 0.2% Congo red in a freshly prepared and filtered alkaline alcoholic saturated sodium chloride for 30 minutes. The slides were rinsed with three rapid changes of 100% et hanol, cleared thro ugh three rapid changes of xylene, and finally cover slipped with Permount.
142 Quantification of A deposition. The Oncor V150 image analysis system was used to quantify stained brain se ctions. This software used hue, saturation and intensity (HIS) to segment objects within the image fi eld. Standard slides were used, which contain extremes of staining intensities, to establish thresholds for object segmentation. The operator for image analysis remained bli nd with respect to ge notype or behavioral assessment. For each animal, frontal cortex and hippocampus was qua ntified from four horizontal sections spaced 600 m apart beginning at 2000 m ventral to bregma. The frontal cortex measurement used an 80x field wi th one limit as the edge of the cortex and the other limit as the midline in the most an terior position possible. The total cortical measurement area consisted of a rectangular 850,000 m2. This area was primarily comprised of the middle two-thirds of the corti cal mantle (e.g., cortical layers 1 and 6 are excluded). The hippocampal an alyses involved three sub-ar eas: CA1, CA3, and dentate gyrus. The areas measured were comprised of 6-8 horizontal sections equally spaced within each region of CA1, CA3 and dentat e gyrus. Region CA1 was defined as the pyramidal cells on the opposite side of the hippocampal fissure from the dentate gyrus. The pyramidal cell region adjacent to the de ntate gyrus, not incl uding the granule cell layers, encompassed region CA3. The objective was positioned so that the hilus of the dentate gyrus was in the center fo r analysis of this region. A load referred to the percent area within the measurement field occupied by the reaction product, while Congo red staining referred to the percent ar ea stained with Congo red. All values for a given mouse and given brain area were averaged to represent a single value for that animal that was used in the statistical analysis.
143 VI. Statistical Analysis A total of 19 behavioral measures were obtained from the 9-behavioral task battery previously described. The tasks were divided into single day tasks and tasks that involved multiple test days. All of the si ngle day tasks (open field, balance beam, Ymaze and elevated plus maze) except stri ng agility were analyzed using one-way ANOVA. String agility was analyzed using the Kruskal-Wallis non-parametric test and post hoc Mann-Whitney U test. For multi-day tasks (Morris water maze, circular platform, platform recognition and radial arm water maze) both one-way ANOVAs and two-way repeated measure ANOVAs were empl oyed. Prior to statis tical analysis of multi-day measures, data was grouped into 2-5 day blocks (except for platform recognition) to facilitate analysis and presentation. Following ANOVA analysis, posthoc pair-by-pair differences between groups (planned comparisons) was resolved using the Fisher LSD test. Swim speed from the Morris water maze retention trial was calculated using the Mouse Clocker software and analyzed by one-way ANOVA; annulus crossings during this trial were analy zed by Mann-Whitney U-tests. All group comparisons were considered significance at p<0.05. Any non-performers (e.g. repeated circulars, consistent floaters, etc.) in a part icular task were eliminated from statistical analysis in that task. To group behavioral and/or pathologic measures based on their common factors, Factor Analysis (FA) was performed using Systat software. FA used all collected data, regardless of genotype or treatment, to relate measures into individual factors. Each factor measured a different component of be havior or cognition (i.e. reference memory, sensorimotor function, etc.). In this way, behavioral meas ures related to one another
144 could be determined, as well as how performa nce in one task might be predictive of performance in another task (see Leighty et al., 2004 for more de tailed explanation of FA). For Study I (Life-long Vaccination Study), an FA was performed using all 19 behavioral measures collected at both th e 4-6 month and 15-16 month time points. An additional DFA was carried out at the 15-16 month time point that included 8 A pathologic measures in addition to the 19 behavioral measures. The 8 pathology measures included 4 measures for total A burden (diffuse + compact) within 3 hippocampal areas (CA1, CA3 & dentate gyrus) and frontal cortex (CX), and 4 measures for Congo red staining (compact A only) within the same brain areas. To determine if the three experime ntal groups (NT, Tg+/Con and Tg+/A ) were distinguishable behaviorally from one anothe r, discriminant functi on analysis (DFA) was performed using all 19 behavioral measures for both the 4-6 month and 15-16 month time points. For the 4-6 month time point DFA was carried out with only the 7 measures (all cognitive-based) that loaded within Factor 1. Similarly, the 15-16 month time point DFA was also performed with only the 9 measures (all cognitive-based) that loaded within Factor 1. All DFA analyses were performed using the Systat software with two different DFA methods direct entry and stepwise-f orward. The direct entry method used all behavior measures availa ble, while the stepwise-forward method selected measures based on their variance cont ribution to best disc riminate between the three groups (see Leighty et al., 2004 fo r more detailed information on DFA methodology.
145 For Study II, Part B (Omega-3 Behavioral Study), statistical analysis of behavior was similar to Study I, as just described. In addition, separate behavioral analyses including only APP/PS1 and NT mice on either the standard or high n-3 fatty acid diets were evaluated and included after the initial Omega-3 Fatty Acid Study Part BI section and entitled Omega-3 Fatty Acid Study Part BII. For the Omega-3 Fatty Acid Study II ne urochemical and histological analyses, groups were compared using one-way ANOVAs, w ith post-hoc Fishers test to determine significance and included only the NT a nd APP/PS1 mice on any of the three experimental diets.
146 Results I. Life-long Vaccination Study Significant differences of transgenic and A immunotherapy effects are summarized in Table 4. Table 4. A summary of transgenic and imm unotherapy at both 4-6 and 15-16 month test points.
147 Sensorimotorand Anxiety-based Tasks. Open Field and Y-maze Entries. At 4-6 months of age, both Tg+/Con and Tg+/A exhibited increased open field activity (F ig. 6a) compared to NT control mice ( P < 0.02). However, at the later time point 15-16 months of age, there were no differences between the groups. Within th e Y-maze task for activity/exploration, arm entries of the Tg+/A group were elevated at both early ( P < 0.01) and late ( P < 0.05) test points compared to both Tg+/Con and NT groups (Fig. 8a). Balance Beam. In the balance beam task (Fig. 6b), 4-6 month old Tg+/A exhibited poorer performa nce than the NT mice; the Tg+/Con group showed an intermediate balance ability. However, at 15-16 months, all th ree groups exhibited poor balance ability as indicated by short tr ial times prior to falling from the beam. String Agility. Despite differences in activ ity (from open field or Y-maze entries) and balance performance, there were no differences at either test point in string agility (Fig. 6c). This shows that any ageand genotype/treatment-related deficits exhibited in sensorimotor tasks are task-s pecific and did not de leteriously affect performance in cognitive-based tasks. Elevated Plus-Maze. In the elevated plus-maze task for anxiety and/or emotionality at 4-6 months of age, an increased number of both open and closed arms entries was exhibited by Tg+/A mice, but not the Tg+/Con mice, compared with NT control mice (Fig. 7b and c). This is more re flective of an increase in activity level (in open field and Y-maze entries) and less of a decrease in anxiety. This is especially evident since arm entries in elevated plus-maze generally load with other activity measures in Factor Analysis (see below). De spite the increased number of entries, there
148 were no differences seen between all three groups in the percent time spent in the open arms at the younger test age (Fig. 7a). Howe ver, at the aged test point, Tg+/Con mice exhibited less anxiety by spe nding an increased percent time in the open arms compared to NT controls ( P < 0.02). Though there were no group differences in the number of entries into either closed or open arms at this age (Fig. 7b and c).
149 Fig, 6. Comparison of sensorimotor f unction in NT, Tg+/Con, and Tg+/A mice at 4 6 and 15-16 behavioral test poi nts. Data represent the mean S.E.M. (a) Open field activity, determined by open field line crossi ngs, was increased in young adult Tg+/Con and Tg+/A mice (b) Equilibrium/agility, as measured by time on a balance beam, was impaired in young Tg+/A mice. (c) String agility, as measured by forepaw grip suspension. *Significantly different from NT at the age indicated, with P < 0.05 or higher level of significance.
150 Fig. 7. Anxiety/emotionality, as determined in the elevated plus-maze by percent time in open arms and arm entries. NT, Tg+/Con, and Tg+/A mice were evaluated at adult (4-6 month) and aged (15-16 month) test points. (a) Aged Tg+/Con mice spent an increased percent time in open arms compared with NT controls. (b) The number of open arm entries was elevated in adult Tg+/A mice in comparison to both NT and Tg+/Con mice. (c) Adult Tg+/A mice also had an increased number of closed arm entries. Data represent the means S.E.M. *Significantly different from NT mice at P <0.05, **significantly different from adult NT and Tg+/Con mice at P <0.05.
151 Cognitive-based Tasks Y-maze Alternation. There was no transgenic effect at either test point for percent alternation within the Y-maze task (Fig. 8b), thus no prot ective effect of A vaccination could be observed for this task. Morris Water Maze Acquisition. To facilitate presentation and statistical analysis, the Morris maze acquisition data was evaluated as two 5-day blocks (Fig. 9). At the adult test point, there wa s an overall groups effect across both blocks [F(2,19)=7.37; P < 0.005], with the Tg+/Con exhibiting impair ment compared to the two remaining groups. Also across both blocks, the Tg+/C on were found to have significantly higher escape latency, compared to both NT ( P <0.005) and Tg+/A ( P <0.02) mice; the latter two groups did not differ in thei r acquisition performance. At the aged test point, 15-16 months, there was nearly an ove rall groups effect [F(2,15)=3.28; P =0.07). Planned comparisons revealed a significant impairment across both blocks for Tg+/Con mice compared with NT controls ( P <0.05). In contrast, th e performance of Tg+/A mice was no different from that of NT control mice. Additionally, across bot h blocks at both the young and aged test point, the Tg+/A were no different from the NT controls. Overall, A immunotherapy protected APP/PS1 transgen ic mice against memory impairment that was present in control Tg+ mice at an adult ag e and this protection wa s preserved into old age.
152 Fig. 8. Y-maze arm entries (a) and percent s pontaneous alternation (b) for adult and aged NT, Tg+/Con, and Tg+/A mice. Data represent the mean S.E.M. (a) Tg+/A mice had significantly more arm entries (e.g. in creased activity) compared with both NT and Tg+/Con mice at the 4 -6 month adult te st point. At the 1516 month aged test point, Tg+/A mice continued to exhibit an increas ed number of arm entries compared with NT controls. (b) There were no group differences at either age for percent spontaneous alternation. *Significantly different fr om NT and Tg+/Con at P <0.05, or higher level of significance.
153 Fig. 9. Morris water maze acquisition at adult (4-6 months) and aged (15-16 months) time points for NT, Tg+/Con, and Tg+/A mice. The 10 days of acquisition, as measured by latency to find a submerged stationary plat form, are presented in two 5-day blocks. For both blocks and at both test points, Tg+/Con mice were significantly impaired, while performance of Tg+/A mice was no different than NT mice. **Significantly different from both other groups at P <0.02 or higher level of significan ce, *significantly different from NT group at P <0.05.
154 Morris Water Maze Retention. The Morris water maze probe trial is presented below in Fig. 10 for both adult (4-6 mont hs) and aged (15-16 months) test points. Both adult NT and Tg+/A groups showed an exclusive pr eference for the quadrant that formerly contained the submerged platform (Q2). In contrast, the adult Tg+/Con mice showed no quadrant preference as seen by their having similar percent time spent in each quadrant. Also, the Tg+/Con mice had signifi cantly fewer annulus crossings compared to the NT mice ( P <0.05), while the Tg+/A mice had similar annulus crossings to the NT group. This shows that at the adult test point, A immunotherapy protects against cognitive impairment of memory retention th at was present in Tg+/Con mice. At the aged test point, similar to the adult mice, the NT group showed an exclusive quadrant preference for the former platform-containing qu adrant (Q2). However, neither group of Tg+ mice showed such a preference. In addi tion, at the aged test point, there were no group differences in annulus cr ossings; neither were there any group differences in swim speed at either test point exhi bited during this probe trial.
155 Fig. 10. Probe trial testing for reference memo ry retention, assessed on the day following the completion of water maze acquisition. A single 1-minute trial was done, with the percent of time spent in the quadrant former ly containing the submerged platform (Q2) and the number of annulus crossings determin ed. At the 4-6 mont h test point, NT and Tg+/A mice showed an exclusive quadrant pr eference for Q2 compared with all other quadrants, while Tg+/Con mice exhibited no quadr ant preference. *Significantly higher than all other quadrants at P<0.0001 (NT group) or P<0.01 (Tg+/A group). Tg+/Con mice also had significantly fewer annulus cr ossings compared with NT mice (P<0.05), while there was no difference between the ot her two groups. At 15-16 months of age, only NT mice showed a quadrant preference fo r Q2 (*significantly higher than all other quadrants at P<0.0001). There were no group differences in annulus crossings at this age.
156 Circular Platform. There were no impairments in latency to find the escape hole (Fig. 11a and b) or number of head pokes/errors into non-escape holes (Fig. 12a and b) for Tg+/Con mice at either test point in the circular platform task for spatial learning/memory. Therefore, no protective effect of A immunotherapy could be observed in this task. In addition there were no differences at either the adult or aged test points between any of the groups; however, stro ng overall effects of learning across days were present at both adult [F(7, 133)=10.75; P <0.0001] and aged [F(7,112)=7.24; P <0.0001] test point in circ ular platform latency.
157 Fig. 11. Spatial learning/memory, as determined by the circular platform task latency to find the escape hole across 4 blocks consisting of 2 days each. NT, Tg+/Con, and Tg+/A were behaviorally evaluated at bot h adult (4-6 months) and aged (15-16 month) test points. There we re no group differences at either adult (a) or aged (b) test points. However, there was a strong overall effect of learning acro ss days at both adult and aged test points.
158 Fig. 12. Spatial learning/memory, as determined by the circular platform task the number of errors, indexed as the nu mber of head pokes into holes which did not contain the escape hole, across four 2-day bl ocks. NT, Tg+/Con, and Tg+/A were behaviorally evaluated at both adult (4-6 months) and aged (15-16 month) test points. There were no group differences at either adult (a) or aged (b) test points.
159 Platform Recognition. At the 4-6 month test point, all groups collectively exhibited a strong learning effect across days [F(3. 54)=20.91; P <0.0001], and there were no overall group differences across the 4 days of testing (Fig. 13). C ontrary, there was an overall groups effect over all 4 days of testing at the 15-16 month time point [F(2,15)=3.97; P <0.05]. At this age, th e Tg+/Con mice were impaired overall versus NT controls ( P <0.02) and specifically on da ys 1 and 3 compared to the NT mice. In contrast to the impaired performance of the Tg+/Con mice, the aged Tg+/A mice performed similar to NT controls overall and for each of the 4 test days (Fig. 13). However, by the last day of testing, there were no group differences in escape latency. This indicates that the Tg+/Con were eventually able to reduce their latencies comparable to the remaining two groups. In a comparison across bot h test points, the NT and Tg+/A mice showed similar latencies indicating that their performa nce did not decline with increased age. In contrast, the escape latency of the Tg+/Con mi ce was significantly higher at the aged test point compared to the adult test point ( P <0.05). This indicates that A immunotherapy protected Tg+ mice from an age-dependent impairment in switching from the spatial strategy of Morris maze to the recognition/ identification strategy of platform recognition.
160 Fig. 13. Platform recognition testing for the abi lity to search/ identify a variably-placed and conspicuously marked platform over 4 days of testing, with latency to swim to the platform being measured. At the 4-6 m onth test point, there were no differences between NT, Tg+/Con, and Tg+/A mice to find the variably placed platform, with all groups reducing their escape latency over da ys. At the 15-16 month test point, however, aged Tg+/Con mice were impaired ove rall, as well as during days 1 and 3 of testing. By contrast, Tg+/A mice performed identically to NT mice. *Significantly different from NT mice at P <0.005, **significantly different from both NT and Tg+/A mice at P <0.025 or higher level of significance.
161 Radial Arm Water Maze. The number of errors at th e 4-6 month test point in the Radial Arm Water Maze (RAWM) task is s hown in Fig. 14 below. At this age, the data were collected across three 3-day blocks for T1(randomized initial trial), T4 (final acquisition trial), and T5 (delay ed retention trial). For bl ock 1 (B1), both groups of Tg+ mice showed memory impairment in the final acquisition trial in addition to the delayed retention trial compared to the NT group, whic h was effective in lowering their errors by T4 and T5. In contrast, by block 2, the Tg+/A mice had similar performance to the NT mice on T5, whereas the Tg+/Con mice remained impaired compared to the NT controls. By block 3, the Tg+/A group performed similar to the NT control mice on both T4 and T5 trials of working memory; in sharp cont rast the Tg+/Con exhibited marked memory impairment on both T4 and T5 compared to the NT mice. Over all three blocks of testing, there was a significan t effect of groups for both T4 [F(2,18)=11.30; P<0.001] and T5 [F(2,18)=9.28; P<0.005] in that both Tg+ groups were impaired compared to the NT controls overall for both T4 and T5 ( P <0.02 or higher level of significance). This overall impairment of Tg+/A mice reflects the fact that it re quired them several blocks of testing to reduce their T4 and T5 errors to the level of the NT c ontrols. Fig. 15 below illustrates the latency in seconds across three 3-day blocks of RAWM behavioral testing at the adult test point. Initially within B1, Tg+/A mice had significantly higher latencies on T4 and T5 compared to the NT mice. However, by the remaining blocks, they had similar latencies/working memory to the NT control group. Although also not statistically different from NT controls or Tg+/A mice, Tg+/Con mice exhibited consistently higher latencies than those two groups for T4 and T5 within the last two blocks of testing.
162 Fig. 14. Working memory in the RAWM task at the 4-6 month test point, with errors being evaluated for T1, T4, and T5 over three 3-day blocks. T4 and T5 are indices of working memory. By the last block of testing, Tg+/A mice were no different than NT mice in working memory performance, wh ile Tg+/Con mice consistently performed poorly over all three blocks in being unable to reduce their number of errors by T4 and T5 in any block. All data are means (S.E.M. ). *NT group significantly better than both Tg+/Con and Tg+/A groups at P <0.05 or higher level of significance, NT group significantly better th an Tg+/Con group at P <0.05 or higher level of significance.
163 Fig. 15. Working memory in the RAWM task at the 4-6 month test point, with latency to find the hidden platform being evaluated fo r T1, T4, and T5 over three 3-day blocks. Within the first block of testing (B1), the NT mice performed significantly better than the Tg+/A mice on T4 and T5, with the Tg+/Con mice having intermediate performance. However, by the last two blocks of testing, Tg+/A mice were no different than NT mice in working memory performance. Although not statistically different from NT controls, Tg+/Con mice consistently had hi gher latencies and we re unable to reduce their latencies by T4 and T5 in any block. All data are means (S.E.M.). *NT group significantly better than Tg+/A group at P <0.05 or higher level of significance.
164 The number of errors at the 15-16 month te st point in the RAWM task is shown in Fig. 16 below using errors for T1, T4, a nd T5. Comparisons between the two test points for RAWM can not be made because prior to the aged test point, the walls at the end of each swim arm of the RAWM aluminum insert were painted black. For the first three blocks of behavioral te sting, all groups performed sim ilarly with the NT controls having consistently fewer erro rs than the two Tg+ groups. By block 4, however, the Tg+/Con mice exhibited significan tly more errors compared to the NT controls for T4 ( P <0.01) and T5 ( P <0.01), while the Tg+/A mice reduced their errors similar to that of the NT mice. In addition, Tg+/Con mice we re significantly impaire d by having a higher number of errors in overall T5 performance ( P <0.05), in contrast to the performance of Tg+/A mice which was identical to that of NT mice. The protective effect of A immunotherapy to Tg+ mice is underscored by RAWM escape latency (Fig. 17), particularly during the last block. During trial 4 of block 4, Tg+/A mice performed identical to NT mice, while Tg+/Con we re impaired compared to both Tg+/A and NT mice. Likewise, during T5, the Tg+/A mice had similar latencies to the NT mice, while the Tg+/Con mice were significantly impaired compared to the NT mice. In addition, Tg+/Con mice were impaired by having substa ntially higher latencies in overall T4 ( P <0.02) and overall T5 ( P <0.02) across all four blocks in contrast to Tg+/A which performed similar to NT mice in those trials.
165 Fig. 16. RAWM testing for working memory at the 15-16 month test point, with number of errors indicated for T1, T4, and T5 over four 3-day blocks of testing. For the first three blocks of testing, all groups perf ormed similarly. By the final block, however, Tg+/A mice reduced their T4 and T5 errors down to the level of NT controls, while Tg+/Con mice made significantly higher numbers of errors compared with NT mice. Tg+/A group significantly diffe rent from NT group at P <0.05, *Tg+/Con group significantly different from NT group at P <0.05. All data are means (S.E.M.).
166 Fig. 17. RAWM testing for working memory at the 15-16 month test point, as indexed by escape latency, across four 3-day blocks of testing. For the fi rst three blocks of testing, all groups performed similarly except that both Tg+ groups had higher latencies within the first block, T5. Also, the Tg +/Con showed significantly higher latencies compared to NT mice on T4 in blocks 2 and 3, while Tg+/A mice performed similar to the NT mice. From T1 to T4, within the final block, Tg+/A mice performed identical to NT controls and were significantly bette r than Tg+/Con mice on T4. On T5, Tg+/A mice were again no different in performance vs. NT controls, while Tg+/Con mice were substantially impaired (high latencies). NT mice significantly lower latency compared to both Tg+/Con and Tg+/A groups at P <0.05 or higher level of significance, *Tg+/Con group significantly worse than NT group at P <0.05. **Tg+/Con significantly worse than both NT and Tg+/A groups at P <0.05 or higher level of si gnificance. All data are means (S.E.M.).
167 A Histopathology At around 15 months into A immunotherapy (and at 17 months of age), A immunostaining and Congo Red staining were performed on brain sections from the frontal cortex, dentate gyrus and both CA1 and CA3 regions of the hippocampus in Tg+ mice. The results shown in Tabl e 5 illustrate that long-term A immunotherapy did not alter total or compact A deposition in any of the brain regions analyzed. However, one of the four Tg+/A mice did have appreciable reductions in mean A levels ranging from 76 to 86% compared to Tg+/Con mice.
168 Table 5. Total A (diffuse and compact) and compact A (Congo Red) for Tg+/Con (n=4) and Tg+/A (n=4) at 17 months of age within frontal cortex and hippocampus
169 FA and Correlation Analyses Table 6 shows the factor analysis of behavioral measures with and without A histopathologic measures in order to determ ine the underlying relati onships between such behavioral measures and A deposition. At the adult (46 months) test point, of the 19 behavioral measures analyzed, 15 measures loaded on three principle factors which together account for more than 70% of the to tal variance. All measures of RAWM and Morris Water maze loaded heavily under fact or 1, which was clearly cognitively-based and accounted for more variance (3 2.5%) than either of the othe r two factors. In contrast, factor 2 was strongly sensorim otor/anxiety based, including activity/exploratory, elevated plus maze and string agility measures. At the aged (15-16 month) test point, many of the above 19 behavioral measures loaded into five principle factors (Table 6). Similar to the adult test point, factor one was largely cognitively based and f actor 2 was primarily sensorimotor/anxiety based. Factor 1 agai n included all measures of RAWM and Morris water maze in addition to both measures of pl atform recognition; this factor, therefore, encompassed working memory, reference learning/memory and recognition/ identification. Also, both measures of platform recognition loaded separately into factor 3. Sensorimotor/ anxiety-based factor 2 reta ined three measures present at the earlier time point and gained an additional elevated plus-maze measure (number of closed arm entries). Two other measures (time in pl us-maze open arms and string agility) which previously loaded together in factor 2 at the ad ult test point, loaded separately into factors 4 and 5, respectively. Inclusion of the eight histopathologic measures at the aged test point resulted in five of these measures load ing into factor 1 with essentially all of the RAWM and Morris water maze measures. The fi ve histopathologic measures in factor 1
170 included all four brain ar eas stained for total A and one area that was stained for compact A deposition. Compact A deposition in region CA3 of the hippocampus loaded with various sensorimotor and cognitive -based measures in fa ctor 2. Factor 3 was largely activity/exploratory-based, while factor 4 included only Congo Red staining in the cortex and dentate gyrus. Lastly, factor 5 solely contained the time in plus-maze open arms. Therefore, similar factor load ings were seen with and without A histopathology, although the pathologic measures did modi fy the behavioral loadings somewhat. At the aged test point, correlation an alyses were performed between the nine cognitive-based measures from factor 1 and all eight of the A histopathologic measures. This analysis involved all Tg+ mice (4 Tg+/A mice and 4 Tg+/Con mice). Of a total of 72 total correlations that were done, ther e were no significant correlations found.
171 Table 6. Factor Loadings of behavioral measur es, with and without pathologic measures.
172 DFA Discriminant function analysis (DFA) wa s used in order to determine if the behavioral performance of each of the three groups (NT, Tg+/Con and Tg+/A ) could be distinguished from each other at both 4-6 m onth and 15-16 month test points. Table 1.4 shows a summary of two DFA methods including the direct entry method (including all behavioral measures) and th e step-wise forward method (selects behavioral measures from all the measures ba sed on their contribution to the variance) at both test points. The direct entry DFA could not distinguish the three gr oups at either test point based on their behavioral performance. However, at the 4-6 month test point, the step-wise forward DFA could completely dist inguish between all three groups such that their rank order of performance was as follows: NT>Tg+/A >Tg+/Con. Additionally, at the 15-16 test point, the step-wise forward DFA successfully distinguished between the two Tg+ groups, even with the lesser number of animals. At both test points a sensorimotor measure (Y-maze entries) and a cognitive-based measure (either RAWM trial 5 on last block or platform recognition av eraged latency) were retained as providing the maximal discriminability. Additional DFAs were performed utilizi ng only the cognitive-ba sed measures that loaded on factor 1 in FA (see Table 1.3). Th e direct entry DFA at the adult test point resulted in significant discrimination betw een the NT group and both Tg+ groups using all seven cognitive measures in factor 1. In contrast, direct entry DFA at the aged test point was unable to successfully distinguish between any of the three groups using the nine behavioral measures from the factor 1. However, all three groups were completely
173 distinguished from each other at the adu lt test point by step-wise forward DFA (rank order by performance was NT>Tg+/A >Tg+/Con). The measures retained as providing maximal discrimination were cognitive-bas ed, including measures from RAWM and Morris water maze. Similar to the adult test point, step-wise forward DFA at the aged test point could successfully discriminate the NT group from the Tg+/Con group. More importantly, step-wise forward DFA could not distinguish the NT group from the Tg+/A group based on their beha vioral performance. Th ree measures, including two from RAWM and one from platform recogni tion, were retained as providing maximum discrimination between the NT group and the Tg+/Con group. Fig. 18 shows canonical scores plot of both the sevenand nine-measure stepwise forward DFAs at 4-6 month and 15-16 test points, respectively. The measures that provided the maximal discriminati on between groups were RAWM trial 5 performance overall and RAWM trial 5 perfor mance on the last block of testing (Table 7). Between 85-88% of the mice at the adult test point and 69-75% of the mice at the aged test point were correctly classi fied by the step-wise forward DFA.
174 Table 7. Summary of discriminant function analyses.
175 Fig. 18. Canonical score plots of step-wise forw ard discriminant function analyses used to compare the overall cognitive pe rformance of NT, Tg+/Con, and Tg+/A groups. Each symbol represents the cognitive perfor mance of one animal graphed from the two linear functions derived in the DFA. At th e 4-6 month test point involving all seven cognitive measures from factor 1 (upper), all groups could be distinguished from one another in cognitive performance (rank order: NT>Tg+/A >Tg+/Con). At the 15-16 month test point involving a ll nine cognitive measures fr om factor 1 (lower), the performance of the NT and Tg+/A mice could not be disti nguished from one another, while NT mice were clearly distinguis hable (better than) Tg+/Con mice.
176 II. Omega-3 Fatty Acid StudySurvival Analyses Fig. 19 shows profound differences in the su rvivability of F1 generation APP/PS1 (n=13) and PS1 (n=7) mice that were fed th e omega-3-deficient (omega-6 only) diet beginning at 6 weeks of age. All of the PS1 mice survived on this diet for 120 days, at which time the mice were euthanitized. In sh arp contrast, the APP/ PS1 mice experienced 100% mortality, mostly within 20 days of begi nning the diet. This resulted in a highly significant difference in the cumulative propor tion surviving, as determined by the CoxMantel test ( P < 0.0001). In addition to the survival analysis w ithin the F1 generation, F2 generation APP (n=7), PS1 (n=9) and NT (n=9) were started on the omega-3-deficient diet at an average of 6 weeks of age. As shown in Fig. 20, a ll genotypes showed similar survival on the experimental diet ( P > 0.05), with the majority of th e mortality occurring between 5 and 10 days after beginning the diet. By comparing survivability within the PS1 transgenic groups for F1 versus F2, it is evident that it is a combination of the interaction between the PS1 transgene and the background strain that determines survival on the omega-3-deficient di et (Fig. 21). There is a significant difference in the cumulative proportion surviving ( P < 0.0001), with the F1 PS1 mice having complete survival in cont rast to the 100% mortality seen in the F2 generation of PS1 mice. In reference to the difference in the background between generations, the F1 generation was composed of a higher amount of B6 with a lesser amount of C57 present as compared to the F2 generation mice. This suggests that the interaction between the B6 background and the PS1 transgene provided limited
177 protection against a diet that comp letely lacks omega-3 fatty acids.
178 Fig. 19. Survivability analysis showing days on the omega-3-deficient diet wherein the F1 PS1 mice had no mortality through 3 months after starting on the experimental diet. However, the F1 APP/PS1 mice had no su rvival on the omega3-deficient diet. 020406080100120140Days -0.2 0.0 0.2 0.4 0.6 0.8 1.0Cumulative Proportion Surviving P P S S 1 1 A A P P P P / / P P S S 1 1
179 Fig. 20. Survivability analysis showing days on th e omega-3-deficient diet. All 3 groups of the F2 mice had similar mortality after starting on the experimental diet. 051015202530Days -0.2 0.0 0.2 0.4 0.6 0.8 1.0Cumulative Proportion Surviving P P S S 1 1 A A P P P P N N T T
180 Fig. 21. Survivability analysis comparing F1 PS1 mice to F2 PS1 mice, showing a significant difference in survival between the two generation within the same genotype. As shown, the PS1 transgene is not completely protective against mortality on a diet that is devoid of omega-3 fatty acids. 020406080100120140Days -0.2 0.0 0.2 0.4 0.6 0.8 1.0Cumulative Proportion Surviving F F 1 1 P P S S 1 1 F F 2 2 P P S S 1 1
181 III. Omega-3 Fatty Acid Study: Behavioral, Pathologic, and Neurochemical Assessments Following a pre-diet baseline determin ation of body weight, all mice were weighed weekly throughout the 5-7 month duration of the study. Averaging all weekly weight determinations, no overall differences were found between any of the six groups of mice, nor were there any differences in fi nal weights at the end of the study (data not shown). Significant differences in behavioral perf ormance due to transgenicity (standard diet NT vs. standard diet Tg +) or diet are summarized in Table 8. Table 9 presents differences in behavioral performance due to transgenicity or diet (standard vs. high n-3) for only APP/PS1 and NT mice; all APP mice and also all high n-6 diet groups were excluded.
182 Table 8. A summary of transgenic and dietary behavioral eff ects (high n-3 and high n-6) in NT and Tg+ mice.
183 Table 9. A summary of transgenic and dietary be havioral effects (high n-3 vs. standard diet) in NT and APP/PS1 mice.
184 Sensorimotorand Anxiety-Based Tasks For all sensorimotorand anxiety-based ta sks, there were no effect of high n-3 or high n-6 diets on performance of NT mice (Fig s. 22-24). Thus, neither diet affected sensorimotor abilities or level of anxiety in NT mice. Some effects of transgenicity and diet were, however, evident for Tg+ mice, as indicated below. Open Field and Y-maze Entries. The Tg+ high n-3 mice had increased open field activity (Fig. 22a) comp ared to all NT groups ( P <0.01) and Tg+ standard mice ( P <0.05). Similarly, within the Y-maze task for activity/exploration, arm entries of the Tg+ high n-3 group were elevated compared to all NT groups ( P <0.01) and Tg+ standard mice ( P <0.02) (Fig. 24a). Thus, a high n-3 diet increased activity/exploration in Tg+ mice. Balance Beam. In the balance beam task (Fi g. 22b), all three groups of Tg+ mice exhibited equally poor balance ab ility compared to NT standard-fed mice, as indicated by short trial times prior to falling from the b eam. A strong overall eff ect of transgenicity was present [F(1,47)=22.47; p<0.0001]. These re sults indicate ther e was no beneficial effect on balance for either hi gh n-3 or n-6 diets in Tg+ mice. String Agility. Although a significant overall effect of transgenicity was evident (p<0.002, Mann-Whitney U test], neither of th e Tg+ groups fed the high n-3 or high n-6 diets exhibited deficits in stri ng agility compared to NT standard-fed mice (Fig. 22c). In contrast, the standard Tg+ mice did show a defi cit in string agility co mpared to standard NT mice ( P <0.05). Nonetheless, all 3 Tg+ groups showed the same level of agility. Elevated Plus-Maze. In the elevated plus-maze task, there were no significant differences in either the percent of time sp ent in the open arms or the number of open arm
185 entries between any of the Tg+ and NT groups (Fig. 23a and b). As well, there were no overall effects of transgenicity for either of these measures. Despite the lack of difference in open arm entries, both the high n-3 and high n-6 had in creased closed arm entries compared to the standard Tg+ mice, but not vs. NT standard-fed mice (Fig. 23c). Within the NT mice, there was no difference in closed arm entries between any of the diet groups.
186 Fig, 22. Comparison of sensorimotor function in NT and Tg+ mice fed either standard, high n-3 or high n-6 diets. Data represent the mean S.E.M. (a) Open field activity, determined by open field line crossings, wa s increased in high n-3 Tg+ mice. (b) Equilibrium/agility, as measured by time on a balance beam, was impaired in all three Tg+ diet groups. (c) String agility, as measured by forepaw grip suspension was impaired only in Tg+ standard mice, although there was no difference in agility between Tg+ groups. *Significantly different from NT standard group, with P < 0.05 or higher level of significance. Significantly different from all NT mice and standard Tg+ mice. a ) b ) c )
187 Fig. 23. Anxiety/emotionality, as determined in the Elevated Plus-maze by percent time in open arms and arm entries. No differences were evident in either percent time in open arms (a) or number of open arm entries (b) within any of the NT and Tg+ groups. However, both high n-3 and high n-6 Tg+ mice had an increased number of closed arm entries vs. standard Tg+ mice. Data repr esents the mean S.E.M. Significantly different from standard Tg+ mice at P < 0.05. a ) c ) b )
188 Cognitive-based Tasks As shown in Figures 24 37, there wa s no effect of a high n-3 diet on performance of NT mice in any of the cognitive -based tasks. This was also the case for NT mice on a high n-6 diet, with the excepti on of n-6 induced de leterious effects on Morris maze retention and circular platform errors. Thus, these diets had minimal impact on cognitive performance in normal mice. For Tg+ mice, however, there were some effects of diet on cognitive performance. These effects were largely only seen in comparison to standard NT mice, but not standard Tg+ mice. Y-Maze Alternation. In Y-maze testing for spontaneous alternation, there was no overall effect of transgenicity, nor were there any differences among the NT and Tg+ groups (Fig. 24b). Morris Water Maze Acquisition. In addition to the lack of impairment of Tg+ mice in Y-maze spontaneous alternation, ther e also was no transgenic impairment in Morris water maze acquisition over 10 days of testing. For both 5-day blocks, all NT groups and all Tg+ groups performed similarly to each other (Fig. 25). Even on the final day of testing, there were no group differences in acquisition (data not shown). Morris Water Maze Retention. Within the NT group, the standard and high n-3 diet mice, but not the high n-6 mice, show ed an exclusive quadrant preference; nonetheless, all three NT groups had a similar number of annulus crossings. Despite no impairment in Morris maze acquisition, sta ndard Tg+ mice showed reference memory impairment compared to standard NT mice in both time spent in the former platformcontaining quadrant and annulus crossings (Fig. 26). Surpri singly within the Tg+ groups, only the high n-6 Tg+ mice had an exclusiv e quadrant preference, whereas the high n-3
189 Tg+ mice only showed a partial quadrant pref erence for Q2 versus Q1. Similarly, both the standard and high n-3 Tg+ mice exhibi ted a lower number of annulus crossings during the memory retention trial compared to the standard NT mice. However, there were no differences in annulus crossing when the 3 Tg+ groups we re compared directly With inclusion of only APP/PS1 mice in both standard and high n-3 groups (Fig. 27), only the standard Tg+ group remained co mpletely impaired in memory retention versus standard NT mice this, for both qua drant preference and annul us crossings. In contrast, the high n-3 APP/PS1 mice showed a partial quadrant preference for Q2 compared to Q1 and Q4. Consistent with this modestly improved retention of n-3 APP/PS1 mice, they also exhibited a simila r number of annulus crossings compared to standard NT mice (Fig. 27). By contrast standard APP+PS1 mice exhibited a lower number of annulus crossings vs. standard NT mice. The only modest benefit of a high n3 diet in APP+PS1 mice is underscored by the fa ct that there was no difference in annulus crossings between APP+PS1 mice on the high n3 diet vs. those on a standard diet. Consistent with their lack of quadrant pref erence, APP/PS1 standard mice exhibited a lower number of annulus crossings vs. standard NT mice. The Morris maze retention data suggest that both n-3 and n-6 fatty acid supplementation can partially correct memory re tention deficits that are otherwise present in standard-fed Tg+ mice, although the high n-6 di et was actually delete rious in NT mice. During the probe trial, there were no differe nces in swim speed among the Tgand Tg+ groups
190 Fig. 24. Y-maze arm entries (a) and percent s pontaneous alternation (b) for standard, high n-3 and high n-6 diet groups within NT and Tg+ mice. Data represents the mean S.E.M. (a) High n3 Tg+ mice had significantly mo re arm entries (e.g., increased activity) compared to al l three NT groups and standard Tg+ mice. (b) There were no group differences for either NT or Tg+ gr oups in percent spontaneous alternation. *Significantly different from st andard-fed Tg+ mice at P< 0.02. a ) b )
191 Fig. 25. Morris water maze acquisition for standard, high n-3 and high n-6 diet groups of NT and Tg+ mice. The 10 days of acqui sition, as measured by latency to find a submerged stationary platform, are presented in two 5-day blocks. For both blocks and for all three NT and three Tg+ groups, there were no di fferences in acquisition.
192 Fig. 26. Probe trial testing for reference memo ry retention, assessed on the day following the completion of water maze acquisition. A single 1-min trial was done, with the percent of time spent in the quadrant former ly containing the submerged platform (Q2) and the number of annulus crossings dete rmined. Within the NT groups, both the standard-fed and high n-3 mi ce showed an exclusive quadrant preference for Q2 compared to all other quadrants. In sh arp contrast, the high n-6 NT mice showed no quadrant preference. Within the Tg+ mice, the standard-fed mice showed no quadrant preference, while the high n-3 mice showed a partial quadrant preference for Q2 compared to Q1 ( = significantly higher than Q1 at P <0.05). Also, the high n-6 Tg+ mice showed a significant exclusive quadran t preference for Q2 versus all other quadrants. (* = significantly high er than all other quadrants at P <0.05 or higher level of significance). No differences were seen in number of annulus crossings within the NT mice, however the standard and high n-3 Tg + had significantly fewer annulus crossing compared to standard NT mice (# signifi cantly different from standard NT at P <0.05), although there were no differences in a nnulus crossings among the 3 Tg+ groups.
193 Fig. 27. Probe trial testing for reference memo ry retention, assessed on the day following the completion of water maze acquisition, including only NT and APP/PS1 groups. Within the NT groups, both the standard-fed and high n-3 mice showed an exclusive quadrant preference for Q2 compared to all other quadrants (* si gnificantly higher than all other quadrants at p<0.05 or higher level of significance). With in the APP/PS1 mice, the standard-fed mice showed no quadrant preference, while the high n-3 mice showed a partial quadrant preference for Q2 compared to Q1 and Q4 ( = significantly higher than Q1 and Q4 at P <0.05 and P <0.02, respectively). No differences were seen in number of annulus crossings within the NT mice, however the standard APP/PS1 mice had significantly fewer annulus crossing compared to standard NT mice (# significantly different from standard NT at P <0.02), while there was no difference between the high n3 APP/PS1 mice and standard NT mice.
194 Circular Platform. Assessment of spatial learning/memory in the circular platform task showed marked differences betw een the diet groups on the last day of this 8-day task. Although there was no transgenic effect in either number of head pokes/errors into non-escape holes or latency to find the escape hole, there were dietary differences within the NT and Tg+ groups. Fi g 28 shows that the high n-6 NT mice had impaired spatial learning/memory, as evidenced by increased errors compared to both standard NT mice and high n-3 NT mice. Howe ver, there were no differences within the NT groups in latency. Within the Tg+ groups the high n-6 mice were also impaired in errors (vs. standard Tg+ mice) and latency (vs. standard NT mice) to find the escape hole. For both NT and Tg+ genotypes, high n-3 mice performed similar to standard fed mice. Since there were no differences in performa nce between genotypic groups for any diet, the two genotypes were combined. Fig 29 shows the resultant strong eff ect of diet in the circular platform task. Together, NT and Tg+ mice fed the high n-6 diet had increased errors compared to both standard and high n-3 diet groups. Similarly, the combined high n-6 diet group had increased latency compar ed to the combined standard diet group. These data indicate that a di et high in n-6 fatty acids induces impairment in spatial reference learning/memory, while a diet high in n-3 fatty acids does not induce such an impairment.
195 Fig. 28. Spatial learning/memory, as determined by the circular platform task by number of head pokes/ errors (a) into non-escape holes and latency (b) to find the escape hole across 8 days of testing, with the last day of testing shown above. (a) High n-6 NT mice showed impairment by an increase of errors compared to both standard and high n-3 mice (** significantly different from st andard and high n-3 NT mice at P <0.02). Similarly, high n-6 Tg+ mice had increased errors compar ed to standard Tg+ (*p<0.05), with the high n-3 Tg+ mice having similar performance to the standard Tg+ group. (b) Despite a difference in errors, there were no differen ces between the NT mi ce with respect to latency to find the escape hole. However, the high n-6 Tg+ were impaired in latency compared to the standard NT mice ( significantly different from standard NT mice at P 0.05). a ) b)
196 Fig 29. The last day of circular platform shows a strong effect of diet by combining both genotypes. Together, NT and Tg+ mice on the high n-6 diet were impaired, as shown by a high number of errors into non-escape holes versus both standard and high n-3 groups (** = P <0.01 or higher level of significance). Also, the high n-6 mice had increased latency to find the escape hole compar ed to the standard diet group (* = P <0.05).
197 Platform Recognition. Within the NT groups, all mice collectively showed a strong learning effect across the four days of testing, with no differences between the diet groups overall or at individual days (Fig. 30a). In contrast, the standard Tg+ mice were impaired overall versus the standard NT mi ce (Fig. 30b; p<0.05). Even in comparison to the standard Tg+ groups poor performance, hi gh n-6 Tg+ were signif icantly worse than all five other groups, including both standard and high n-3 Tg+ mice. In contrast to the impaired high n-6 Tg+ performance, the high n-3 Tg+ group performed significantly better overall (p<0.005). Although performan ce of high n-3 Tg+ mice was no different from standard NT mice overall, their performance still was not improved enough to be significantly better than sta ndard Tg+ mice. High n-3 Tg + mice did have significantly lower latencies on days 2, 3, and 4 versus hi gh n-6 Tg+ mice. The platform recognition data indicate that a diet high in n-3 fa tty acids protects AD transgenic mice from impairment in recognition/id entification. In addition, a di et with a high amount of n-6 fatty acids induces an impairment in recognitio n/identification that is even greater than that normally present in Tg+ on a standard diet. Inclusion of only APP/PS1 mice for bot h standard and high n-3 Tg+ groups revealed marked differences in recognition/identification within this task. Standard APP/PS1 mice were impaired overall (p<0.02) and specifically on days 3 and 4 versus the standard NT mice (Fig. 31b; p<0.02). In sharp contrast to the impaired standard APP/PS1 performance, the high n-3 APP/PS1 group performance was no different from the standard NT mice on all days and over all 4 days of testing. Moreover, by day 4 of the task, the high n-3 APP/PS1 mice performe d significantly better (had lower latency) than the standard APP/PS1 and were no differe nt from standard NT mice. These data
198 indicate that a diet high in n-3 fatty acids protects AD transgenic mice from impairment in recognition/identification and by the end of the task fu lly reversed the transgenic impairment present in standard-fed mice.
199 Fig. 30. Platform recognition testing for the ability to search/ identify a variably-placed and conspicuously marked platform over 4 days of testing, with latency to swim to the platform being measured. (a) There were no differences in latency between the diet groups of the NT mice, with all groups reducing their latency over days. (b) The Tg+ standard-fed mice, however, were impaired overall vs. NT standard-fed mice. In addition, the high n-6 Tg+ mice were signifi cantly more impaired compared to the standard Tg+ mice. In sharp contrast, the high n-3 Tg+ mice performed significantly better compared to the high n-6 Tg+ groups overall and specifically during days 2 through 4 of testing, although their performance wa s not statistically better than that of standard Tg+ mice. Significantly different from standard NT mice overall at P <0.05, significantly different fr om high n-3 Tg+ mice at P <0.02 or higher level of significance, **significantly different from all othe r groups in overall performance at P <0.05 or higher level of significance.
200 Fig. 31. Platform recognition, as assessed within NT and APP/PS1 standard and high n-3 diet groups. (a) There were no differences in latency between the diet groups of the NT mice, with both groups reducing their latency ov er days. (b) The Tg+ standard-fed mice, however, were impaired overall compared to standard NT mice. In sharp contrast, the high n-3 Tg+ mice performed significantly bett er compared to the standard APP/PS1 mice by day 4 of testing. Significantly different fr om high n-3 APP/PS1 mice at P <0.05, significantly different from standard NT mice at P <0.05 or higher level of significance, **significantly differe nt from standard NT overall at P <0.02. a ) b )
201 Radial Arm Water Maze. RAWM data are presented in terms of errors (Figures 32-34) and latency (Figures 35-37) to find the goal arm. All of the NT groups performed similar to each other across all three blocks (Fig. 32a) e xhibiting no differences due to dietary intervention. In addition, all NT groups showed learning by exhibiting a significant reduction of errors from T1 to T4 and retaining that low number during T5 for both blocks 2 and 3 as well as overall three blocks. In contrast, the standard-fed Tg+ mice (Fig. 32b) had significantly more erro rs (were impaired) as compared to the standard-fed NT mice for T4 and T5 during bl ock 2, and during T5 for the last block of testing. Within the Tg+ groups, there was no si gnificant benefit of either experimental diet. However, the Tg+n-3 did make signi ficantly more errors on block 1 T5 as compared to the standard Tg+ mice. W ith that exception, there were no other group differences within the Tg+ groups. There were no differences in overall T4 or T5 errors within the NT groups (Fig. 33). Regarding Tg+ mice, only the standard Tg+ mice were impaired overall on T4 compared to the standard NT mice, while the high n-3 and high n-6 Tg+ performed similar to standard NT mice (Fig. 33a). More importantl y, however, the overall performance of high n-3 and high n-6 Tg+ mice wa s not significantly better than standard Tg+ mice. Moreover, all 3 groups of Tg+ mi ce were equally impaired in overall T5 errors versus standard NT mice (Fig. 33b). Comparison between only the standard a nd high n-3 dietary groups including only NT and APP/PS1 mice is presented as overall T4 and T5 errors in Figure 34. Regarding the NT mice, both groups performed similar to each other for both T4 and T5. Analysis pertaining solely to APP/PS1 mice showed th at for both T4 and T5 overall errors, only
202 the standard APP/PS1 mice were impaired compared to the standard NT mice ; performance of the high n-3 APP/PS1 mice was statistically not different from the same standard NT mice (Fig. 34). As was the cas e for all mice included (Figure 33), however, a high n-3 diet did not affect working me mory (T4 and T5) performance of APP+PS1 mice compared to standard APP+PS1 mice. These data involving RAWM errors indicate that a diet high in n-3 and/or n-6 fatty aci ds has very minimal to no effect on working memory in Alzheimers transgenic mice.
203 Fig 32. Working memory in the RAWM task in NT (a) and Tg+ (b) mice being evaluated for T1, T4 and T5 errors over thr ee 3-day blocks. T4 and T5 are indices of working memory. All three NT groups perfor med similarly in being able to reduce their number of T4 and T5 errors across blocks. By contrast, standard Tg+ were impaired versus standard NT mice for T4 and T5 during block 2 and T5 only during block 3. Within the Tg+ groups, the n-3 mice were impaired as compared to standard Tg+ mice for T5 within block 1. Otherwise there were no differences due to diet within the Tg+ mice. *Standard Tg+ mice significantly different from standard NT mice at P <0.05 or higher level of significance. High n-3 Tg+ significantly di fferent from standard Tg+ mice at P <0.05. a ) b )
204 Fig 33. RAWM errors in overall T4 (a) and T5 (b) for both NT and Tg+ groups. All NT groups had similar overall performance in bot h T4 and T5. Although the standard Tg+ mice had increased errors in overall T4 compar ed to standard-fed NT mice, there were no differences in T4 performance between the three Tg+ groups. In overall T5 errors, all three Tg+ groups displayed similar working memory impairment in making an increased number of errors versus standard NT mice. *Significantly different from standard NT at P <0.05.
205 Fig 34. RAWM errors in overall T4 and T5 for both NT groups and APP/PS1 groups. All NT groups had similar overall performa nce in both T4 and T5. The standard APP/PS1 mice (but not the high n-3 mice) ha d increased T4 and T5 overall errors compared to standard-fed NT mice. Howeve r, the high n-3 diet had no effect on T4 and T5 working memory when the two APP+PS1 groups were compared directly to one another. *Significantly diffe rent from standard NT at P <0.05.
206 Latencies to find the hidden platform fo r T1, T4 and T5 across the three 3-day blocks of the RAWM task are graphed for both NT (Fig. 35a) and Tg+ (Fig. 35b) mice. Across all nine days of RAWM testing, there were no differences between any of the NT groups with respect to latency to find the hidden platform. As was the case for RAWM errors, by the final two blocks of testing (B2 and B3), all three NT groups showed excellent working memory, as evidenced by thei r low T4 and T5 latencies. By contrast, all three Tg+ groups had increased latencies for both working memory trials (T4 and T5) in blocks 1 and 2 compared to standard NT mice. This impairment in T4 and T5 performance continued into the final block for standard Tg+ mice vs. standard NT mice. Compared to standard Tg+ mice, high n-3 a nd high n-6 Tg+ mice had slightly lower T4 and T5 latencies in Block 3 (Fig. 35b), whic h resulted in both of these groups not being different from standard NT c ontrols on those trials. However, in the more important direct comparison between the three Tg+ groups no significant differe nces were noted in T4 and T5 performance during block 3 (or any other block). Underscoring the lack of a robust effect of high n-3 or high n-6 diets in Tg+ mice are results of overall T4 and T5 performance (Fig. 36). All three NT groups performed similar in showing excellent overall T4 and T5 performance (e.g., low late ncies), while all three Tg+ groups showed similar overall T4 and T5 impairment compared to standard NT controls. Despite the somewhat improved performance of the high n-3 and high n-6 Tg+ groups by the last block, neither of these groups performed si gnificantly better than standard-fed Tg+ controls for any RAWM latency measure and both groups were impaired by having higher overall latencies in T4 (Fig. 36a) and T5 (Fig. 36b) overall.
207 To determine any effects a high n-3 di et might have specifically on APP+PS1 mice, RAWM performance of only APP/PS1 mi ce was compared to NT controls in the last block of testing (Fig. 37). As with all other indexes of RAWM performance, a high n-3 diet did not improve the already excellent working memory performance of NT mice during this final block. However, standa rd Tg+ mice were unable to improve their performance between T1 and T4/5 while high n-3 mice were able to do so. Because of this better performance by high n-3 Tg+ mice, they were no different in T4 or T5 latencies compared to standard NT mice, wh ile standard Tg+ mice were impaired on both T4 and T5 (Fig. 37). By T5 of the last block, there was a strong trend for the high n-3 APP/PS1 mice to have improved performance compared to the impairment of the standard APP/PS1 group ( P =0.06). As with RAWM errors the RAWM latency analysis suggests that dietary supplementation with n-3 or n-6 fatty acids resu lts in slight or no improvement in working memory performance.
208 Fig 35. Working memory in the RAWM task in NT (a) and Tg+ (b) mice being evaluated as latency to find the platform fo r T1, T4 and T5 over three 3-day blocks. There were no differences in latency between any of the NT groups across all days of RAWM testing or at individual blocks. In block 1 and block 2, a ll three Tg+ groups had increased latencies in both T4 and T5 compared to standard-fed NT mice ( = all Tg+ groups significantly different from standard NT at P <0.05 or higher level of significance). However, by block 3 both high n-3 and high n-6 Tg+ mice performed somewhat better, such that their performance was not different from standard NT mice, whereas the standard Tg+ group remained im paired. Standard Tg+ group significantly different from sta ndard NT group at P <0.05.
209 Fig 36. RAWM latency to find the hidden platform in overall T4 (a) and T5 (b) for both NT and Tg+ groups. All NT groups had simila r overall performance in both T4 and T5, however, all three Tg+ groups had increased late ncies in both T4 and T5 compared to the standard-fed NT mice. *Significantly different from standard NT at P <0.02 or higher level of significance.
210 Fig 37. RAWM latency to find the hidden platform within the last bl ock of testing for both NT and APP/PS1 on either the standard or high n-3 fatty acid diet s. Both NT groups performed similar to each other and exhibited excellent learning by a decrease in latency from T1 to T4/T5. Likewise, the high n-3 APP/PS1 showed this learning effect, whereas the standard APP/PS1 could not improve thei r performance between T1 and T4/5. Also, the standard APP/PS1 had increased latencies ve rsus standard NT mice for T4 and T5 (* P <0.05 or greater level of significan ce vs. standard NT controls)
211 Cytokine Levels At 7-9 months of age (5 -7 months into dietary treatments), pro-inflammatory and anti-inflammatory cytokines from plasma were measured in both NT (Fig. 38a) and APP/PS1 (Fig. 38b) mice from both the high n3 and high n-6 diet groups. For NT mice, there were no differences in plasma levels of any cytokine for hi gh n-3 vs. high n-6 diet groups (Fig. 38a). Among APP/PS1 mice, ther e was a consistent trend for high n-6 mice to have reduced levels of both proand an ti-inflammatory cytokines compared to high n3 mice (Fig. 38b). However, only levels of IL-1 and IFNwere significantly reduced by a high n-6 diet vs. a high n-3 diet. Because there were no significant differences between the two genotypes for any cytokine ir respective of diet, cytokine levels from both genotypes were combined. As shown in Figure 39, the high n-6 diet mice (NT & APP/PS1) had consistently lower levels of a ll cytokines measured compared to all mice on the high n-3 diet. However, the only m easure that was significantly reduced by the high n-6 diet was IL-1 Surprisingly, the anti-inflammato ry cytokines, IL-4 and IL-10 were also slightly, but not significantly, lo wer in the high n-6 gr oup. For both NT and APP/PS1 mice, these results indicate that the plasma cytokine profile is similar following a lengthy period of a high n-3 vs. a high n-6 diet. Hippocampal A levels As measured by ELISA, there were no differe nces in levels of either soluble or insoluble A within hippocampal tissu e between any of the three APP/PS1 dietary groups (Fig. 40). Thus, neither a high n-3 or a hi gh n-6 diet, administer ed over 5 months, affected brain A levels. There was however, a st rong trend for the high n-6 APP/PS1
212 group to have a lower level of insoluble A 42 compared to both standard and high n-3 diet groups.
213 Non-transgenic APP/PS1 Fig. 38. Standardized mean signal intensity of plasma cytokines in both NT (a) and APP/PS1 (b) mice on either the high n-3 or high n-6 fatty acid diets. No differences were seen in any of the cytokine levels between the two NT groups for either proor antiinflammatory cytokines. However, with in the APP/PS1 groups, the high n-6 mice expressed lower levels of IL-1 and IFNcompared to high n-3 APP/PS1 mice. Also, the high n-6 group had redu ced expression of TNFcompared to both NT groups. Significantly different compared to high n-3 APP/PS1 mice at P <0.05. a) b )
214 Fig. 39. Standardized mean si gnal intensity of plasma cytoki nes for all mice on either the high n-3 or high n-6 fatty acid diets irrespective of genotype Although the high n-6 diet group had consistently lower leve ls across all cytokines, IL-1 in high n-6 diet mice was the only cytokine that was significantly re duced vs. high n-3 mice. *Significantly different from high n-3 group at P <0.05.
215 Fig. 40. Hippocampal A levels (pmol/g), as measur ed by ELISA, of insoluble and soluble A 1-40 (a, b) as well as insoluble and soluble A 1-42 (c, d). No differences were found between any of the dietary groups for any A marker.
216 Fatty Acid Brain Tissue Levels Saturated and Mono-unsaturated Fatty Acids. Fatty acid levels including saturated, mono-unsaturated and polyunsaturated fatty acids were measured from frontal cortex and expressed as mean percentage of the total fatty acids. There were no differences in any of the saturated or monounsaturated fatty acids between any of the three NT groups (data not shown). Figure 41 illu strates that there were no differences in either total saturated (Fig. 41a) or total m ono-unsaturated (Fig. 41b) fatty acids between the NT groups. Similarly, there were no differe nces in any of the saturated fatty acids within the three APP/PS1 groups or in their total saturated fatty acid level (Fig. 41a). There was also no difference in total mono-uns aturated fatty acids between any of the APP/PS1 groups (Fig. 41b).
217 Fig. 41. Cortical levels of tota l saturated (a) and mono-unsatur ated (b) fatty acid levels of standard, high n-3 and hi gh n-6-fed NT and APP/PS1 mi ce. No differences were found between any of the groups for saturated or mono-unsaturated fa tty acids. Values expressed as mean percentage of total fatty acids SEM. a) b)
218 Omega-6 and Omega-3 Fatty Acids. As seen in Figure 42a for NT mice, the high n-6 group had elevated co rtical levels of 18:2n-6 and 20:2n-6 as compared to both other groups. Likewise, the high n-3 mice had reduced leve ls of 20:4n-6 and 22:4n-6 as compared to both standard and high n-6 mice. Surprisingly, however, the high n-3 mice showed an elevation of one n-6 fatty acid, 20: 3n-6. Despite this elevation, the high n-3 mice had reduced total n-6 fatty acids compar ed to both standard and high n-6 NT mice (Fig. 43a). Therefore, in the important dir ect comparison of the hi gh n-3 diet versus the high n-6 diet, the high n-6-fed NT mice had in creased levels of four of seven omega-6 fatty acids. Of the four different n-3s th at were measured, 20: 5n-3 and 22:5n-3 were both significantly elevated in the high n-3 NT mice compared to both other groups (Fig. 42a). Likewise, total n-3 fatty acids were higher in the n-3 fed NT group as compared to the standard-fed and high n-6 fed mice (Fig. 43b). Thus, in a direct comparison between the two experimental diets, the high n-3 diet resulted in increased omega-3 fatty acids within the frontal cortex in the NT mice. Within the dietary groups of APP/PS1 mi ce, the results were less consistent. Figure 42b shows that only 22:5n-6 was signi ficantly increased in the n-6 fed mice compared to the n-3 fed mice. However, re ductions in 3 of the seven n-6 fatty acids (18:2n-6, 20:2n-6 and 20:3n-6) were found in the standard-f ed mice compared to both experimental diet groups. As Figure 45a shows, there were no differences in total omega-6 fatty acids between any of the diet groups within the APP/PS1 mice. There were also no differences in any of the n-3 fa tty acids measured in the APP/PS1 mice (Fig. 42b). Thus, the total n-3 fatty acids were si milar between the three APP/PS1 diet groups, as seen in figure 43b.
219 Comparisons were also made within the individual diet groups to determine the effect of genotype on the fatty acid content in the frontal cortex (data not shown). For the standard-fed mice, there were no significant differences between any of the fatty acids measured for NT vs. APP/PS1 mice, except fo r elevations in both 22:5n-6 and 22:5n-3 in APP/PS1 mice. However, for the high n-3 diet mice, 5 of the seven n-6 fatty acids were elevated in APP/PS1 mice compared to NT mice (exceptions were 18:3n-6 and 20:3n-6), resulting in significantly higher total n-6 fatty acid levels in cortex of APP/PS1 mice vs. NT controls. In addition, 20:5n-3 and 22:5n3 were reduced in the APP/PS1 mice versus NT mice on the high n-3 diet. The reduction in these two n-3 fatty acids resulted in a significant reduction in total n-3 fatty acids in the APP/PS1 mice. Lastly, for mice fed the high n-6 diet, only 20:2n-6 was elevated in APP/PS1 mice compared to NT mice. Thus, for mice fed the high n-6 diet, there we re no genotypic differences in any of the other six n-6 fatty acids, or a ny genotypic differences for any of the n-3 fatty acids.
220 Fig. 42. Cortical levels of n-6 and n-3 fatty acids for standard, hi gh n-3 and high n-6 NT (a) and APP/PS1 (b) mice. Within the NT mice, the high n-6 diet group showed increases in four of seven n-6 fatty acid s compared to the high n-3 diet group. Conversely, high n-3 diet mice had elevations in two of four omega-3 fatty acids, 20:5n-3 and 22:5n-3, as compared to both other NT groups. Within the APP/PS1 mice, the high n-6 group had elevations in onl y one n-6 fatty acid (22:5n-6) versus the high n-3 mice. However, the standard-fed mice had lower leve ls of three n-6 fatty acids compared to both other groups. There were no differences in omega-3 fatty acids between any of the three APP/PS1 groups. Values expressed as m ean percentage of total fatty acids SEM. *Significantly different from both other diet groups within that genotype at P <0.05 or greater level of significance. # Significantly different from standard-fed group within that genotype at P <0.05 or higher level of significance.
221 Fig. 43. Total n-6 (a) and total n-3 (b) fatty acid levels in standard, high n-3 and high n-6 NT and APP/PS1 mice. High n-3 NT mice had significantly lower total n-6 as compared to both other NT groups. Similarly, high n-3 NT mice had significantly elevated n-3 as compared to both standard and high n-6-fed mice. There were no differences between the APP/PS1 groups for either total n-6 or n-3 fa tty acid levels in frontal cortex. Values expressed as mean percentage of total fatty acids SEM. Signifi cantly different from both standard-fed and high n-6-fed NT groups at P <0.05 or higher level of significance. a) b)
222 Correlations and Multimetric Analyses Correlations were performed involving bot h NT and APP/PS1 mice for behavioral measures versus fatty acid levels (satur ated, mono-unsaturated and poly-unsaturated) from the frontal cortex, plasma cytokines, and hippocampal A levels. First, behavioral measures of NT mice, combining all three diet groups, were correlate d with cortical fatty acid levels. Then, NT mice from only the high n-3 and high n-6 groups were used to correlate behavior with both fatty acid le vels and plasma cytokines. Because all pathology factors were only measured in APP/ PS1 mice, a final two sets of correlations excluded all APP mice and ther efore also excluded behavior al measures from the high n6 mice. Therefore, correlations were perf ormed between standard and high n-3 APP/PS1 mice behavioral measures and both co rtical fatty acids and hippocampal A levels. Lastly, A levels, cytokines and fatty acid levels were correlated within the high n-3 and high n-6 APP/PS1mice. Correlations involving all NT groups. Significant correlati ons involving all NT mice combined from the three diet groups reve aled associations between n-6 or n-3 fatty acids in the frontal cortex with cognitive impairment (Table 10). More specifically, 20:2n-6 (di-homo linoleic acid) was nega tively correlated with Morris water maze retention, such that an increased amount of this fatty acid in the brai n was associated with less time spent in the former platform-contai ning quadrant during th e probe trial. In addition, levels of this same fatty acid as well as 22:5n-6 (docosapentaenoic acid) were correlated with increased latency to find th e hidden platform in RAWM. Surprisingly, 22:6n-3 (DHA) and total N-3 PUFAs were positively correlated to impairment in both Morris water maze acquisition and RAWM.
223 Correlations involving high n3 and high n-6 NT groups. Additional correlations, including only the high n-3 a nd high n-6 fatty acid diets were also performed with the NT mice involving behavior al performance, brain fatty acid levels, and plasma cytokines (Table 10). Many significant correlations arose between impairment in platform recognition and all se ven n-6 fatty acids, as well as 2 of 4 n-3 fatty acids, indicating a clear association be tween high brain lipid levels and impaired search/recognition and identif ication learning. Despite no differences between the n-3 and n-6 NT dietary groups in saturated or mono-unsaturated fatty acid levels, there were significant and consistent associations between high levels of eight out of ten saturated fatty acids and impairment in the platform recognition task for both NT groups combined. Also, six out of ten mono-unsaturated fatty acid levels were signi ficantly correlated to impairment in the same task for both n-3 a nd n-6 NT groups combin ed. In addition, all n-6 fatty acids measured and two of four n-3 fatty acids were significantly correlated with impairment in the platform recognition task for both n-3 and n-6 NT groups combined. There were essentially no correlations betw een plasma cytokine levels and cognitive performance for the combined high n-3 and high n-6 NT diet groups.
224 Table 10. Correlations for all NT mice combined between cortical fatty acid levels and behavioral measures or for only high n-3 a nd high n-6 NT mice between fatty acid levels, plasma cytokines and behavioral performance. All abbreviations defined below in Table 12.
225 Correlations involving standard and high n-3 APP/PS1 groups. Significant correlations were found between combined standard-fed and high n-3-fed APP/PS1 groups versus behavioral meas ures, brain fatty acid, and A levels (Table 11). Most notably, elevated hippocampal soluble A 40 was significantly associated with Morris water maze acquisition and retention impairment in addition to increased latencies for platform recognition and RAWM. As well, insoluble A 40 also significantly correlated with impairments in Morris water maze and platform recognition. Soluble and insoluble A 42 were not as well correlated with cogni tive measures, although high levels of A 42 were associated with poorer RAWM performance. In addition to A levels, increased levels of seve ral cortical n-6 fatty acids also correlated with cognitive deficits in Morris water maze acquisition and retention, as well as cognitive impairment in RAWM. Surprisingly, cortical levels of n-3 fatty acids were essentially not associated with improvement in any of the cognitive-based measures. Correlations involving high n-3 and high n-6 APP/PS1 groups. Significant correlations were found between hippocampal A levels, plasma cytokines and cortical fatty acid levels in high n-3 and high n6 APP/PS1 mice (Table 11). There were significant positive correlations between insoluble A 42 and plasma cytokines, IFNand IL-10, indicating both pro-inflammatory and anti-inflammatory associations with A However, 20:2n-6, 20:4n-6 and 22:5n-6 all nega tively correlated with pro-inflammatory cytokines such as IFN, TNF, IL-2 or IL-6. In additi on, the n-3 fatty acid, 18:3n-3 (ALA) also negatively correlated with many of the pro-inflammatory cytokines as well as with the anti-inflammatory cy tokine, IL-10. Together, thes e correlations indicate that
226 increased cortical levels of n-6 and n-3 fatty acids are asso ciated with an overall antiinflammatory effect.
227 Table 11. Correlations for combined standard and high n-3 APP/PS1 mice between behavioral measures, cortical fa tty acid levels and hippocampal A levels or for combined high n-3 and high n-6 APP/PS1 mice between hippocampal A levels, plasma cytokines and brain fatt y acid levels. InsolA 42 = insoluble A 42 within hippocampus; all other abbreviati ons defined below in Table12.
228 Table 12 shows the factor an alysis of behavioral measures including all animals in all 6 treatment/genotype groups on a standard, high n-3 or high n-6 fatty acid diet. The table also includes a factor analysis for only the NT and APP/PS1 mice on either the standard or high n-3 diet, with and wit hout pathologic measures. Including all 6 experimental groups, 16 of the 19 behavioral me asures analyzed loaded on four principle factors, which together account for more than 65% of the total varian ce. All measures of RAWM errors and platform recognition, plus two measures of Morris water maze, loaded heavily in Factor 1, which was clearly c ognitively-based and accounted for 33% of the total variance. In contrast, Factor 2 was strongly sensorimotor/anxiety based, including both elevated plus maze and balance beam m easures. Factor 3 included both circular platform measures as well as Y-maze activit y, while Factor 4 contained a single measure the number of closed arm choices in elevated plus maze testing. Utilizing only two diet groups (standard and high n-3) and including only NT and APP/PS1 animals, many of the above 19 behavioral measures loaded similarl y in Factors 1 and 2 (Table 10). Factor 1 was again cognitivelybased in loading al l measures of RAWM errors, platform recognition, and all three measures of Morri s water maze. This factor, therefore, encompassed working memory, reference learning/ memory and recognition/ identification. Factor 2 was anxiety-based in loading only measures of elevated plus maze, while Factor 3 loaded two measur es of activity/exploratory behavior. Including neurochemical meas ures of cortical fatty acids (total saturated, total mono-unsaturated, total n-3, tota l n-6, and individual n-3 and n6 fatty acids), as well as hippocampal soluble and insoluble A levels to the behavioral measures for the NT and APP/PS1 groups is also reported in Table 12. Loading on Factor 1 were 10 cognitive
229 measures, soluble A 40 and A 42, and four n-6 fatty acid measures (including total n-6 fatty acid levels). Factor 1 contributed over 31% to the total variance. Factor 2 contained only fatty acid measures two individual n-6 fatty acids and three individual n-3 fatty acids. Factor 3 contained the same measures of elevated plus maze as Factor 2 without the pathology measures. In addition, string ag ility and two measures of n-3 fatty acids also loaded on Factor 2. Las tly, total mono-unsaturated fatty acids loaded separately on Factor 4. In summary, factor loadings w ith and without pathologic measures were similar for Factor 1, though cl early there was an association between these measures and both A pathology and cortical fatty acid levels.
230 Table 12. Factor loadings of behavioral measures, with and without pathologic measures.
231 Discriminant function analysis (DFA) wa s used in order to determine if the overall behavioral performance of each of the three diet groups (standard, high n-3 and high n-6) could be distinguished from each other for both NT and Tg+ groups. Table 11 shows a summary of two DFA me thods: the direct entry me thod (which included all 19 behavioral measures) and the step-wise fo rward method (which selected behavioral measures from all the measures based on thei r contribution to the va riance). The direct entry DFA could not distinguish between the three groups for either NT or Tg+ mice based on their overall behavior al performance. However, the step-wise forward DFA, iteratively selecting from the 19 behavioral measures, could distinguish between the NT high n-6 group and both other NT groups. Th is discrimination was based on a single behavioral measure circular platform errors on the final day of testing, wherein high n-6 NT mice were markedly impaired vs. the other two groups. Within the Tg+ groups, the step-wise forward DFA was successful in distinguishing between all thr ee diet groups so that the percent of animals correctly classified was 100% according to the classification matrix. Nine of the 19 behavioral measures were retained as providing the highe st level of discriminability within the Tg+ groups: 8 cognitive measures and 1 anxiety/activity measure. A Canonical plot of these stepwise-forward DFA results involving Tg+ mice is presented in Figure 44. Similar discrimination was seen when including onl y APP/PS1 mice and comparing standard and high n-3 dietary groups (Table 12). A step-wise forward DFA using all behavioral measures could successfully discriminate be tween both standard and high n-3 APP/PS1 groups. Similar to the DFA that utilized all three Tg+ groups, the percent of animals correctly classified by the st ep-wise forward DFA including only APP/PS1 animals was
232 100%. Although both step-wise forward DFAs did provide discrimi nation between Tg+ groups behaviorally, no group was clearly superi or across all of the behavioral measures iteratively selected therein. Moreover, dir ect entry DFAs were completely unsuccessful in distinguishing overall behavi oral performance of the three Tg+ groups. Thus, there was no overall benefit or impairment provided by either the n-3 or n-6 diets in Tg+ mice. Four additional DFAs were performed, which included only the eight cognitivebased measures that loaded into Factor 1 from Table 10. For the NT mice, neither the direct entry nor stepwise-for ward DFAs could distinguish between any of the three dietary groups (Fig. 45a). Li kewise, the direct entry DFA method could not discriminate between any of the Tg+ groups (Fig. 45b). Ho wever, a stepwise-forward DFA, selecting from the 8 cognitive-based measures from F actor 1 revealed discrimination between Tg+ high n-6 mice and both Tg+ standard and Tg+ hi gh n-3 mice. The measures retained as providing maximal discrimination were platform recognition on the last day of testing (a search recognition/identification-based measure) and T5 latency during the last block of RAWM testing. When repeating the above DFAs for only APP/PS1 mice and only for standard diet vs. high n-3 diet, the two A PP/PS1 groups could be distinguished when a stepwise-forward DFA was utilized and only Factor 1 cognitive-based measures were included. The two measures retained as pr oviding this discrimination were RAWM last block T5 and RAWM overall T5 latency. For both of these cognitive measures, high n-3 mice had better performance. The percent of animals correctly classified by the step-wise forward DFA for Factor 1 for all Tg+ anim als and only APP/PS1 animals was 72% and 90%, respectively. Nonetheless, for all th ree other DFAs that were done involving APP/PS1 mice alone, there was no discriminatio n between high n-3 and standard diet
233 groups, again underscoring that a high n-3 diet did not provide substantially better overall performance across a variety of tasks.
234 Fig. 44. Canonical scores plot of stepwise-f orward DFA using all 19 behavioral measures for the three Tg+ groups. Each symb ol represents the cognitive performance of one animal as graphed from the two linear f unctions derived in the DFA. For the Tg+ animals, all groups could be distinguished from one another in cognitive performance.
235 Fig. 45. Canonical scores plots of di rect entry DFAs for the three groups of NT mice (a) and the three groups of Tg+ mice (b). The eight cognitive-bas ed measures that loaded on Factor 1 of Factor Analysis were u tilized. The DFA could not successfully discriminate between any of the three dietary groups within the NT animals. Similarly, none of the three Tg+ groups could be discriminated from the others based on their overall cognitive performance across the 8 cognitive measures utilized. a ) b )
236 Table 13. Summary of discriminant function analyses including NT and Tg+ mice.
237 Table 14. Summary of discriminant function an alyses including NT and APP/PS1 mice for the standard and high n-3 groups.
238 Discussion In the present work, we evaluated the beha vioral and pathologi cal effects of active A immunotherapy and dietary administration of omega-3 and omega-6 fatty acids in APP/PS1 and NT mice. Previous work has shown the significance of using transgenic animals to test various therapeutics for AD, such as environmental enrichment, dietary supplementations or vaccination protocols. Within the present work, there was a greater cognitive benefit from A vaccination as compared to omega-3 dietary intervention. However, neither study resulted in any significant alteration of A levels, indicating that any cognitive benefits observed through either A immunotherapy or dietary fatty acid administration were A independent. I. A Vaccination Discussion Immunotherapeutic approaches to AD first began in the 1996 with in vitro work using monoclonal antibodies against A (Solomon et al., 1996). It wasnt until 1999, however, that the first study was performed on an AD transgenic mouse model (Schenk et al., 1999). Within the last 6 years, many studies have focused on the use of either active or passive vaccinations in transgen ic mice; however, few have behaviorally assessed the efficacy of such immunotherapy. The current study tested the effects of lifelong active A vaccination on behavioral and A deposition measures in AD transgenic
239 mice. Following initiation of immunotherapy at 2 months of age, these mice were behaviorally evaluated at an early time point (4-6 months ) and a late time point (1516months) in an extensive battery that an alyzed multiple sensorimotor and cognitive domains. Thus Tg+ were vaccinated over a 14 month period (essentially most of their adult life) to determine if life-long vaccinations protect ag ainst cognitive impairment in both adulthood and old age. This study re presents the longest period of time any AD vaccination protocol has utilized in any species. Briefly, th is protracted protocol of immunotherapy protected AD transgenic mice from impairment in a variety of cognitive measures that were otherwise present at both test points. This improvement in cognition, however, occurred without any change in compact A deposition within the brains of these mice. Nevertheless, factor analysis revealed an underlyi ng relationship between compact A and multiple cognitive measures, indi cating an interaction between brain A and cognitive performance. Behavioral Effects Sensorim otor and Anxiety-based Tasks Within this study, at the early time point, the Tg+/A group exhibited increased activity over multiple measures. The increased activity could be due to higher levels of soluble A or an interaction of the A /antibody complex within the hippocampus, which has been shown to be influential on lo comotor activity (Laghmouch et al., 1997). Alternatively, the increased activity coul d be due to glutamate activation by A 1-42, resulting in over-activation of NMDA receptors (Brunskill et al., 2005). Over activation of NMDA receptors in the hippocampus has prev iously been shown to result in increased motor activity (Rogers et al., 1989).
240 Subjects with AD have been shown to e xhibit reduced anxiety levels (Gauthier, 1998; Stout et al., 2003), as did the Tg+/C mice at the late test point. However, because the Tg+/A mice had corrected their level of anxiet y to the level of the NT mice, the A vaccination corrected their disi nhibition. Comparable to fi ndings by Arendash et al. (2001), the Tg+/C group at either test point had similar overa ll activity to the NT mice. In addition, both balance beam and string agili ty tasks did not result in any difference between the Tg+/C mice and the NT mice. Th is indicates that any cognitive impairment in the Tg+/C mice was not due to sensorimotor impairments. Behavioral Effects Cognitive-based Tasks Similar to the balance beam and string agility tasks, both Y-maze and circular platform tasks did not reveal any difference in performance between the Tg+/Con and NT groups at either test point. Therefore, th ere was no potential fo r the immunotherapy to correct the performance within these tasks. These findi ngs are similar to those by Arendash et al. (2001a) and King and Ar endash (2002) that found no change in performance between NT and APP Tg+ mice at similar ages for the Y-maze task. Importantly, measures from both Y-maze and circ ular platform tasks loaded independent from other cognitive-based measures in the factor analysis, as our laboratory has previously reported (Leighty et al., 2004). This implies that these two tasks measure separate cognitive domains as compared to the water-based memory tasks. Most active vaccination studies that repor ted behavioral findings only used single tasks to identify any cognitive improvement. Janus et al. (2000) reported that 17 weeks of A 1-42 injections to 15 week old TgCRN D8 mice resulted in improved spatial
241 reference learning, compared to control mice. Similarly, the present study also found a significant improvement at a young age in AD transgenic mice due to repeated vaccinations (see below). However at the aged test point, the results were less impressive, possibly due to the loss of Tg+ mi ce between the two test points, resulting in lower group size and consequently less power fo r statistical analysis It is noteworthy that the survival rate of APP/PS1 mice in the present study was comparable to the survival rate of APP mice reported by King a nd Arendash (2002). Nevertheless, at both test points, Tg+/Con mice were impaired vers us NT mice in spatial reference learning during Morris maze testing. In contrast, Ar endash et al. (2001a) found that young adult APP/PS1 mice were not impaired in spatial re ference learning in Morris water maze. A possible explanation for the discrepancy could be the slight alteration in A processing/aggregation after generations of crossbreeding, thereby making the APP transgene more sensitive to the cogni tive domain required for Morris water maze acquisition. At both young a nd aged timepoints, A immunotherapy was effective at protecting against reference learning impairment particularly at the earlier test point, as evidenced by Morris water maze acquisition results. Similar to Morris water maze acquisition, A immunotherapy protected against the memory retention deficit otherwise present at the young time point, as shown in th e Morris water maze probe trial. However, by the late time point, the immunotherapy was unable to protect against the continued memory impairment exhibited by these same mice. The lack of effect of the vaccination could be partially due to the reduced number of animals in both Tg+ control and treatment groups. Alternatively, it may be due to the high plaque load these animals have accumulated by 16 months of age, which was not prevented by A immunotherapy and
242 that adversely affected their memory rete ntion. In contrast, Kotilinek et al. (2002) reported an improvement in both Morris wate r maze acquisition and retention memory in adult Tg2576 mice following 1 week of passive immunizatio n. However, the results indicated a significant treatment effect only for pre-treatment ve rsus post-treatment testing of the Tg+ mice and not for a direct comparison of post-treatment performance of Tg+ controls vs. Tg+/A treated mice. The platform recognition task, perfor med immediately following Morris water maze, measured the animals ability to switc h strategies. During Morris water maze, the mice were required to use a spatial strategy to find a submerged platform; however, the platform recognition task required the animals to use a different strategy identification/recogn ition learning. At the early time point, no effect of the A immunotherapy could be seen because all groups performed well. At the aged test point, however, the performance of Tg+/C mice was im paired vs. NT controls. However, the performance of A -treated Tg+ mice was no different fr om the NT mice or from any of the groups at the early test point, indicating that A immunotherapy protected these mice from impairment in switching strategies. This immunization-induced protection is especially relevant to patients with AD, w ho commonly have deficits in attention-shifting tasks (Amieva et al., 2004; Ta les et al., 2004; Dorion et al., 2002). Deficits in attentionshifting tasks have been shown to be regiona lized in the parietal cortex (Buck et al., 1997). Adult Tg+ mice show A plaques in the cortex (Dodart et al., 2000b), which could contribute to the impairment in strategyswitching of the Tg+/C mice. Therefore, it is evident that A immunotherapy administered early in life and continued as a prophylactic can protect AD Tg+ mice from impa irment in strategy switching.
243 Previously, the radial arm water maze (RAWM) task has been shown to be sensitive to working memory impairment and A deposition (Arendash et al., 2001a; Gordon et al., 2001; Nilsson et al., 2004; Leight y et al., 2004). By th e last block of the task, and for both time points, the pres ent studys Tg+/C mice exhibited poor performance. This is in conjunction with short-term memory impairments, which progress throughout the disease process in patients with AD. In contrast, A immunotherapy improved the performance of Tg+ at both ages such that the Tg+/A mice were no different from NT mice by the la st block of testing. Morgan et al. (2000) had similar results after 8 months of A 1-42 injections, such that A -treated mice performed significantly better than control Tg+ mice and no different from NT mice. Likewise, Sigurdsson et al (2003) injected 6-8 m onth old Tg2576 mice with A 1-30 for 13 months and found an overall signifi cant improvement in working memory performance in an 8-arm radial maze. However, the use of A immunotherapy in aged APP/PS1 mice did not reverse any of the wo rking memory impairments exhibited in RAWM testing, as reported by Austin et al. (2003). In that later study, the lack of a treatment effect could be due to the much shorter period of vaccination and/or the much greater A burdens that the aged APP/PS1 had when A immunotherapy was begun. Results from the present study indicate that long-term A immunotherapy is effective at protecting against working me mory impairments in AD Tg+ mice if implemented at a young or adult age. Summarizing the behavioral data in this study, A immunotherapy administered for 15 months effectively prot ected AD Tg+ mice against cognitive deficits
244 in spatial learning/reference memory (Morri s water maze), strategy switching/recognition (platform recognition), and working memory (RAWM). A Histopathology Surprisingly, all of the cognitive improvements previously reported occurred without any change in compact A deposition. Therefore, all of the cognitive benefits from life-long A immunotherapy were independent of alterations of A deposition. Several prior vaccination studies have also shown cognitive benefits of immunotherapy in behavioral tasks withou t concurrent changes in A measures. Morgan et al. (2000) only found a reduction in A in the cortex, not hippo campus, despite significant improvements in working memory. In addition, passive immunotherapy has been reported to result in improvements in cognition with no alterations in amyloid deposition. Regarding the latter, Dodart et al. (2002) found that admini stration of m266 (specific to A 13-28) improved object recognition memory after a single injection, with no concurrent change in A deposition. The authors suggested th at the behavioral benefit from the vaccination is more likely due to peripheral clearance or sequestration of soluble A and not by altering deposited forms of A within the brain. In e ffect, the pres ent study and others suggest that behavioral improvements resulting from A immunotherapy are possibly due to neutralization of soluble A oligomers from the brain. This effect occurs early in A aggregation, prior to manifest A fibrillar formation, such that improvements in cognition could occur at an early age, prior to overt A deposition, as was seen the present study for the early time point.
245 Correlations and Multimetric Analyses Previous work has demonstrated sign ificant correlations between cognitive measures and A deposition; however, the present study found no such correlations. Most notably this could be due to the lack of change in A deposition between the transgenic groups, despite the improved cognitive performance of Tg+/A mice. Also, because previous work from the same mouse line (Arendash et al., 2001; Gordon et al., 2001) found significant correlations be tween cognitive performance and A pathology measures, generations of crossbreeding for the present study could have affected the behavioral phenotype of the transgenic mice, as well as the fact that the present study involved a full test battery (r un twice). Despite the lack of correlations, FA analysis revealed significant underlying relationshi ps between multiple cognitive measures and A deposition. In prior work from our laborat ory, factor analysis has been successfully used in transgenic AD mouse models to identify underlying relationships between cognitive and pathological measures (Lei ghty et al., 2004). Studies that employ a comprehensive test battery that encompasses multiple cognitive and sensorimotor domains have the advantage of utilizing multimetric analyses, such as FA and DFA, in order to characterize and distinguish between different treatment groups based on their performances. Leighty et al. (2004) demonstrated signi ficant associations between cognitive performance and A deposition measures within four separate PDAPP mouse lines. At both the early and late test poi nt the current study found one primary factor (factor 1) that was largely cognitively based a nd one factor (factor 2) that loaded mainly sensorimotor measures. At the late test point, both with and without A pathologic measure, one measure of elevated plus maze (E P-TO) loaded separately in its own factor.
246 While the remaining two factors (EP-#C and EP -#O) loaded with activity or exploratory measures, the percent time in open arms cl early measures more of a true anxiety component. Also at the late test point, A pathologic measures loaded primarily within factor 1, which remained cognitively-based. However, the findings from factor 2 associated cognitive and sensorimotor components in addition to one measure of A ; whereas without pathologic measures, factor 2 did not include any cognitive measures. This indicates that A interacts with measures of activ ity in addition to measures of cognition. In addition to FA, discriminant functi on analysis (DFA) has also been used successfully in transgenic mouse models studies. This analysis attempts to discriminate between groups of mice based on their perfor mance. Within the present study, although the direct entry method was relatively in effective in discriminating between the 3 groups, the DFA step-wise forward method was very effective at di stinguishing between the NT and Tg+ groups. At the later test poin t, starting with all 19 behavioral measures, this DFA was able to discriminate between the Tg+ groups. This indicates that even after 15 months of active A immunizations, A vaccinated Tg+ were still clearly distinguishable from Tg+ controls based on their sensorimotor measures and cognitive performance. In addition, a stepwise forward DFA using only the top cognitive measures from factor 1 significantly separated th e NT and Tg+/Con groups, but could not distinguish cognitive performance of the NT and Tg+/A groups. The measures retained for providing the discrimination between NT and Tg+/Con mice included two measures of RAWM, and one from platfo rm recognition, indicating that the transgenic behavioral
247 effect at 15-16 months is largely based on working memory and recognition/ identification. In summary, life-long active A immunotherapy was eff ective at protecting AD Tg+ mice from impairment in a variety of c ognitive domains This protection provided to Tg+ mice extended across all of the waterbased cognitive tasks and well into older age. Since this cognitive protection occurred without a decrease in A deposition, the mostly likely mechanism of A immunotherapys action invol ves neutralization/removal of small A oligomers from the brain. Because of the effectiveness of several A immunotherapy studies in AD transgenic models of Alzheimers disease, cl inical trials were performed. A Zurich cohort of AD subjects that had r eceived active vaccinations of A 1-42 showed stabilization of memory a nd preserved hippocampal functi on in conjunction with the production of antibodies (Hock, 2003). Als o, antibodies from these inoculated AD subjects were reactive to A plaques from AD Tg+ mouse brains and A from human brains (Hock, 2002). However, 5 % of a se parate cohort of AD subjects that were injected with A 1-42 developed a severe inflammatory response (e.g., meningoencephalitis) to the vaccin e or to the vaccines adjuvant Therefore, these trials were halted and less risky approaches for AD immunotherapy are currently being explored. Alternative vaccina tion therapies which might offe r less negative side effects include passive vaccinations, immunization with partial A fragments, or injection of Tcells.
248 II. Omega-3 Discussion Numerous epidemiologic studies suggest that a higher intake of fish oils is associated with a reduced risk of dementia la ter in life (Kalmijn et al., 1997a; Kalmijn et al., 1997b, Kalmijn et al., 2004). This reduced ri sk is thought to be more specifically contributed by the n-3 fatty acids, DHA and/or EPA, which are highly concentrated in fish (Kalmijn et al., 1997a; Kalmijn et al., 1997b). Despite this notion, there have been no long-term clinical studies that have admini stered either DHA or EPA to subjects with dementia as a treatment, or prior to onset of cognitive impairment in the form of a prevention-based study. In addition, the majority of rodent studies that have utilized fish oils or DHA supplementation have largely fi rst involved DHA or n-3 depletion prior to supplementation (Gamoh et al., 2001; Ikemoto et al., 2001; Calon et al., 2004; Hashimoto et al., 2005). In non-transgenic rodent s, there was no cognitive benefit to supplementation with DHA after multiple gene rations of n-3 deficiency (Gamoh et al., 2001; Ikemoto et al., 2001). Likewise, Ca lon et al. (2004) found no benefit of recognition/identification memory or memory retention in either non-transgenic or APP transgenic mice supplemented with DHA. However, APP mice enriched with DHA showed an improvement in memory acquisiti on in the Morris water maze (Calon et al., 2004). In addition, Hashimoto et al. (2005) showed an improvement in working and reference memory in rats supplemented with DHA, after being fed a fish oil deficient diet for 3 generations and infused with A 1-40. Thus, the effects of n-3 fatty acid supplementation in rodents is equivocal, with the later two studies related to Alzheimers
249 disease not designed to address the central question of whether n-3 fatty acids are cognitively protective against AD. The major aims of the present study were to : 1) determine the cognitive effects of a high n-3 diet in both NT and AD transgenic mice, 2) determine the cognitive effects of a typical American diet (high in n-6 and low in n-3 fatty acids) in both NT and AD transgenic mice, and 3) determine the effects of these two diets on bl ood cytokine levels, brain fatty acid levels, and brain A levels (Tg+ mice only). The high n-3 diet mimicked humans having fish twice weekly, with a dietar y n-6/n-3 ratio of 4/1. In essence, this study investigated whether an n-3 enriched diet, begun in young adulthood, can protect against inevitable cognitive impairment in AD transgenic mice and whether a high n-6 enriched diet with an n-6/n-3 ratio of 50/1 can contribute to greater cognitive impairment in AD transgenic mice. Moreover, since bot h diets were also gi ven to non-transgenic mice, cognitive effects in normal animals could also be tested. Survival Analysis. Prior to an assessment of the behavioral impact of a high n-3 or high n-6 diet in Tg+ mice (see below), a study was performed to determine the effects of an n-3 deficient diet on survivability. Briefly, this stu dy found that an interaction between the background and transgene profile de termined the survivability of mice that were fed an n3 deficient diet. F1 generation PS1 mice had increased survival on an n-3 deficient diet as compared to F1 APP mice and F2 gene ration APP, PS1 and NT mice. The F1 generation had a relatively high amount of B6 background (25%), which primarily contributed to the increased survival of the PS1 mice. Mice with 100% B6 background
250 have been shown to be less sensitive to diet ary alterations in fat, and have minimal changes in metabolism when shifting between high fat to low fat diets (Biddinger et al., 2005). In addition, mice with a high B6 bac kground also have elevated expression of fatty acid synthase (FAS), malic enzyme and -5 desaturase, all of which convert available fatty acids into long-chain fatty ac ids (Coleman, 1992; Bi ddinger et al., 2005). FAS is critical in de novo lipogenesis to ultimately c ontribute to the production of palmitate. Palmitate is a precursor to downs tream production of long-chain fatty acids. Malic enzyme also contributes to the de novo production of palmitate. Thus, the B6 background appears to have contributed to increased lipid metabolism in order to compensate for the n-3 deficiency and reduced n-6 dietary content. A combination of the increased expression of FAS, malic enzyme and -5 desaturase due to the increased B6 background from the F1 versus the F2 genera tion clearly contributed to the increased survivability of PS1 mice on a diet devoid of n-3 fatty acids. As to why the PS1 transgene contributed to survival whereas the APPsw transgene did not, there is no clear associati on between either of these transgenes and lipid metabolism. Previous work has shown that the presence of a mutant APP transgene in mice leads to increased inflammation, oxidative damage and dystrophic neurites (Irizzary et al., 1997; Benzing et al., 1999; Melhorn et al., 20 00; Pratico et al., 2001). These factors alone are not detrimental to th e survival of APP mice at their young age. However, the combination of the severe reducti on in n-6 fatty acids, the lack of n-3 fatty acids in the omega-3 deficient diet, and th e presence of mutant APP could account for their low survival. In the present study, th e F2 generation mice had a reduced percentage of B6 background (12.5%), which was insufficien t to contribute to survival irrespective
251 of genotype. In summary, the higher B6 background and PS1 genotype of F1 generation mice was able to compensate for the a diet low in n-6 and deficient in n-3 fatty acids. Behavioral, Pathologic and Neurochemical Assessments. Briefly, this study determined that in NT mice, long-term n-3 supplementation did not provide any cogniti ve benefits. Similarly, in Tg+ mice, n-3 supplementation largely resulted in cognitive performance that was no be tter than that of mice fed a standard diet or a diet high in n-6 fatty acids. For bot h NT and Tg+ mice, the high n-6 diet did not result in an overall worse c ognitive performance as compared to the standard-fed mice. These results suggest that DHA and EPA s upplementation may only provide limited, or no cognitive protection against AD. It is po ssible, however, that di etary fish contains additional beneficial nutrients besides DHA and EPA, that could offer protection against development of AD in humans. Additionally, there were no major diet-induced changes in plasma cytokine levels or brain A deposition (the later for Tg+ mice). However NT mice did show some marked diet-induced change s in cortical fatty acids levels. Within the high n-6 fed NT mice, most of the n-6 fatty acids were elevated as compared to the high n-3 fed NT mice. Similarly, half of the n-3 fatty acids were elevated in the high n-3 fed NT mice as compared to the high n-6 fe d NT mice. However, in the APP/PS1 mice, there were minimal diet-induced ch anges in cortical fatty acid levels with either a high n3 or high n-6 diet. Additionally, there we re no diet-induced changes in soluble or insoluble hippocampal A or plasma cytokines between any of the three APP/PS1 groups. Despite the minimal changes in cogni tive and pathologic measures, multimetric analyses revealed significant separations be tween the dietary groups of the Tg+ and
252 more specifically between the standard and high n-3 APP/PS1 mice. However, within the NT mice, only the high n-6 fed mice were discriminated from both the standard and high n-3 NT mice. Behavioral Effects Sensorim otor and Anxiety-based Tasks Prior rodent studies that have evaluated sensorimotor and anxiety effects from fish oil supplementation found both an increase (Carrie et al., 2000) and decrease (Chalon et al., 1998) in locomotor activ ity; however no significant ch anges in anxiety-based tasks (Chalon et al., 1998; de Wild e et al., 2002). The presen t study did not reveal any differences in sensorimotor/anxiety measures among the three dietary groups of NT mice. Thus, no effects of high n-3 or high n-6 diets were evident on sensorimotor function and anxiety level in normal mice. These results are in contrast to several prior reports involving normal rodents. Ca rrie et al. (2000) showed increased locomotor activity in mice that were supplemented with fish oil. However, Chalon et al. (1998) found reduced activity in young rats that had been supplemented with fish oil for two generations. Chalon et al. (1998) also measured anxiety in the elevated plus maze task and found no effect with fish oil supplementation. In conj unction with this, de W ilde et al. (2002) also found no difference in elevated plus maze meas ures with n-3 PUFA enrichment in rats. The results from prior studies are therefor e inconsistent, perhaps reflecting different methodologies. In contrast to the lack of di etary effects on sensorim otor function/anxiety in the present studys NT mice, Tg+ mice given n-3 supplementation had increased activity in two separate tasks (open fiel d and Y-maze). The reason(s) for why an
253 increased level of activity results from n3 supplementation in Tg+ mice, but not NT mice, are unknown and would be based on conjecture without further study. No previous studies have evaluated th e effects of n-3 or n-6 supplementation on balance beam or string agility in either normal or Alzheimers transgenic mice. For both NT and Tg+ mice of the present study, there we re no effects of either the n-3 or the n-6 diet on either of these tasks. These results underscore that neithe r diet had deleterious effects on sensorimotor function and that a ny dietary effects on cognitive-based tasks within the NT or Tg+ groups could not be at tributable to sensorimotor impairment. Behavioral Effects Cognitive-based Tasks Previous work involving a lterations of dietary n-3 or n-6 fatty acids evaluated cognitive performance primarily in the Mo rris water maze task (Jensen et al., 1996; Wainwright et al., 1999; Carrie et al., 2000; de Wilde et al., 2002; Barcelo-Coblijn et al., 2003; Calon et al., 2004). A few studies (all from the same group) have reported enhanced performance with n-3 supplement ation in maze learning using a simple maze construction with only one entry/exit and many blind alleys (Suzuki et al., 1998; Lim & Suzuki, 1999; Lim & Suzuki, 2000). One study that utilized the Tg2576 APP transgenic mouse model found marginal improvement in spatial memory learning from Morris water maze acquisition with DHA enrichment (Calon et al., 2004). It is important to underscore, however, that the Calon et al (2004) study involved DHA supplementation to Tg2576 mice that were previously on a DHA-def icient diet. As mentioned earlier, this experimental model has marginal relevance to human-related studies. Within the present study there were minimal effects of either diet across a variety of cognitive measures in
254 NT or Tg+ mice. The few effects of high n-3 or high n-6 diets can primarily be summarized as follows: n-6 enrichment in Tg + and NT mice resulted in impaired spatial learning/memory in the circular platform task as compared to standard-fed groups within their genotypes, n-6 supplementation in Tg+ mice resulted in improved memory retention as seen in the Morris water m aze, and n-3 supplementation in APP/PS1 mice resulted in improvement in recognition/identif ication in the platform recognition task, as compared to the standard-fed APP/PS1 mice. Because there was no difference in Ymaze percent altern ation between the standard Tg+ mice and the standard NT mice, no improvement due to fatty acid administration could be seen in the Tg+ mice. Also, in the NT mice, supplementation of n-3 fatty acids did not result in any improveme nt in alternation pe rformance; as well, supplementation of n-6 fatty acids did not resu lt in a lower percent al ternation in either Tg+ or NT mice. Even though no other previous work has evaluated the effects of n-3 or n-6 fatty acid supplementation on spatial learni ng/memory in the circular platform task, the present study showed that dietary enrichment of either NT or Tg+ mice with n-6 fatty acids resulted in cognitive impairment in this task. Although the circular platform task measures spatial reference learning and memo ry similar to Morris water maze, circular platform measures loaded on factors independent of other cognitive measures in factor analysis, indicating a separate cognitive domain used in this task as compared to Morris maze and water-based cognitive tasks in general. Clearly, circular platform uses unique cognitive domains that were sensitive to n-6 supplementation in both Tg+ and NT mice. Likewise to the above-mentioned tasks, the standard-fed Tg+ mice performed similar to the standard-fed NT mice in Morr is water maze. Despite this lack of a
255 transgenic impairment, there was ample opportunity for the n-3 supplementation to improve upon Morris maze performance in Tg+ mice, and yet no such improvement occurred with n-3 supplementation. The lack of a transgenic impairment in Morris acquisition is in opposition to the impairment seen in 5 month old APP mice (Arendash et al., 2004) and 4-6 month old APP/PS1 mice pr eviously reported in this dissertation (Jensen et al., 2005). The lack of an improvement or impairment in Morris maze acquisition with either n-3 or n-6 dietary intervention in NT mice is consistent with all other NT rodent studies that have used Morr is water maze to evaluate cognitive effects of fatty acid supplementation. Jensen et al (1996) found no improvement in spatial learning/memory after 4 generations of fish oil supplementation in either acquisition or retention memory. Additional work invol ving supplementation of young rats with DHA or fish oil did not result in improved spatia l memory (Wainwright et al., 1999; BarceloCoblijn et al., 2003). Also, de Wilde et al (2002) showed that adult rats did not behaviorally benefit from n-3 supplementation in Morris water maze acquisition or retention. Thus, the present studys lack of any n-3 dietary effect on Morris maze acquisition in normal rodents is consis tent with the prior literature. Within Tg2576 APP mice Calon et al ( 2004) reported that DHA enrichment to an n-3 deficient diet resulted in improved acqui sition in Morris water maze. However, the improvement in acquisition was only noted in the latter phase of Morris maze acquisition. Thus, there was no effect overall of transgenic ity, DHA enrichment or low dietary DHA. The design of the Calon study done in AD Tg + mice and the present studys design are not comparable. Nonetheless, both studies found no effects of a hi gh n-3 diet on Morris maze retention this is, if all 6 groups of th e present study are included in the analysis.
256 If only APP/PS1 mice are compared for a high n-3 vs. standard diet, however, the present study found that n-3 supplemented Tg+ mice showed a partial improvement in memory retention. Surprisingly, the high n-6 Tg+ mice showed a complete restoration in memory retention to the same level as the NT mice. Within the NT mice, opposite results were seen, such that the high n-6 mice showed impairment in memory retention. Thus, the same diet involving high n-6 fatty acids had opposite effects in NT and Tg+ mice. Several previous studies have shown that NT mice supplemented with n-3 fatty acids show no benefit in Morris water maze acquisition or retention (Jensen et al., 1996; Wainwright et al., 1999; Carri e et al., 2000), while NT mice fed low dietary n-3 fatty acids are not impaired in Morris water maze (C alon et al., 2004). Since no prior Morris maze studies involved administratio n of a high n-6 diet to AD Tg + mice, there is a lack of information in the available l iterature that could account fo r the discrepancy between the response of NT and Tg+ to a high n-6 diet. The platform recognition task requires th at mice switch strategies from the spatial memory of Morris water maze to a search/id entification strategy. Calon et al. (2004) found no effect of dietary DHA in NT mice in th is task, similar to the results from the present study. However, Calon et al. ( 2004) also found no diff erence overall between APP and NT mice, nor with th e addition of DHA to either genotype. However, the present study found a clear impairment of Tg + mice fed a high n-6 diet as compared to mice fed a high n-3 fatty acids. Although, n3 fatty acids have been shown to be important in visual development (Marin et al., 2000), this is most lik ely not the reason for the impairment of high n-6 Tg+ mice. All Tg+ and NT groups performed similar in MWM acquisition in which they were required to use spatial cues, illustrating that all
257 groups had similar visual abilities. Thus, there was a clear cognitive protection offered by n-3 fatty acids to AD Tg+ mice compared to the impairment otherwise evident in Tg+ mice fed a diet high in n-6 fatty acids, indepe ndent of visual abiliti es. These results are quite significant in relation to subjects with AD that have difficulty in strategy switching or attention shifting, as previously discussed in a prior section. The radial arm water maze (RAWM) task ha s been shown to be a sensitive index of working memory and also sensitive to hippocampal A deposition (Arendash et al., 2001; Gordon et al., 2001; Nilsson et al., 2004; Le ighty et al., 2004). To date, there have been no other studies that eval uated the effects of either omega-3 or omega-6 fatty acid supplementation to NT or AD Tg+ mice on radial arm water maze (RAWM) working memory performance. However, Sugimoto et al., (2002) found that adult mice that were supplemented with DHA had improved working memory, as indexed by the 8-arm radial maze task. In contrast, Gamoh et al. (2001) found no improvement in working memory in the same task from aged rats that were given dietary DHA after 3 generations on a fish oil deficient diet. Similarly, the present study illustrated th at there were no, or minimal effects of diet in either NT or Tg+ mi ce on RAWM working memory performance. Although, both high n-3 and high n-6 diet Tg+ mice performed similar to standard NT mice by the last block of testing, neither diet group provided any significant improvement or impairment in memory as compared to the standard Tg+ mice. This indicates that n-6 enrichment did not result in a greater deficit as compared to either a standard or high n-3 diet in either NT or Tg+ mice. Thus, high di etary n-6 fatty acids in normal individuals, or those predisposed to AD would not appear to induce an increased risk of developing a working memory deficit. In addition, the current study showed that n-3 supplementation
258 was not able correct working memory impair ments in Tg+ mice and provided essentially no benefit to working memory in Tg+ mice. A Pathology Puskas et al. (2003) and Barcelo-Coblijn et al. (2003) found that fish oil supplementation to rats increased expression of transthyretin (TTR). TTR has been previously shown to solubulize fibril A and inhibit aggregation of A in vitro (Schwarzman & Goldgaber, 1996; Redondo et al ., 2000). TTR has the potential to both inhibit initial formation of A deposition and clear compact am yloid from the brains of rodents fed fish oil. However, the present study found no alterations of either soluble or insoluble hippocampal A 1-40 or A 1-42 levels from APP/PS1 mi ce fed a high n-3 diet. Also, there were no differences in hippocampal soluble or insoluble A 1-40 or A 1-42 in high n-6 APP/PS1 mice compared to either the high n-3 or standard-fed APP/PS1 mice. In contrast, Lim et al. (2005) found th at DHA enrichment to aged AD Tg+ mice previously fed an n-3 deficient diet reduced total insoluble A in addition to total (insoluble + soluble) A 42 and total A 40 in the cortex. Despite the altered A levels, TTR expression was unchanged in those mice, indicating an alternative mechanism by which DHA altered amyloid levels. Cytokine Levels Previous works shows conf licting results indicating bot h pro-inflammatory and anti-inflammatory expression resulting from n-3 supplementation. Billiar et al. (1988) and Renier et al. (1993) show ed that macrophages from rodents fed dietary fish oil
259 resulted in an anti-inflamma tory effect. By contrast, Watanabe et al. (1993) and Petursdottir et al. (2002) show ed that rodents fed a diet enriched with fish oil had increased pro-inflammatory cytokines from peritoneal macrophages. The present study found no significant differences in plasma cyt okine levels of NT mice fed a high n-3 vs. high n-6 diet. These results are in contrast to all four of th e above studies, perhaps in part due to their use of macrophages to measure cy tokines rather than the current studys use of plasma; alternatively their tr eatment groups were compared to standard diet animals, in contrast to the n-3 vs. n-6 group comparisons For the most part, human studies have supported anti-inflammatory actions of n-3 fatty acids. Blok et al. (1996) reviewed many of the human studies involving n-3 fatty acid with inflammatory diseases. To summarize his review, administration of dietary n-3 fatty acids re sulted in improvements in rheumatoid arthritis, Systemic Lupus Er ythematosus (SLE), psoriasis and ulcerative colitis. The primary mechanism to reduce inflammation is through EPA metabolism. EPA, from fish oil, rapidly incorporates into cell membranes and replaces AA in the phospholipids. This replacement increases th e production of a less active prostaglandin (E3) and leukotriene (B5), thereby reducing the inflammatory response. In the present study, plasma cytokine le vels from Tg+ mice indicated a trend toward reduction in all cytokines (proand anti-inflammatory) in plasma for Tg+ mice fed a high n-6 diet vs. a high n-3 diet. Therefore, the pr esent study supports both proand anti-inflammatory enhancement in Alzheimers Tg+ mice fed a high n-6 diet.
260 Brain Fatty Acid Levels AD subjects have been found to have a reduced amount of EPA, DHA and total n3 fatty acids in brain tissue as compared to elderly control subjects (Corrigan et al., 1998; Conquer et al., 2000). This reduction in n-3 PUFAs as compared to control elderly persons further provides evidence that the pa thology associated with AD is independent of the normal aging process. Although the cause of the alterations in fatty acid content in the brain of AD patients is unknown, some studies suggest that there is an alteration in fatty acid metabolism by the liver, increased oxidative damage, or a reduction in dietary intake of specific fatty acids (Soderberg, et al., 1991; Skinner et al ., 1993; Corrigan et al., 1998; Prasad et al., 1998). Fatty acid metabo lism proceeds primarily such that dietary fatty acids are metabolized by the liver and transported via the blood to target organs, such as the brain. The present study found th at alterations in br ain fatty acids were affected by both the AD genotype and dietary in take. However, all three diets (standard, high n-3 and high n-6) contained similar amoun ts of saturated and monounsaturated fatty acids. As such, there were no significan t differences in cortical saturated or monounsaturated fats within the three NT or three APP/PS1 groups. There were also no differences between the NT or APP/PS1 mice on any of the three di ets, indicating that there is no difference in metabolism of satu rated or monounsaturated fats in APP/PS1 mice as compared to NT mice. For the NT mice fed a high n-6 diet, ma ny of the cortical n-6 fatty acids (including total n-6 fatty acids) were elevated versus the high n-3 fe d mice. Likewise for the NT mice fed a high n-3 diet, half of the n3 fatty acids were elevated versus the high n-6 NT mice. Thus, in the NT mice, the frontal cortex content of fatty acids is reflective
261 of their dietary intake. However, despite the significant change s in both n-3 and n-6 cortical levels, minimal changes in behavior al performance were seen between the NT groups. These results provide further evid ence for a lack of relationship between DHA and/or EPA intake and cognitive function. W ithin APP/PS1 mice, only one n-6 fatty acid (22:5n-6) was elevated in th e high n-6 group versus the high n-3 group. Also, there were no differences in cortical n-3 content be tween any of the three APP/PS1 groups. Therefore, unlike the NT mice, despite the si gnificant difference in dietary intake of n-3 and n-6 fatty acids, the concurrent changes in brain fatty acid levels were not evident. In contrast to the current study, Calon et al. (2005) found signi ficant decreases in frontal cortex n-6 fatty acids and increases in n-3 fatty acids in aged Tg2576 APP mice fed a low n-3 diet supplemented with DHA. Th e same results were also seen in aged NT mice on a low n-3 diet that was also suppl emented with DHA. The authors suggested that the difference in frontal cortex fatty acid levels of the low n-3 fed APP mice was most likely due to the initial n-3 depletion in combination wi th the presence of the APP transgene. Calon et al. ( 2005) suggested that for the lo w n-3 group, the APP transgene provided oxidative stress that rapidly depleted the brain of DHA. However in a direct comparison between the genotypes, the pr esent study found that among the high n-6 mice, there were minimal differences in brai n fatty acid levels when comparing NT to APP/PS1 mice. However, the current study al so found significant incr eases in several n6 fatty acids within the APP/PS1 high n-3 mice versus the high n-3 NT mice. Thus, despite identical dietary intake, APP/PS1 mice show a deficit in metabolizing polyunsaturated fatty acids for incorporation in to brain tissue as compared to NT mice. Therefore enhanced dietary intake of n-3 or n-6 fatty acid in normal mice resulted in
262 elevated brain levels of n-3 and n-6 fatty acids, respectively; however, these same elevations were not seen in APP/PS1 mice. Correlations Two sets of correlations were performed involving NT mice 1) for all three diet groups or 2) for only the high n-3 and high n-6 groups. Briefly, these correlations revealed a positive relationship between sa turated, monounsaturated, n-3 and n-6 fatty acids in the frontal cortex and cognitive im pairment in multiple tasks, primarily RAWM and platform recognition. These correlations show that brain total fatty acid levels, irrespective of their fo rm, are indicative of cognitive impairment. However, Greenwood and Winocur (1996) showed that learning impair ment is directly related to intake of saturated fats, and independent of monounsaturated or polyunsaturated fatty acid intakes. More recently, Winocur and Greenwood (2005) showed that rats fed a diet high in saturated fats were impaired on memory a nd learning tasks. Li kewise, Morris et al. (2005) found that increased intake of n-6 and n3 fatty acids in addition to saturated fats plays an important role in cognitive function a nd risk of dementia. Similar to the present study, Suzuki et al., (1998) found that mice that had increased n-6 fatty acids in the brain stem were poorer performers as compared to mice that had increased n-3 content in the brain stem. However, the authors did not perform any direct correlations. Similarly, Ikemoto et al. (2001) also showed that an increa se in brain n-6 fatty ac id levels is related to learning impairment in a br ightness-discrimination task. Similar to NT mice, correlations am ong APP/PS1 mice revealed a significant relationship between cortical n-6 levels a nd cognitive impairment in multiple tasks.
263 Previous work has also shown that elevated le vels of brain n-6 fatty acids are associated with impaired learning or memory in AD Tg+ mice (Calon et al., 2004) or AD-like rats (Hashimoto et al., 2002). Studies of human AD subjects also show a relation between a reduction in brain n-6 fatty acid s and reduced prevalence of dementia in AD (Corrigan et al., 1998; Prasad et al., 1998). Prasad et al. (1998) not only showed a correlation between AD and reduction of brain n-6 fatty acids, but also an association of -amyloid plaques in the regions that had reduced n-6 fatty acid le vels. Likewise, the present study showed a significant correlation between hippocampal A levels and cognitive impairment. This is in agreement with previous work show ing multiple correlations between brain A deposition and a variety of cognitive measur es (Arendash et al., 2001a; Gordon et al., 2001; Leighty et al., 2004). La stly, both cortical n-3 and n-6 fatty acid levels in frontal cortex were overall correlated with plasma cytokines higher brain levels of n-3 and n-6 fatty acids were associated with lower levels of pro-inflammatory cy tokines (e.g., an antiinflammatory effect). As mentioned in one of the below sections, the relation between polyunsaturated fatty acids an d cytokine levels is conf licting, as supported by the correlation results. FA and DFA In agreement with the A vaccination study of this dissertation and previous studies from our laboratory, the factor anal ysis results involving behavioral measures from the present study show that factor 1 is loaded exclusively by cognitive measures, whereas factor 2 is sensorimotor/anxiety-based (Leighty et al., 2004; Jensen et al., 2005). Whether or not APPsw mice were included with the APP/PS1 mice in the factor analysis,
264 the factor loadings remained relatively sim ilar. Inclusion of neurochemical measures showed associations between mu ltiple cognitive domains and both A measures and n-6 fatty acids. In support, similar associations were also seen within several correlations as previously discussed. In addition to FA, a number of discri minant function analyses (DFAs) were performed in order to determine if the separa te dietary groups within each genotype (Tg+ or NT) could be significantly distinguishe d based on their sensorimotor and cognitive performances. Direct entry DFA was comple tely unsuccessful at discriminating between any of the dietary groups of NT or Tg+ (with or without APP mice) mice. However, the step-wise forward DFA could discriminate between the high n-6 NT group and both the standard and high n-3 NT groups using all measures (sensorimotor and cognitive). In addition, step-wise forward DFA was ineffectiv e at discriminating be tween the three NT groups using only the top seven cognitive measur es from the FA factor 1, or the two NT groups (standard and high n-3) using either all measures or the top nine cognitive measures from FA factor 1. This indicates that the fatty acid diets partially contributed to the overall performance of the NT mice, de spite the lack of significant difference between these groups on any given task. However, there were several significant differences in cortical fatt y acid levels within the sepa rate NT groups, which, according to the correlation results were significantly a ssociated with multiple cognitive domains. Clearly, the cortical fatty acid content in NT mice has an underlying relationship to cognitive performance. The step-wise forward DFAs for Tg+ and APP/PS1 mice were more effective at discriminating between dietary groups than the DFAs for the NT mice. Step-wise
265 forward DFA was successfully used to comple tely separate all three Tg+ groups and both standard and high n-3 APP/PS1 groups using all 19 behavioral measures. This indicates an overall contribution of dietary fatty acids to performance on all tasks collectively (sensorimotor, anxiety, and cognitive) in Tg+ mice despite a lack of effect on individual tasks between the 3 Tg+ dietary treatment gr oups. Also, Tg+ mice fed the high n-6 diet were able to be separated from both the st andard and high n-3 Tg+ mice with the stepwise forward DFA using only the top seven cognitive measures from the FA factor 1. This is the same separation seen from the step-wise forward DFA using all measures in NT mice. Lastly, step-wise forward DFA completely discriminated standard-fed APP/PS1 mice from high n-3-fed APP/PS1 mi ce using their top nine cognitive measures from the FA factor 1. Thus, despite a lack of difference in many of the behavioral tasks, n-3 fatty acid supplementation contributed significantly to the overall behavioral performance of APP/PS1 mice. Taken togeth er with the correla tion results and FA measures, it is evident that dietary intake of fatty acids are closely associated with multiple cognitive domains in both NT and Tg + mice. Although none of the direct entry DFAs were successful, all of the step-wise fo rward DFAs were completely successful at separating the high n-3 and hi gh n-6 diet groups for Tg+ mice (with or without APP mice). Overall, this indicates that the increase in n-3 fatty acids in the high n-3 diet was sufficient to alter the behavior of AD Tg+ mi ce such that they coul d be separated based solely on their behavioral performances.
266 Clinical Implications of Omega-3 Fa tty Acids for Prevention Against AD Within NT mice, there was no overall cognitive benefit of dietary n-3 supplementation. Conversely, there were no overall impairments due to the n-6 enrichment. Thus, the high n-6 content of a typical American diet does not by itself lead to cognitive impairment in individuals not pr e-disposed to AD. However, since a high n6 fatty acid diet is typically li nked with a diet high in saturated fatty acids, it is more likely that the saturated fats in a typical Am erican diet increase the risk of cognitive impairment. Supportive of this premise, Morr is et al. (2005) found little evidence for an association between fish intake and rate of cognitive declin e. Instead, Morris found that the positive relationship between fish in take and cognitive protection was most significant when including ALA supplemen tation over a long-term period, thus the protection was not solely due to the fish intake In addition, rodents that have been fed a high fat diet are clearly impaired in spatial learning in the Morris water maze(Zhao et al., 2004) and learning/memory in th e variable-interval delayed attention task (Winocur & Greenwood, 1996; Greenwood & Winocur, 1999), underscoring the increased risk of cognitive impairment induced by a high fat diet. Although in the AD transgenic mice, there were limited cognitive benefits of high n-3 fatty acid supplementation compared to a standard diet, a high n3 diet did not result in significantly better cognitive performance in any task compared to a high n-6 diet. Indeed, Tg+ mice on a high n-6 diet were not worse overall on any cognitive task compared to those on a high n-3 diet. Theref ore, even in subjects predisposed to AD, a diet high in n-6 fatty acids doe s not necessarily lead to in creased cognitive impairment. Rather, it would appear that th e high saturated fat content often associated with a high n-6
267 diet is the determining factor in increased ri sk of AD for a typical American diet. Along this line, Morris et al. (2005) proposed that overall fat consumption had a higher link to cognitive decline and Alzheimers disease than a diet simply low in n-3 fatty acids. The present work suggests that DHA, EPA, and/or fish oil supplements may provide only limited, or no cognitive protection agai nst AD, as well as no reduction in A levels in the brain. Also, the protective effects of a high fish diet, as shown in previous epidemiological studies, may involve other be neficial nutrients in fish besides a high content of n-3 fatty acids nutrients that could confirm some protection against ADrelated dementia. Alternatively, the postulate d protection of a high fish diet could be related to the otherwis e healthful lifestyle adopted by ma ny fish eater (e.g., a lower fat diet, high intake of fruits and vegetables, increased exercise). IV. Overall Conclusions Life-long active vaccination and dietary omega-3 fatty acids to AD Tg+ mice resulted in two different levels of cognitive protection, with A vaccinations clearly being more beneficial. There was a more extensive and consistent long-term cognitive protection offered by the A immunotherapy across multiple cognitive domains, whereas long-term omega-3 supplementation results in very limited cognitive benefit in selected cognitive measures. Both long-term treatm ents resulted in no change in brain A levels, indicating that any cogn itive benefits were i ndependent of either A deposition (A immunotherapy study) or A generation (dietary study) in th e brain This suggests that the cognitive protection from the A vaccine may be due to neutralization of soluble
268 oligomers or a completely A -independent mechanism. Nonetheless, the multimetric analyses for both studies showed a clear association/relations hip between brain A levels and cognitive factors. Because there were no di fferences in plasma cytokine levels in the n-3 dietary study, any cognitive protection offered by omega-3 supplementation was independent of a global inflammatory re sponse, although the possi bility of a more localized brain inflammatory response cannot be eliminated because brain levels of cytokines were not evaluated. Nonetheless, there were numer ous alterations of cortical fatty acid levels that significantly correlated wi th several behavioral tasks. Thus, there is a clear link between brain fatty acid leve ls and cognitive function, which was also underscored by the factor analysis. Las tly, both studies approached AD therapeutics from a prevention-based arena in utilizing Alzheimers Tg+ mice to determine the longterm effects of active A immunotherapy and dietary fatt y acid manipulation in highly controlled prospective studies where all other va riables were controlled for. Since such longitudinal protective-based studies in hum ans are impractical and cannot be tightly controlled, testing various ther apeutic and/or prophylactic in terventions in AD transgenic models represents a critically important venue for developi ng effective interventions against Alzheimers Disease.
269 References 1. Agnew, S. 1996. Genetic subtypes of Alzh eimers disease and cognitive functioning. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 278-288. Oxford: Oxford University Press. 2. Aid, S., Vancassel, S., Poums-Ballihaut, C., Chalon, S., Guesnet, P. and Lavialle, M. 2003. Effect of a diet-induced n-3 PUFA depletion on cholinergic parameters in the rat hippocampus. Journal of Lipid Research 44: 1545-1551. 3. Aisen, P., Davis, K., Berg, J., Schafer, K., Campbell, K., Thomas, R., Weiner, M., Farlow, M., Sano, M., Grundman, M. and Thal, L. 2000. A randomized controlled trial of prednisone in Alzh eimers disease. Alzheimers Disease Cooperation Study. Neurology 54: 588-593. 4. Aisen, P. 2002a. The potential of anti-i nflammatory drugs for the treatment of Alzheimers disease. Lancet Neurology 1: 279-84. 5. Aisen, P., Schmeidler, J. and Pasinet ti, G. 2002b. Randomized pilot study of nimesulide treatment in Alzheimers disease. Neurology 58: 1050-1054. 6. Aisen, P., Schafer, K., Grundman, M., Pfeiffe r, E., Sano, M., Davis, K., Farlow, M., Jin, S., Thomas, R. and Thal, L. 2003. Effects of rofecoxib or naproxen vs placebo on Alzheimer disease progression: a randomized controlled trial. JAMA 289: 2819-2826. 7. Akbar, M. and Kim, H. 2002. Protectiv e effects of docosah exaenoid acid in staurosporine-induced apopt osis: involvement of phospha tidylinositol-3 kinase pathway. J Neurochem 82: 655-665. 8. Alessandri, J., Poumes, Ballihsut, C., La ngelier, B., Perruchot, M., Raguenez, G., Lavialle, M and Guesnet, P. 2003. Incor poration of docosahexaenoic acid into nerve membrane phospholipids: bridging the gap between animals and cultured cells. Am J Clin Nutr 78: 702-710. 9. Almkvist, O., Axelman, K., Basun, H., Je nsen, M., Viitanen, M., Wahlund, L. and Lannfelt, L. 2003. Clinical findings in nondemented mutation carriers predisposed to Alzheimers disease: a model of mild cognitive impairment. Acta Neurol Scand Suppl 179: 77-82.
270 10. Akbar, M. and Kim, H. 2002. Protectiv e effects of docosahexaenoic acid in staurosporine-induced apopt osis: involvement of phospha tidylinositol-3 kinase pathway. J Neurochem 82: 655-665. 11. Amieva, H., Lafont, S., Rouch-Leroyer, I., Rainville, C., Dartigues, J., Orgogozo, J. and Fabrigoule, C. 2004. Evidencing inhibi tory deficits in Alzheimers disease through interference effects and shifting disabilities in the Stroop test. Arch Clin Neuropsychol. 19: 791-803. 12. Andreasen, N., Vanmechelen, E., Vanderstich ele, H., Davidsson, P. and Blennow, K. 2003. Cerebrospinal fluid levels of totaltau, phosphor-tau and A beta 42 predicts development of Alzheimers disease in patients with mild cognitive impairment Acta Neurol Scand Suppl 179: 47-51. 13. Arai, H., Terajima, M., Miura, M., Higuchi S., Muramatsu, T., Machida, N., Seiki, H., Takase, S., Clark, C., Lee, V. et al. 1995. Tau in cerebrospinal fluid: a potential diagnostic marker in Alzheimers disease. Ann Neurol 38: 649-652. 14. Arendash, G., Gordon, M., Diamond, D., Austi n, L., Hatcher, J., Jantzen, P., DiCarlo, G., Wilcock, D. and Morgan, D. 2001a. Behavioral assessment of Alzheimers transgenic mice following long-term A vaccination: task specificity and correlations between A deposition and spatial memory. DNA and Cell Biology 20: 737-744. 15. Arendash, G., King, D., Gordon, M., Morgan, D., Hatcher, J., Hope, C. and Diamond, D. 2001b. Progressive, age-related behavi oral impairments in transgenic mice carrying both mutant amyloid precursor pr otein and presenilin-1 transgenes. Brain Research 891: 42-53. 16. Arendash, G., Garcia, M., Costa, D., Cracch iolo, J., Wefes, I. and Potter, H. 2004. Environmental enrichment improves cogni tion in aged Alzheimers transgenic mice despite stable beta-amyloid deposition. Neuroreport 15: 1751-1754. 17. Arispe, N., Rojas, E. and Pollard, H. 1993. Alzheimer disease amyloid protein forms calcium channels in bilayer membranes: blockade by tromethamine and aluminum. PNAS 90: 567-571. 18. Ashe, K. 2001. Learning and Memory in Transgenic Mice Modeling Alzheimers Disease. Learning & Memory 8: 301-308. 19. Auer, S., Sclan, S., Yaffee, R. and Reisberg, B. 1994. The neglected half of Alzheimer disease: cognitive and functional concomitants of severe dementia. J Am Geriatr Soc 42: 1266-1272.
271 20. Austin, L., Arendash, G., Gordon, M., Diam ond, D., DiCarlo, D., Dickey, C., Ugen, K. and Morgan, D. 2003. Short-term beta -amyloid vaccinations do not improve cognitive performance in cogniti vely impaired APP+PS1 mice. Behav Neurosci 117: 478-484. 21. Bagga, D., Wang, L., Farias-Eisner, R., Glaspy, J. and Reddy, S. 2003. Differential effects of prostaglandin derived from -6 and -3 polyunsaturated fatty acids on COX-2 expression and IL-6 secretion. PNAS 100: 1751-1756. 22. Bales, K., Verina, T., Cummins, D., Du, Y ., Dodel, R., Saura, J., Fishman, C., DeLong, C., Piccardo, P., Petegnief, V ., Ghetti, B. and Paul, S. 1999. Apolipoprotein E is essential for am yloid deposition in the APP (V171F) transgenic mouse model of Alzheimers disease. PNAS 96: 15233-15238. 23. Barcel-Coblijn, G., H gyes, E., Kitajka, K., Pusks, L. Zvara, A., Hackler, L., Nyakas, C., Penke, Z. and Farkas, T. 2003. Modifications by docosahexaenoic acid of age-induced alterations in gene expression and molecular composition of rat brain phospholipids. PNAS 100: 11321-11326. 24. Bard, F., Cannon, C., Bargour, R., Burke, R., Games, D., Grajeda, H., Guido, T., Hu, K., Huang, J., Johnson-Wood, K., Kh an, K., Kholodenko, D., Lee, M., Lieberburg, I., Motter, R., Nguyen, M., So riano, F., Vasquez, N., Weiss, K., Welch, B., Seubert, P., Schenk, D. and Yednock, T. 2000. Peripherally administered antibodies against amyloid -peptide enter the central nervous system and reduce pathology in a mouse model of Alzheimer disease. Nature 6: 916-919. 25. Baxter, M., Frick, K., Price, D., Breckler S., Markowska, A. and Gorman, L. 1999. Presynaptic markers of cholinergic function in the rat brain: relationship with age and cognitive status. Neuroscience 89: 771-779. 26. Beard, C., Kokmen, E., Sigler, C., Smith, G., Petterson, T. and OBrien, P. 1996. Causes of death in Alzheimers disease. AEP 6: 195-200. 27. Becker, J., Lopez, O. and Butters, M. 1996. Episodic memory: differential patterns of breakdown. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 71-88. Oxford: Oxford University Press. 28. Behl, C. and Moosmann, B. 2002. Serial Review: Causes and consequences of Oxidative Stress in Alzheimers Disease. Free Radical Biology & Medicine 33: 182-191. 29. Benzing, W., Wujek, J., Ward, E., Shaffe r, D., Ashe, K., Younkin, S. and Brunden, K. 1999. Evidence for glial-mediated inflammation in aged APPsw transgenic mice. Neurobiology of Aging 20: 581-589.
272 30. Biddinger, S., Almind, K., Miyazaki, M., Kokkotou, E., Ntambi, J. and Kahn, C. 2005. Effects of diet on genetic backgr ound on sterol regulatory element-binding protein-1c, stearoyl-CoA de saturase 1, and the development of the metabolic syndrome. Diabetes 54: 1314-1323. 31. Billiar, T., Bankey, P., Svingen, B., Curran, R., West, M., Holman, R., Simmons, R. and Cerra, F. 1998. Fatty acid intake and Kupffer cell function: fish oil alters eicosanoids and monokine production to endotoxin stimulation. Surgery 104: 343-349. 32. Birks, J., Melzer, D. and Beppu, H. 2002a. Donepezil for mild and moderate Alzheimers disease (Cochrane Review). In: The Cochrane Library 33. Birks, J., Grimley, E., Iakovidou et al. 2002b. Rivastigmine for Alzheimers disease (Cochrane Review) In: The Cochrane Library 34. Blennow, K. and Hampel, H. 2003. CSF mark ers for incipient Azlheimers disease. The Lancet: Neurology 2: 605-613. 35. Blennow, K. and Vanmechelen, E. 2003. CSF markers for pathogenic processes in Alzheimers diseae: diagnostic implications for use in clinical neurochemistry. Brain Research Bulletin 61: 235-242. 36. Blok, W., Katan, M. and van der Meer, J. 1996. Modulation of inflammation and cytokine production by dietary (n-3) fatty acids. J Nutr 126: 1515-1533. 37. Boller, F., Verny, M., Hugonot-Diener, L. a nd Saxton, J. 2002. Clinical features and assessment of severe dementia. A review. European Journal of Neurology 9: 125-136. 38. Bonin, A. and Khan, N. 2000. Regulation of calcium signaling by docosahexaenoic acid in human T-cells. Implication of CRAC channels. J Lipid Res 41: 277-284. 39. Borchelt, D., Ratovitski, T., van Lare, J., Lee, M., Gonzales, V., Jenkins, N., Copeland, N., Price, D. and Sisodia, S. 1997. Accelerated amyloid deposition in the brains of transgenic mice coexpressi ng mutant presenilin 1 amyloid precursor proteins. Neuron 19: 939-945. 40. Boston, P., Bennett, A., Horrobin, D. and Bennett, C. 2004. Ethyl-EPA in Alzheimers disease-a pilot study. Prostaglandins, Leukotrienes and Essential Fatty Acids 71: 341-346. 41. Brunskill, E., Ehrman, L., Williams, M., Kl anke, J., Hammer, D., Schaefer, T., Sah, R., Dorn, G., Potter, S. and Vorhees C. 2005. Abnormal neurodevelopment,
273 neurosignaling and behavior in Npas3-deficient mice. Eur J Neurosci 22: 12651276. 42. Buck, B., Black, S., Behrmann, M., Caldwell, C. and Bronskill, M. 1997. Spatialand object-based attentional deficit in Alzheimers disease. Relationship to HMPAO-SPECT measures of parietal perfusion. Brain 120: 1229-1244. 43. Buerger, K., Teipel, S., Zinkowski, R., Bl ennow, K., Arai, H., Engel, R., HofmannKeifer, K., McCulloch, C., Ptok, U., He un, R., Andreasen, N., DeBernardis, J.,Kerkman, D., Moeller, H., Davies, P. and Hampel, H. 2002a. CSF tau protein phosphorylation at threonine 231 correlat es with cognitive decline in MCI subjects. Neurology 59: 627-629. 44. Buerger, K., Zinkowski, R., Teipel, S ., Tapiola, R., Arai, H., Blennow, K., Andreasen, N., Hofmann-Keifer, K., DeBe rnardis, J., Kerkman, D., McCulloch, C., Kohnken, R., Padberg, F., Pirttila, T., Schapiro, M., Rapoport, S., Moeller, H., Davies, P. and Hampel, H. 2002b. Differe ntial diagnosis of Alzheimer disease with cerebrospinal fluid levels of ta u protein phosphorylati on at threonine 231. Arch Neurol 59: 1267-1272. 45. Burns, M., Gaynor, K., Olm, V., Merken, M ., LaFrancois, J., Wang, L., Mathews, P., Noble, W., Matsuoka, Y. and Duff, K. 2003. Presenilin redistribution associated with aberrant cholesterol transport en hances beta-amyloid production in vivo. J Neurosci 23: 5645-5649. 46. Cai, L., Chin, F., Pike, V., Toyama, H., Li ow, J., Zoghbi, S., Modell, K., Briard, E., Shetty, H., Sinclair, K., Donohue, S ., Tipre, D., Kung, M., Dagostin, C., Widdowson, D., Green, M., Gao, W., Herma n, M., Ichise, M. and Innis, R. 2004. Synthesis and evaluation of two 18F-lab eled 6-iodo-2-(4`-N ,N-dimethylamino) phenylimidazo [1,2-a] pyridine derivatives as prospective radi oligands for betaamyloid in Alzheimers disease. J Med Chem 47: 2208-2218. 47. Calhoun, M., Wiederhold, K., Abramowski, D. Phinney, A., Probst, A., SturchlerPierrat, C., Staufenbiel, M., Sommer, B. and Jucker, M. 1998. Neuron loss in APP transgenic mice. Nature 395: 755-756. 48. Calon, F., Lim, G., Yang, F., Morihara, T., Teter, B., Ubeda, O., Rostaing, P., Triller, A., Salem, N., Ashe, K., Frautschy, S. and Cole, G. 2004. Docosahexaenoic acid protects from dendritic pa thology in an Alzheimers disease mouse model. Neuron 43: 633-645. 49. Calon, F., Lim, G., Morihara, T., Yang, F., Ubeda, O., Salem, N., Frautschy, Sl. And Cole, G. 2005. Dietary n-3 polyunsatur ated fatty acid depletion activates caspases and decreases NMDA receptors in the brain of a transgenic mouse model of Alzheimers disease. European Journal of Neuroscience 22: 617-626.
274 50. Carlson, G., Borchelt, D., Dake, A., Turner, S., Danielson, V., Coffin, J., Eckman, C., Meiners, J., Nislen, S., Younkin, S. a nd Hsiao, K. 1997. Genetic modification of the phenotypes produced by amyloid pr ecursor protein overexpression in transgenic mice. Hum Mol Genet 6: 1951-1959. 51. Carrie, I., Guesnet, P.., Bourre, J. and Francs, H. 2000. Diets containing long-chain n -3 polyunsaturated fatty acids affect be haviour differently during development than ageing in mice. British Journal of Nutrition 83: 439-447. 52. Casu, M., Wong, T., De Koninck, Y., Ribi ero-da-Silva, A. and Cuello, A. 2002. Aging causes a preferential loss of cho linergic innervation of characterized neocortical pyramidal neurons. Cereb Cortex 12: 329-337. 53. Chalon, S., Delion-Vancassel, S., Belz ung, C., Guilloteau, D., Leguisquet, A., Besnard, J. and Durand, G. 1998. Dietar y fish oil affects monoaminergic neurotransmission and be havior in rats. Nutritional Neurosciences 2512-2519. 54. Champeil-Potokar, G., Denis, I., Goustard-L angelier, B., Alessandri, J., Guesnet, P. and Lavialle, M. 2004. Astrocytes in culture require docos ahexaenoic acid to restore the n-3/n-6 polyunsaturated fa tty acid balance in their membrane phospholipids. Journal of Neuroscience Research 75: 96-106. 55. Chauhan, N. and Siegel, G. 2003. Intracere broventricular passive immunization with anti-A antibody in Tg2576. Journal of Neuroscience Research 74: 142-147. 56. Chen, G., Chen, K., Knox, J., Inglis, J., Bernard, A., Martin, S., Justice, A., McConlogue, L., Games, D., Freedman, S. and Morris, R. 2000. A learning deficit related to age and -amyloid plaques in a mouse model of Alzheimers disease. Nature 408: 975-979. 57. Cohen-Mansfield, J. 2001. Nonpharmacol ogic interventions for inappropriate behaviors in dementia: a revi ew, summary, and critique. Am J Geriatr Psychiatry 9: 361-381. 58. Coleman, D. 1992. The influence of ge netic background on th e expression of mutations at the diabetes (db) locus in the mouse. VI: Hepatic malic enzyme activity is associated with diabetes severity. Metabolism 41: 1134-1136. 59. Coleman, P. and Yao, P. 2003. Synaptic slaughter in Alzheimers disease. Neurobiology of Aging 24: 1023-1027. 60. Conquer, J., Tierney, M., Zecevic, J., Bettg er, W. and Fisher, R. 2000. Fatty acid analysis of blood plasma of patients with Alzheimers disease, other types of dementia and cognitive impairment. Lipids 35: 1305-1312.
275 61. Cook, S., Miyahara, S., Bacanu, S., Perez-Madrinan, G., Lopez, O., Kaufer, D., Nimgaonkar, V., Wisniewski, S., DeKosky, S. and Sweet, R. 2003. Psychotic symptoms in Alzheimer disease: evidence for subtypes. American Journal of Geriatric Psychiatry 11: 406-413. 62. Corder, E., Saunders, A., Strittmatter, W., Schmechel, S., Gaskell, P., Small, G., Roses, A., Haines, J. and Pericak-Vance, M. 1993. Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimers disease in late onset families. Science 261: 921-923 63. Corrigan, F., Horrobin, D., Skinner, E., Besson, J. and Cooper, M. 1998. Abnormal content of n-6 and n-3 long-chain unsatu rated fatty acids in the phosphoglycerides and cholesterol esters of parahippocampal cortex from Alzheimers disease patients and its relationship to acetyl CoA content. The International Journal of Biochemistry & Cell Biology 30: 197-207. 64. Coslett, H. and Saffran, E. 1996. Visuospatial functioning. Chap. In The Cognitive Neuropsychology of Alzh eimer-type Dementia. 193-205. Oxford: Oxford University Press. 65. Cotman, C., Tenner, A. and Cummings, B. 1996. -Amyloid Converts and Acute Phase Injury Response to Chronic Injury Responses Neurobiology of Aging 17, 5: 723-731. 66. Crum, Anthony, Bassett and Folstein. 1993. Population-based norms for the MiniMental State Examination by age and educational level. JAMA 269, 18: 2386-91. 67. Das, U. 2003. Long-chain polyunsaturated fatt y acids in the growth and development of the brain and memory. Nutrition 19: 62-65. 68. DeKosky, S. 2003. Pathology and Pathways of Alzheimers Disease with an Update on New Developments in Treatment J Am Geriatr Soc 52:S314-320. 69. DeKosky, S. and Marek, K. 2003. Looking Backward to Move Forward: Early Detection of Neurodegenerative Disorders. Science 302: 830-834. 70. DeMattos, R., Bales, K., Cummins, D., D odart, J., Paul, S. and Holtzman, D. 2001. Peripheral anti-A antibody alters CNS and plasma A clearance and decreases brain A burden in a mouse model of Alzheimers disease. PNAS 98: 8850-8855. 71. Demetriades, A. 2002. Functional neuroi maging in Alzheimers type dementia. Journal of Neurological Sciences 203-204: 247-251.
276 72. de Wilde, M., Farkas, E., Gerrits, M., Kili aan, A. and Luiten, P. 2002. The effects of n-3 polyunsaturated fatty acid-rich di ets on cognitive and cerebrovascular parameters in chronic cerebral hypoperfusion. Brain Research 947: 166-173. 73. Dickey, C., Morgan, D., Kudchodkar, S., We iner, D., Bai, Y., Cao, C., Gordon, M. and Ugen, K. 2001. Duration and specifici ty of humoral immune responses in mice vaccinated with the Alzheimers disease-associated -amyloid 1-42 peptide. DNA and Cell Biology 20: 723-729. 74. Dodart, J., Meziane, H., Mathis, C, Unge rer, A., Bales, K. and Paul, S. 1999. Behavioral Disturbances in Transgenic Mice Overexpressing the V717F Amyloid Precursor Protein. Behavioral Neuroscience 113: 982-990. 75. Dodart, J., Mathis, C., Bales, K., Paul, S. and Ungerer, A. 2000a. Behavioral deficits in APPV717F transgenic mice deficient for the apolipoprotein E gene. Clinical Neuroscience and Neuropathology 11: 603-607. 76. Dodart, J., Mathis, C., Saura, J., Bales, K., Paul, S. and Ungerer, A. 2000b. Neuroanatomical Abnormalities in Be haviorally Characterized APPV717F Transgenic Mice. Neurobiology of Disease 7: 71-85. 77. Dodart, J., Bales, K., Gannon, K., Greene, D., DeMattos, R., Mathis, C., DeLong, C., Wu, S., Wu, X., Holtzman, D. and Paul, S. 2002. Immunization reverses memory deficits without reducing brain A burden in Alzheimers disease model. Nature 5: 452-457. 78. Doraiswamy, P., Krishnan, K., Anand, R ., Sohn, H., Danyluk, J., Hartman, R. and Veach, J. 2002. Long-term effects of ri vastigmine in moderately severe Alzheimers disease: does early initiati on of therapy offer sustained benefits? Prog Neuropsychopharmacol Biol Psychiatry 26: 705-712. 79. Dorion, A., Sarazin, M., Hasboun, D., HahnBarma, V., Dubois, B., Zouaoui, A., Marsault, C. and Duyme, M. 2002. Relationship between attentional performance and corpus callosum morphom etry in patients with Alzheimers disease. Neuropsychologia 40: 946-956. 80. Du, C., Sato, A., Watanabe, S., Wu, C., Ikemoto, A., Ando, K., Kikugawa, K., Fujii, Y. and Okuyama, H. 2003. Cholestero l synthesis in mice is suppressed buy lipofuscin formation is not affected by long-term feeding of n-3 fatty acidenriched oils compared with lard and n-6 fatty acid-enriched oils. Biol. Pharm. Bull. 26: 766-770. 81. Duncan, G., Miyamoto, S., Gu, H., Lieberma n, J., Koller, B. and Snouwaert, J. 2002. Alterations in regional brain metabolism in genetic and pharmacological models of reduced NMDA receptor function. Brain Res 951: 166-176.
277 82. Eckert, A., Keil, U., Marques, C., Bonert, A., Claudia, F., Schssel, K. and Mller, W. 2003. Mitochondrial dysfunction, apopt otic cell death, and Alzheimers disease. Biochemical Pharmacology 66: 1627-1634. 83. Eikelenboom, P., Bate, C., Gool, W., Hoozeman s, J., Rozemuller, J., Veerhuis, R. and Williams, A. 2002. Neuroinflammation in Alzheimers Disease and Prion Disease. Glia 40: 232-239. 84. El Khoury, J., Hickman, S., Thomas, C., Cao, L., Silverstein, S. and Loike, J. 1996. Scavenger receptor-mediated adhesion of microglia to -amyloid fibrils. Nature 382: 716-719. 85. Emery, V. 1996. Language functioning. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 166-192. Oxford: Oxford University Press. 86. Engelhart, M., Geerlings, M., Ruitenberg, A., van Swieten, J., Hofman, A., Witteman, J. and Breteler, M. 2002a. Diet and ri sk of dementia: Does fat matter? The Rotterdam Study. Neurology 59: 1915-1921. 87. Engelhart, M., Geerlings, M., Ruitenberg, A., van Swieten, J., Hofman, A., Witteman, J. and Breteler, M. 2002b. Dietary intake of antioxidants and risk of Alzheimer disease. JAMA 287: 3223-3229. 88. Eriksen, J., Sagi, S., Smith, T., Weggen, S., Das, P., McLendon, C., Ozols, V., Jessing, K., Zavitz, K., Koo, E. and Golde, T. 2003. NSAIDs and enentiomers of fluribiprofen target gamma-secr etase and lower Abeta 42 in vivo. Clin Invest 112: 440-449. 89. Fassbender, K., Simons, M., Bergmann, C., Stroick, M., Ltjohann, D., Keller, P., Runz, H., Khl, S., Bertsch, T., von Berg mann, K., Nennerici, M., Beyreuther, K. and Hartmann, T. 2001. Simvastatin str ongly reduces levels of Alzheimers disease b-amyloid peptides Ab42 and Ab40 in vitro and in vivo. PNAS 98: 58565861. 90. Fassbender, K., Stroick, M., Bertsch, T ., Ragoschke, A., Kuehl, S., Walter, S., Walter, J., Brechtel, K., Muechlhauser, F., Von Bergmann, K. and Lutjohann, D. 2002. Effects of statins on human cerebra l cholesterol metabolism and secretion of Alzheimer amyloid peptide. Neurology 59: 1257-1258. 91. Favrelire, S., Perault, M., Juguet, F., De Javel, D., Bertrand, N., Piriou, A. and Durand, G. 2003. DHA-enriched phospholipid diets modulate age-related alterations in rat hippocampus. Neurobiology of Aging 24: 233-243.
278 92. Feldmann, H., Gauthier, S., Hecker, J., Vella s, B., Subbiah, P. and Whalen, E. 2001. A 24-week, randomized, double-blind study of donepezil in moderate to severe Alzheimers disease. Neurology 57: 613-620. 93. Ferdinandusse, S., Denis, S., Mooijer, P., Zhang, Z., Reddy, J., Spector, A. and Wanders, R. 2001. Identification of the peroxisomal -oxidation enzymes involved in the biosynthesi s of docosahexaenoic acid. Journal of Lipid Research 42: 1987-1995. 94. Folch, J., Lees, M. and Sloane Stanley, G. 1957. A simple method for the isolation and purification of total lipi ds from animal tissue. J Biol Chem 226: 495-509. 95. Frenkel, D., Balass, M. and Solomon, B. 1998. N-terminal EFRH sequence of Alzheimers -amyloid peptide represents the epitope of its anti-aggregating antibodies Journal of Neuroimmunology 88: 85-90. 96. Frenkel, D., Balass, M., Katchalski-Katzi r, E. and Solomon, B. 1999. High affinity binding of monoclonal antibodies to the sequential epitope EFRH of -amyloid peptide is essential for modulat ion of fibrillar aggregation. Journal of Neuroimmunology 95: 136-142. 97. Frenkel, D., Dewachter, I., Van Leuve n, F. and Solomon, B. 2003. Reduction of amyloid plaques in brain of transgenic mouse model of Alzheimers disease by EFRH-phage administration. Vaccine 21: 1060-1065. 98. Fukutani, Y., Cairns, N., Shiozawa, M., Sasa ki, K., Sudo, S., Isaki, K. and Lantos, P. 2000. Neuronal loss and neurofibrillary de generation in the hippocampal cortex in late-onset sporadic Alzheimers disease. Psychiatry Clin Neurosci 54: 523529. 99. Games, D., Adams, D., Alessandrini, R., Ba rbour, R., Berthelette, P., Blackwell, C., Carr, T., Clemens, J., Donaldson, T., Gillespie, F., Guido, T., Hagopian, S., Johnson-Wood, K., Khan, K., Lee, M., Leibowitz, P., Lieberburg, I., Little, S., Masliah, E., McConlogue, L., Montoya-Za vala, M., Mucke, L., Paganini, L., Penniman, E., Power, M., Schenk, D., Seube rt, P., Snyder, B., Soriano, F., Tan, H., Vitale, J., Wadsworth, S., Wolozin, B. and Zhao, J. 1995. Alzheimer-type neuropathology in transgenic mice overexpressing V717F beta-amyloid precursor protein. Nature 373: 523-527. 100. Gamoh, S., Hashimoto, M., Hossain, S. and Msumura, S. 2001. Chronic administration of docosahexaenoic acid im proves the performance of radial arm maze task in aged rats. Clinical and Experimental Pharmacology and Physiology 28: 266-270.
279 101. Gasic-Milenkovic, J., Loske, C. a nd Munch, G. 2003. Advanced glycation endproducts cause lipid peroxidation in the human neuronal cell line SH-SY5Y. J Alzheimers Dis 5: 25-30. 102. Gauthier, S. 1998. Update on diagnostic methods, natural history and outcome variables in Alzheimers disease. Dement Geriatr Cogn Disord 9; Suppl 3: 2-7. 103. Gee, J., Ding, L., Xie, Z., Lin, M., DeVita, C. and Grossman, M. 2003. Alzheimers disease and frontotempora l dementia exhibit distinct atrophybehavior correlates: a com puter-assisted imaging study. Acad Radiol 10: 13921409. 104. Glckner, F., Meske, V. and Ohm, T. 2002. Genotype-related differences of hippocampal apolipoprotein E levels only in early stages of neuropathological changes in Alzheimers disease. Neuroscience 114: 1103-1114. 105. Gray, C. and Sala, S. 1996. Charting decline in dementia. Chap. In The Cognitive Neuropsychology of Alzh eimer-type Dementia. 23-46. Oxford: Oxford University Press. 106. Greene, J. and Hodges, J. 1996. Semantic processing. Chap. In The Cognitive Neuropsychology of Alzh eimer-type Dementia. 128-148. Oxford: Oxford University Press. 107. Greenwood, C. and Winocur, G. 1996. Cognitive impairment in rats fed high-fat diets: a specific effect of saturated fatty acid-intake. Behav Neurosci 110: 451459. 108. Grossberg, G., Irwin, P., Satlin, A., Me senbrink, P. and Spiegel, R. 2004. Rivastigmine in Alzheimers disease: efficacy over two years. Am J Geriatr Psychiatry 12: 420-431. 109. Grynberg, A., Fournier, A., Sergiel, J. and Athias, P. 1995. Effect of docosahexaenoic acid and eicosapentaenoic acid in the phospholipids of rat heart muscle cells on adrenoceptor re sponsiveness and mechanism. J Mol Cell Cardiol 27: 2507-2520. 110. Gordon, M, King, D., Diamond, D., Jantzen, P ., Boyett, K., Hope, C., Hatcher, J., DiCarlo, G., Gottschalk, P., Morgan, D. and Arendash, G. 2001. Correlation between cognitive deficits and A deposits in transgenic APP+PS1 mice. Neurobiology of Aging 22: 377-386. 111. Gordon, M., Holcomg, L., Jantzen, P., Di Carlo, G., Wilcock, D., Boyett, K., Connor, K., Melachrino, J., OCallaghan, J. and Morgan, D. 2002. Time course
280 of the development of Alzheimer-lik e pathology in the doubly transgenic PS1+APP mouse. Experimental Neurology 173: 183-195. 112. Haag, M. 2003. Essential fa tty acids and the brain. The Canadian Journal of Psychiatry 113. Haan, M. and Wallace, R. 2004. Can dementia be prevented: Brain aging in a population-based context. Annual Review of Public Health 25: 1-24. 114. Haas, C., Lemere, C., Capell, A., Citron, M., Seubert, P., Schenk, D., Lannefelt, L. and Selkoe, D. 1995. The Swedish muta tion causes early-onset Alzheimers disease by beta-secretase cleavage within the secretory pathway. Nature Medicine 1: 1291-1296. 115. Hampel, H., Buerger, K., Zinkowski, R., Teipel, S., Goernitz, A., Andreasen, N., Sjoegren, M., DeBernardis, J., Kerkman, D., Ishiguro, K., Ohno, H., Vanmechelen, E., Vanderstichele, H., McCu lloch, C., Moller, H., Davies, P. and Blennow, K. 2004. Measurement of phosphorylated tau epitopes in the differential diagnosis of Alzheimer dis ease: a comparative cerebrospinal fluid study Arch Gen Psychiatry 61: 95-102. 116. Hampel, H., Goernitz, A. and Buerger, K. 2003. Advances in the development of biomarkers for Alzheimers disease: from CSF total tau and A 1-42 proteins to phophorylated tau protein. Brain Research Bulletin 61: 234-253. 117. Hampson, A., Grimaldi, M., Axelrod, J. and Wink, D. 1998. Cannabidiol and (-) 9tetrahydrocannabinol are ne uroprotective antioxidants. PNAS 95: 8268-8273. 118. Hashimoto, M., Tanabe, Y., Fujii, Y., Kikuta, T., Hitoshi, S. and Shido, O. 2005. Chronic administration of docosahexaenoic acid ameliorates the impairment of spatial cognition learni ng ability in amyloid -infused rats. Nutritional Neuroscience 135: 549-555. 119. Hashimoto, M., Hossain, S., Shimada, T., Sugioka, K., Yamasaki, H., Fujii, Y., Ishibashi, Y., Oka, J. and Shido, O. 2002. Docosahexaenoic acid provides protection from impairment of learning abil ity in Alzheimers disease model rats. Journal of Neurochemistry 81: 1084-1091. 120. Hashimoto, M., Tanabe, Y., Fujii, Y., Kikut a, T., Shibata, H. and Shido, O. 2005. Chronic administration of docosahexaenoic acid ameliorates the impairment of spatial cognition learning ability in amyloid beta-infused rats. J Nutr 135: 549555.
281 121. Harris, M., Hensley, K., Butterfield, D., Leedle, R. and Carney, J. 1995. Direct evidence of oxidative injury produced by the Alzheimers beta-amyloid peptide (1-40) in cultured hippocampal neurons. Exp Neurol 131: 193-202. 122. Hensley, K., Carney, J., Mattson, M., Akse nova, M., Harris, M., Wu, J., Floyd, R. and Butterfield, D. 1994. A model for -amyloid aggregation and neurotoxicity based on free radical generation by the peptid e: relevance to Alzheimers disease. PNAS 91: 3270-3274. 123. Hensley, K., Hall, N., Subramaniam, R., Cole, P., Harris, M., Aksenov, M., Aksenova, M., Gabbita, P., Wu, J., Carne y, J., Lovell, M., Markesberry, W. and Butterfield, A. 1995. Brain regional corre spondence between Alzheimers disease histopathology and biomarke rs of protein oxidation J Neurochem 65: 2146-2156. 124. Heude, B., Ducimetiere, P. and Berr, C. 2003. Cognitive decline and fatty acid composition of erythrocyte membranes-The EVA Study. Am J Clin Nutr 77: 803808. 125. Hirayama, A., Horikoshi, Y., Maeda, M., Ito, M. and Takashima, S. 2003. Characteristic developmental expression of amyloid beta40, 42 and 43 in patients with Down syndrome. Brain Dev 25: 180-185. 126. Hock, C., Koneitxko, U., Papassotiropoulos, A., Wollmer, A., Streffer, J., von Rotz, R., Davey, G., Moritz, E and Nitsch, R. 2002. Generation of antibodies specific for -amyloid by vaccination of patients with Alzheimer disease. Nature 127. Hock, C., Konietzko, U., Streffer, J., Trac y, J., Signorell, A., Muller-Tillmanns, B., Lemke, U., Kenke, K., Moritz, E., Garc ia, E., Wollmer, M., Umbricht, D., de Quervain, D., Hofmann, M., Maddalena, A ., Papassotiropoulos, A. and Nitsch, R. 2003. Antibodies against -amyloid slow cognitive decline in Alzheimers disease. Neuron 38: 547-554. 128. Holcomb, L., Gordon, M., McGowan, E., Yu, X., Benkovic, S., Jantzen, P., Wright, K., Saad, I., Mueller, R., Morgan, D., Sanders, S., Zehr, C., OCampo, K., Hardy, J., Prada, C., Eckman, C., Younkin, S., Hs aio, K. and Duff, K. 1998. Accelerated Alzheimer-type phenotype in transgenic mice carrying both mutant amyloid precursor protein and presenilin 1 transgenes. Nature Medicine 4: 97-100. 129. Holcomb, L, Gordon, M., Jantzen, P., Hsiao, K., Duff, K. and Morgan, D. 1999. Behavioral Changes in Transgenic Mi ce Expressing Both Amyloid Precursor Protein and Presenilin-1 Mutations: Lack of Association with Amyloid Deposits. Behavior Genetics 29: 177-185. 130. Horrocks, L. and Yeo, Y. 1999. Health benefits of docosahexaenoic acid (DHA). Pharmacological Research 40: 211-225.
282 131. Hsiao, K., Borchelt, D., Olson, K., Johannsdot tir, R., Kitt, C., Yunis, W., Xu, S., Eckman, C., Younkin, S., Price, D. et al. 1995. Age-related CNS disorder and early death in transgenic FVB/N mi ce overexpressing Alzheimer amyloid precursor proteins. Neuron 15: 1203-1218. 132. Hui, J., Wilson, R., Bennett, D., Bienias, J ., Gilley, D. and Evans, D. 2003. Rate of cognitive decline and mortality in Alzheimers disease. Neurology 61: 13561361. 133. Huitron-Resendiz, S., Sanchez-Alavez, M ., Gallegos, R., Berg, G., Crawford, E., Giacchino, J., Games, D., Henriksen, S. and Criado, J. 2002. Age-independent and age-related deficits in vis uospatial learning, sleep-wake states, thermoregulation and motor activity in PDAPP mice. Brain Research 928: 126137. 134. Hulstaert, F., Blennow, K., Ivanoiu, A., Schoonderwaldt, H., Riemenschneider, M., De Deyn, P., Bancher, C., Cras, P., Wiltfang, J., Mehta, P., Iqbal, K., Pottel, H., Vanmechelen, E. and Vanderstichele, H. 1999. Improved discrimination of AD patients using beta-amyloid (1-42) and tau levels in CSF. Neurology 52: 15551562. 135. Igbavboa, U., Hamilton, J., Kim, H., Sun, G. and Wood, W. 2002. A new role for apolipoprotein E: modulating transp ort of polyunsaturated phospholipid molecular species in synaptic plasma membranes. Journal of Neurochemistry 80: 255-261. 136. Ikegami, S., Shumiya, S. and Kawamura, H. 1992. Age-related changes in radialarm maze learning and basal forebrain cholinergic systems in senescence accelerated mice (SAM). Behav Brain Res 51: 15-22. 137. Ikemoto, A., Nitta, A., Furukawa, S., Oh ishi, M., Nakamura, A., Fujii, Y. and Okuyama, H. 2000. Dietary n-3 fatty acid deficiency decreases nerve growth factor content in rat hippocampus. Neuroscience Letters 285: 99-102. 138. Ikemoto, A., Ohishi, M., Sato, Y., Hata, N ., Misawa, Y., Fujii, Y. and Okuyama, H. 2001. Reversibility of n-3 fatty acid de ficieny-induced alte rations of learning behavior in the rat: level of n-6 fatty acids as another critical factor. Journal of Lipid Research 42: 1655-1663. 139. in t Veld, B., Ruitenberg, A., Hofman, A., Launer, L., van Duijin, C., Stijnen, T., Breteler, M. and Stricker, B. 2001. Nonste roidal antiinflamma tory drugs and the risk of Alzheimers disease. N Engl J Med 345: 1515-1521.
283 140. in t Veld, B., Launer, L., Breteler, M., Hofman, A. and Stricker, B. 2002. Pharmacologic Agents Associated with a Preventive Effect on Alzheimers Disease: A Review of the Epidemiologic Evidence. Epidemiologic Reviews 24: 248-268. 141. Irizarry, M., McNamara, M., Fedorchak, K ., Hsiao, K. and Hyman, B. 1997. APPsw transgenic mice develop age-related A deposits and neuropil abnormalities, but no neuronal loss in CA1. Journal of Neuropathol ogy and Experimental Neurology 56: 965-973. 142. Itokazu, N., Ikegaya, Y., Nishikawa, M. and Matsuki, N. 2000. Bidirectional actions of docosahexaenoic acid on hi ppocampal neurotransmissions in vivo. Brain Research 862: 211-216. 143. Iuvone, T., Esposito, G., Esposito, R., Santam aria, R., Di Rosa, M. and Izzo, A. 2004. Neuroprotective effect of cannabidi ol, a non-psychoactive component from Cannabis sativa, on beta-amyloid-induced toxicity in PC12 cells. J Neurochem 89: 134-141. 144. Jantzen, P., Connor, K., DiCarlo, G., Wenk, G., Wallace, J., Rojiani, A., Coppola, D., Morgan, D. and Gordon, M. 2002. Microglial activation and -amyloid deposit reduction caused by a nitric oxide-releasing nonsteriodal antiinflammatory drug in amyloid precursor prot ein plus presenilin-1 transgenic mice. J Neurosci 22: 2246-2254. 145. Janus, C., Pearson, J., McLaurin, J., Mathew s, P., Jiang, Y., Schimdt, S., Chishti, M., Horne, P., Heslin, D., French, J., Mount, H., Nixon, R., Mercken, M., Bergeron, C., Fraser, P., St George-H yslop, P. and Westaway, D. 2000. A peptide immunization reduces behavioral impairment and plaques in a model of Alzheimers disease. Nature 408: 979-982. 146. Janus, C. and Westaway, D. 2001. Tran sgenic mouse models of Alzheimers disease. Physiology & Behavior 73: 873-886. 147. Jellinger, K. 2002. Alzheimer disease a nd cerebrovascular pa thology: an update. J Neural Transm 109: 813-836. 148. Jensen, M., Skarsfeldt, T. and Hy, C. 1996. Correlation between level of (n-3) polyunsaturated fatty acids in brain phospholip ids and learning ability in rats. A multiple generation study. Biochimica et Biophysica Acta 1300: 203-209. 149. Jensen, M., Mottin, M., Cracchiolo, J., Leighty, R. and Arendash, G. 2005. Lifelong immunization with human -amyloid (1-42) protects Alzheimers transgenic mice against cogniti ve impairment throughout aging. Neuroscience 130: 667-684.
284 150. Jick, H., Zornberg, G., Jick, S., Seshad ri, S. and Drachman, D. 2000. Statins and the risk of dementia. The Lancet 356: 1627-1631. 151. Jones, C., Arai, T. and Rapoport, S. 1997. Evidence for the involvement of docosahexaenoic acid in cholinergic stimulat ed signal transduction at the synapse. Neurochemical Research 22: 663-670. 152. Jones, R., Soininen, H., Hager, K., Aa rsland, D., Passmore, P., Murthy, A., Zhang, R. and Bahra, R. 2004. A multinationa l, randomized, 12-week study comparing the effects of donepezil and galantamine in patients with mild to moderate Alzheimers disease Int J Geriatr Psychiatry 19: 58-67. 153. Joseph, J., Shukitt-Hale, B., Denisova, N., Prior, R., Cao, G., Martin, A., Taglialatela, G and Bickford, P. 1998. L ong-term dietary strawberry, spinach, or vitamin E supplementation retards the onset of age-related neuronal signaltransduction and cognitive behavioral deficits. Journal of Neuroscience 18: 80478055. 154. Joseph, J., Shukitt-Hale, B., Denisova, N., Bielinski, D., Martin, A., McEwen, J and Bickford, P. 1999. Reversals of age-re lated declines in neuronal signal transduction, cognitive, and motor behavior al deficits with blueberries, spinach, or strawberry dietary supplementation. Journal of Neuroscience 19: 8114-8121. 155. Julin, P., Almkvist, O., Basun, F., Lannf elt, L., Svensson, L., Winblad, B. and Wahlund, L. 1998. Brain volumes and regiona l cerebral blood flow in carriers of the Swedish Alzheimer amyloid protein mutation. Alzheimer Dis Assoc Disord 12: 49-53. 156. Kalmijn, S., Launer, L., Ott, A., Witteman, J., Hofman, A. and Breteler, M. 1997a. Dietary fat intake and risk of incide nt dementia in the Rotterdam Study. Annals of Neurology 42: 776-782. 157. Kalmijn, S., Feskens, E., Launer, L. and Kromhout, D. 1997b. Polyunsaturated fatty acids, antioxidants and cogniti ve function in very old men. American Journal of Epidemiology 145: 33-41. 158. Kalmijn, S., van Boxtel, M., Ocke, M., Ve rschuren, W., Dromhout, D. and Launer, L. 2004. Dietary intake of fatty aci ds and fish in relation to cognitive performance at middle age. Neurology 62: 275-280. 159. Kammoun, S., Gold, G., Bouras, C., Gia nnakopoulos, P. McGee, W., Herrmann, F. and Michel, J. 2000. Immediate causes of death of demented and non-demented elderly. Acta Neurol Scand Suppl 176: 96-99.
285 160. Kanai, M., Matsubara, E., Isoe, K., Urakam i, K., Nakashima, K., Arai, H., Sasaki, H., Abe, K., Iwatsubo, T., Kosaka, T ., Watanabe, M., Tomidokoro, Y., Shizuka, M., Mizushima, K., Nakamura, T., Igeta, Y., Amari, M., Kawarabayashi, T., Ishiguro, K., Harigaya, Y., Wakabayashi, K ., Okamato, K., Hirai, S. and Shoji, M. 1998. Longitudinal study of cerebrospinal fluid levels of ta u, A beta1-40, and A beta1-42(43) in Alzheimers disease: a study in Japan. Ann Neurol 44: 17-26. 161. Kang, J., Wang, J., Wu, L. and Kang, Z. 2004. Fat-1 mice convert n -6 to n -3 fatty acids. Nature 427: 504. 162. Kaufer, D., Cummings, J. and Christine, D. 1996. Effect of tacrine on behavioral symptoms in Alzheimers dise ase: an open-label study. J Geriatr Psychiatry Neurol 9: 1-6. 163. Kelly, P., Bondolfi, L., Hunziker, D., Sc hlecht, H., Carver, K., Maquire, E., Abramowski, D., Wiederhold, K., Sturchle r-Pierrat, C., Jucker, M., Bergmann, R., Staufenbiel, M. and Sommer, B. 2003. Pr ogressive age-related impairment of cognitive behavior in APP23 transgenic mice. Neurobiol Aging 24: 365-378. 164. Kidron, D. and Freedman, M. 1996. Motor functioning. Chap. In The Cognitive Neuropsychology of Alzh eimer-type Dementia. 206-220. Oxford: Oxford University Press. 165. King, D. and Arendash, G. 2002a. Beha vioral characterization of the Tg2576 transgenic model of Alzheimers disease through 19 months. Physiology & Behavior 75: 627-642. 166. King, D., and Arendash, G. 2002b. Mainta ined synaptophysin immunoreactivity in Tg2576 transgenic mice during aging: corr elations with cognitive impairment. Brain Research 926: 58-68. 167. Kivipelto, M., Helkala, E., Laakso, M., Hanninen, T., Hallikaiene, M., Alhainen, K., Iivonen, S., Mannermaa, A., Tuomilehto, J., Nissinen, A. and Soininen, H. 2002. Apolipoprotein E e4 allele, elevated midlife total cholesterol level, and high midlife systolic blood pressure are independent risk factors for late-life Alzheimer disease. Ann Intern Med 137: 149-155. 168. Klegeris, A., Walker, D. and McGeer P. 1994. Activation of macrophages by Alzheimer amyloid peptide. Biochem Biophys Res Commun 199: 984-991. 169. Klunk, W., Engler, H., Nordberg, A., Wang, Y., Blomqvist, G., Holt, D., Bergstrom M., Savitcheva, I., Huang, G., Estrada, S., Ausen, B., Debnath, M., Barletta, J., Price, J., Sandell, J., Lopresti, B., Wall, A., Koivisto, P., Antoni, G., Mathis, C. and Langstrom B. 2004. Imaging brain amyloid in Alzheimers disease with Pittsburgh Compound-B. Ann Neurol 55: 306-319.
286 170. Knittweis, J. 1999. Weight loss in cancer and Alzheimers disease is mediated by a similar pathway. Medical Hypotheses 53: 172-174. 171. Koller, M., Mohajeri, M., Huber, M., Wollm er, M., Roth Zgraggen, B., Sandmeier, E., Moritz, E., Tracy, J., Nitsch, R. a nd Christen, P. 2004. Active immunization of mice with an A -Hsp70 vaccine. Neurodegenerative Diseases 1: 20-28. 172. Komatsu, W., Ishihara, K., Murata, M ., Saito, H. and Shinohara, K. 2003. Docosahexaenoic acid suppresses nitric oxide production and inducible nitric oxide synthase expression in inte rferon-gamma plus lipopolysaccharidestimulated murine macrophages by i nhibitig the oxidative stress. Free Radic Biol Med 34: 1006-1016. 173. Kotilinek, L., Bacskai, B., Westerman, M., Kawarabayashi, T., Younkin, L., Hyman, B., Younkin, S. and Ashe, K. 2002. Reversible memory loss in a mouse transgenic model of Alzheimers disease. The Journal of Neuroscience 22: 63316335. 174. Kruman, I., Kumaravel, T., Lohani, A., Pedersen, W., Cutler, R., Kruman, Y., Haughey, N., Lee, J., Evans, M. and Matts on, M. 2002. Folic acid deficiency and homocysteine impair DNA re pair in hippocampal neurons and sensitize them to amyloid toxicity in experimental models of Alzheimers disease. Journal of Neuroscience 22: 1752-1762. 175. Kurz, A., Erkinjuntti, T., Small, G., Lilienfeld, S. and Damaraju, C. 2003. Longterm safety and cognitive effects of gala ntamine in treatment of probable vascular dementia or Alzheimers disease with cerebrovascular disease. Eur J Neurol 10: 633-640. 176. Laghmouch, A., Bertholet, J. and Crusi o, W. 1997. Hippocampal morphology and open-field behavior in Mus musculus domesticus and Mus spretus inbred mice. Behav Genet 27: 67-73. 177. Lalonde, R., Dumont, M., Staufenbiel, M., St urchler-Pierrat, C. and Strazielle, C. 2002. Spatial learning, exploration, anxiet y, and motor coordination in female APP23 transgenic mice with the Swedish mutation. Brain Res 956: 36-44. 178. Larner, A. and du Plessis, D. 2003. Ea rly-onset Alzheimers disease with presenilin-1 M139V mutation: clinical, meuropsychological and neuropathological study. Eur J Neurol 10: 319-323. 179. Laurin, D., Masaki, K., Foley, D., White, L. and Launer, L. 2004. Midlife dietary intake of antioxidants and risk of latelife incident dementia: the Honolulu-Asia Aging Study. Am J Epidemiol 159: 959-967.
287 180. Le Bars, P., Katz, M., Berman, N., Itil, T., Freedman, A. and Schatzberg, A. 1997. A placebo-controlled, double-blind, randomi zed trial of an extract of Ginkgo biloba for dementia. North American EGb Study Group. JAMA 278: 1327-1332. 181. Le Bars, P., Kieser, M. and Itil, K. 2000. A 26-week analysis of a double-blind, placebo-controlled trial of the gingko biloba extract EGb 761 in dementia. Dement Geriatr Cogn Disord 11: 230-237. 182. Le Bars, P., Velasco, F., Ferguson, J., De ssain, E., Kieser, M and Hoerr, R. 2002. Influence of the severity of cognitive impairment on the effect of the Gnkgo biloba extract EGb 761 in Alzheimers disease. Neuropsychobiology 45: 19-26. 183. Lee, B., Mintun, M., Buckner, R. and Morris, J. 2003. Imaging of Alzheimers Disease. Journal of Neuroimaging 13, 3: 199-214. 184. Leighty, R., Nilsson, L., Potter, H., Costa, D., Low, M., Bales, K. Paul, S. and Arendash, G. 2004. Use of multimetric stas tical analysis to characterize and discriminate between the performance of four Alzheimers transgenic mouse lines differing in Abeta deposition. Behav Brain Res 153: 107-121. 185. Lemere, C., Lopera, F., Kosik, K., Lendon, C., Ossa, J., Saido, T., Yamaguchi, H., Ruiz, A., Martinez, A., Madrigal, L., Hincapie, L., Arango, J., Anthony, D., Koo, E., Goate, A., Selkoe, D. and Arango, J. 1996. The E280A presenilin 1 Alzheimer mutation produces increase d A beta 42 deposition and severe cerebellar pathology. Nature Medicine 2: 1146-1150. 186. Lepage, G. and Roy, C. 1986. Di rect transesterification of all classes of lipids in a one-step reaction J Lipid Res 27: 114-120. 187. Lesne, S., Ali, C., Gabriel, C., Croci, N ., MacKenzie, E., Glabe, C., Plotkine, M,. Marchand-Verrecchia, C., Vivien, D. and Buisson, A. 2005. The Journal of Neuroscience 25: 9367-9377. 188. Lim, S. and Suzuki, H. 1999. Intakes of dietary docosahexaenoic acid ethyl ester and egg phosphatidylcholine improves maze-learning ability in young and old mice. Nutritional Neurosciences 1629-1632. 189. Lim, S. and Suzuki, H. 2000. Changes in maze behavior of mice occur after sufficient accumulation of docosahexaenoic acid in brain. Nutritional Neurosciences 319-324. 190. Lim, G., Yang, F., Chu, T., Chen, P., Beec h, W., Teter, B., Tran, T., Ubeda, O., Hsiao Ashe, K., Frautschy, S. and Cole G. 2000. Ibuprofen suppresses plaque
288 pathology and inflammation in a mouse model for Alzheimers disease. J Neurosci 20: 5709-5714. 191. Lim, G., Calon, F., Morihara, T., Yang, F ., Teter, B., Ubeda, O., Salem, N., Frautshy, S. and Cole, G. 2005. A Diet Enriched with Omega-3 Fatty Acid Docosahexaenoic Acid Reduces Amyloid Burden in an Aged Alzheimer Mouse Model. The Journal of Neuroscience 25: 3032-3040. 192. Ling, Y., Morgan, K. and Kalsheker, N. 2003. Amyloid precursor protein (APP) and the biology of proteolytic processing: relevance to Alzheimers disease. The International Journal of Biochemistry & Cell Biology 35: 1505-1535. 193. Litman, B., Niu, S., Polozova, A. and Mitchell, D. 2001. The role of docosahexaenoic acid containing phospholip ids in modulating G protein-coupled signaling pathways: visual transduction. J Mol Neurosci 16: 237-243. 194. Liu, B. and Hong, J. 2002. Role of Miecroglia in Inflammation-Mediated Neurodegenerative Diseases: Mechanis ms and Strategies for Therapeutic Intervention. The Journal of Pharmocology and Experimental Therapeutics 304, 1: 1-7. 195. Lombardo, J., Stern, E., McLellan, M., Kajd asz, S., Hickey, G., Bacskai, B. and Hyman, B. 2003. Amyloidantibody treatment leads to rapid normalization of plaque-induced neuritic alterations. J Neurosci 23: 10879-10883. 196. Lukiw, W. and Bazan, N. 2000. Neuroinf lammatory Signaling Upregulation in Alzheimers Disease. Neurochemical Research 25, 9/10: 1173-1184. 197. Lukiw, W., Cui, J., Marcheselli, V., Bodker, M., Botkjaer, A., Gotlinger, K., Serhan, C. and Bazan, N. 2005. A role for docosahexaenoic acid-derived neuroprotectin D1 in neural cell su rvival and Alzheimer disease. Journal of Clinical Investigation 1-16. 198. Maccioni, R., Mun oz, J. and Barbeito, L. 2001. The Molecular Bases of Alzheimers Disease and Other Neurodegenerative Disorders. Archives of Medical Research 32: 367-381. 199. Marcon, G., Giaccone, G., Cupidi, C., Balest rieri, M., Beltrami, C., Finato, N., Bergonzi, P., Sorbi, S., Bugiani, O. and Tagliavini, F. 2004. Neuropathological and clinical phenotype of an Italian Alzheimer family with M239V mutation of presenilin 2 gene. J Neuropathol Exp Neurol 63: 199-209. 200. Marin, M., Rey, G., Pedersoli, L., Rodri go, M. and de Alaniz, M. 2000. Dietary long-chain fatty acids and visual respons e in malnourished nursing infants. Prostaglandins, Leukot Essent Fatty Acids 63: 385-390.
289 201. Martin, D., Spencer, P., Horrobin, D. a nd Lynch, M. 2002. Long-term potentiation in aged rats is restored when the age-re lated decrease in poly unsaturated fatty acid concentration is reversed. Prostaglandins, Leukotri enes and Esse ntial Fatty Acids 67: 121-130. 202. Marutle, A., Warpman, U., Bogdanovic, N ., Lannfelt, L. and Nordberg, A. 1999. Neuronal nicotinic receptor deficits in Alzheimer patients with the Swedish amyloid precursor protein 670/671 mutation. Journal of Neurochemistry 72: 1161-1169. 203. Masaki, K., Losonczy, K., Izmirlian, G., Fo ley, D., Ross, G., Petrovitch, H., Havlik, R. and White, L. 2000. Association of vitamin E and C supplement use with cognitive function and dementia in elderly men. Neurology 54: 1265-1272. 204. Masliah, E., Sisk, A., Mallory, M., Muck e, L., Schenk, D. and Games, D. 1996. Comparison of neurodegernative pathol ogy in transgenic mice overexpressing V717F -amyloid precursor protein and Alzheimers disease. J Neurosci 16: 5795-5811. 205. Masterman, D. 2003. Treatment of the Neur opsychiatric Symptoms in Azlheimers Disease. J Am Med Dir Assoc. 4(6 Suppl): S146-54. 206. Mathias, J. 1996. Reading disorder in Alzheimer-type dementia. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 149-165. Oxford: Oxford University Press. 207. Mathis, C., Bacskai, B., Kajdasz, S., Mc Lellan, M., Frosch, M., Hyman, B., Holt, D., Wang, Y., Huang, G., Debna th, M. and Klunk, W. 2002. Bioorg Med Chem Lett 12: 295-298. 208. Matthews, H., Korbey, J., Wilkinson, D. and Rowden, J. 2000. Donepezil in Alzheimers disease: eighteen month re sult from Southampton Memory Clinic. Int J Geriatr Psychiatry 15: 713-720. 209. Maurer, K., Ihl, R., Dierks, T. and Fro lich, L. 1997. Clinical efficacy of Ginkgo biloba special extract EGb 761 in de mentia of the Alzheimer type. J Psychiatr Res 31: 645-655. 210. McGeer, P., Schultzer, M. and McGeer, E. 1996. Arthritis and anti-inflammatory agents as possible protective factors fo r Alzheimers disease: a review of 17 epidemiologic studies. Neurology 47: 425-432. 211. McGeer, P. and McGeer, E. 2002. Is there a future for vaccination as a treatment for Alzheimers disease? Neurobiology of Aging 5787:
290 212. McGahon, B., Martin, D., Horrobin, D. a nd Lynch, M. 1999. Age-related changes in synaptic function: Analysis of the effect of dietary supplementation with -3 fatty acids. Neuroscience 94: 305-314. 213. McLaurin, J., Cecal, R., Kierstead, M., Phinney, A., Manea, M., French, J., Lambermon, M., Darabie, A., Brown, M., Janus, C., Chishti, M., Horne, P., Westaway, D., Fraser, P., Mount, H., Prz ybylski, M. and St George-Hyslop, P. 2002. Therapeutically effectiv e antibodies against amyloidpeptide target amyloid-b residues 4-10 and inhibit cy totoxicity and fibrillogenesis. Nature 214. Mehlhorn, G., Hollborn, M. and Schliebs, R. 2000. Induction of cytokines in glial cells surrounding cortical -amyloid plaques in transgenic Tg2576 mice with Alzgheimers pathology. Int J Devl Neuroscience 18: 423-431. 215. Metcalf-R., James, M., Mantzioris, E. a nd Cleland, L. 2003. A practical approach to increasing intakes on n3 polyunsaturated fatty acids: use of novel foods enriched with n-3 fats. European Journal of Clinical Nutrition 57: 1605-1612. 216. Meydani, S., Endres, S., Woods, M. 1991. Oral (n-3) fatty acid supplementation suppresses cytokine production and ly mphocyte proliferation: Comparison 217. Meydani, S. 1996. Effect of (n-3) pol yunsaturated fatty acids on cytokine production and their biologic function. Nutrition 12: S8-S14. 218. Meydani, M. 2002. Nutrition interventions in aging and age-associated disease. Proceedings of the Nutrition Society 61: 165-171. 219. Meydani, M. 2003. Soluble adhesion molecules: surrogate markers of cardiovascular disease? Nutr Rev 61: 63-68. 220. Michikawa, M. 2003. Cholesterol paradox: Is high total or low HDL cholesterol level a risk for Alzheimers disease. Journal of Neuroscience Research 72: 141146. 221. Miklossy, J., Taddei, K., Suva, D., Verdile, G., Fonte, J., Fisher, C., Gnjec, A., Ghika, J., Suard, F., Mehta, P., McLean, C ., Masters, C., Brooks, W. and Martins, R. 2003. Two novel presenilin-1 mutations (Y256S and Q222H) are associated with early-onset Alzheimers disease. Neurobiol Aging 24: 655-662. 222. Minagar, A., Shapshak, P., Fujimura, R., Ow nby, R., Heyes, M. and Eisdorfer, C. 2002. The role of macrophage/microglia and astrocytes in the pathogenesis of three neurologic disorders: HIV-associat ed dementia, Alzheimer disease, and multiple sclerosis. Journal of the Neurological Sciences 202: 13-23.
291 223. Minami, M., Kimura, S., Endo, T., Hamaue, N., Hirafuji, M., Togashi, H., Matsumoto, M., Yoshioka, M., Saito, H., Watanabe, S., Kobayashi, T. and Okuyama, H. 1997. Dietary docosahex aenoic acid increases cerebral acetylcholine levels in improves passive avoidance performance in stroke-prone spontaneously hypertensive rats. Pharmacology Bioche mistry and Behavior 58: 1123-1129. 224. Mineur, R. and Crusio, W. 2002. Behavior al and neuroanatomi cal characterization of FVB/N inbred mice. Brain Res Bull 57: 41-47. 225. Mitchell, D., Niu, S. and Litman, B. 2 003. Enhancement of G protein coupled signaling by DHA phospholipids. Lipids 38: 437-443. 226. Moechars, D., Lorent, K., De Strooper, B ., Dewachter, I. and Van Leuven, F. 1996. Expression in brain of amyl oid precursor protein mutate d in the alpha-secretase site causes disturbed beha vior, neuronal degenerati on and premature death in transgenic mice. EMBO J 15: 1265-1274. 227. Mohajeri, M., Madani, R., Saini, K., Lipp, H., Nitsch, R. and Wolfer, D. 2004. The impact of genetic background on neurodegenera tin and behavior in seizured mice. Genes Brain Behav 3: 228-239. 228. Monsonego, A., Maron, R., Zota, V., Sel koe, D. and Weiner, H. 2001. Immune hyporesponsiveness to amyloid -peptide in amyloid precursor protein transgenic mice: Implications for the pathogenesis a nd treatment of Alzheimers disease. Immunology 98: 10273-10278. 229. Montine, K., Quinn, J., Zhang, J., Fessel, J., Roberts, L., Morrow, J. and Montine, T. 2004. Isoprostanes and related pr oducts of lipid peroxidation in neurodegenerative diseases. Chem Phys Lipids 128: 117-124. 230. Moosmann, B. and Behl, C. 1999. The an tioxidant neuroprotective effects of estrogens and phenolic compounds are independent from their estrogenic properties. PNAS 96: 8867-8872. 231. Moosmann, B., Skutella, T., Beyer, K. and Behl, C. 2001. Protective activity of aromatic amines and imines agai nst oxidative nerve cell death. Biol Chem 382: 1601-1612. 232. Morgan, D., Diamond, D., Gottschall, P., Ug en, K., Dickey, C., Hardy, J., Duff, K., Jantzen, P., DiCarlo, G., Wilcock, D., C onnor, K., Hatcher, J., Hope, C., Gordon, M. and Arendash, G. 2000. A peptide vaccination prevents memory loss in an animal model of Alzheimers disease. Nature 408: 982-985.
292 233. Mori, C., Spooner, E., Wisniewsk, K., Wi sniewski, T., Yamaguch, H., Saido, T., Tolan, D., Selkow, D. and Lemere, C. 2002. Intraneuronal Abeta42 accumulation in Down syndrome brain. Amyloid 9: 88-102. 234. Moriguchi, T., Loewke, J., Garrison, M., Ca talan, J. and Salem, N. 2001. Reversal of docosahexaenoic acid deficiency in th e rat brain, retina, liver, and serum. Journal of Lipid Research 42: 419-427. 235. Morris, R. 1996a. A cognitive neuropyschology of Alzheimer-type dementia. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 3-10. Oxford: Oxford University Press. 236. Morris, R. 1996b. Attentional a nd executive dysfunction. Chap. In The Cognitive Neuropsychology of Alzh eimer-type Dementia. 49-70. Oxford: Oxford University Press. 237. Morris, R. 1996c. Neurobiological correla tes of cognitive dysfunction. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 223-254. Oxford: Oxford University Press. 238. Morris, M., Beckett, L., Scherr, P., Hebert L., Bennett, D., Field, T. and Evans, D. 1998. Vitamin E and vitamin C supplement use and risk of incident Alzheimer disease. Alzheimer Dis Assoc Disord 12: 121-126. 239. Morris, M., Evans, D., Bienias, J., Ta ngney, C., Bennett, D., Aggarwal, N., Wilson, R. and Scherr, P. 2002. Dietary intake of antioxidant nutrients and the risk of incident Alzheimer disease in a biracial community study. JAMA 287: 32303237. 240. Morris, M., Evans, D., Bienias, J., Tangney, C., Bennett, D., Aggarwal, N., Schneider, J. and Wilson, R. 2003. Diet ary Fats and the Risk of Incident Alzheier Disease. Archives of Neurology 60: 194-200. 241. Morris, M., Evans, D., Tangney, C., Bien ias, J. and Wilson, R. 2005. Fish Consumption and Cognitive Decline With Age in a Large Community Study. Archives of Neurology 62: 1-5. 242. Morrison and Hoff. 1997. Life and d eath of neurons in the aging brain. Science 278, 5337: 412-419. 243. Morrow, J., Harris, T. and Roberts II L. 1990. Noncyclooxygenase oxidative formation of a series of novel prostagl andins: analytical ramifications for measurement of eicosanoids. Anal Biochem 184: 1-10.
293 244. Mudher, A. and Lovestone, S. 2002. Alzhei mers diseasedo tauists and Baptists finally shake hands? TRENDS in Neurosciences 25, 1: 22-26. 245. Munoz, D. and Feldman, H. 2000. Causes of Alzheimers disease. CMAJ 162: 6572. 246. Nicoll, J., Wilkinson, D., Holmes, C., Steart, P., Markham, H. and Weller, R. 2003. Neuropathology of human Alzheimer dis ease after immunization with amyloidpeptide: a case report. Nature Medicine 9: 448-452. 247. Nilsson, L., Arendash, G., Leighty, R., Cost a, D., Low, M., Garcia, M., Cracchiolo, J., Rojiani, A., Wu, X., Bales, K., Pa ul, S. and Potter, H. 2004. Cognitive impairment in PDAPP mice depends on ApoE and ACT-catalyzed amyloid formation. Neurbiology of Aging 248. Nishimura, T., Takeda, M., Nakamura, Y., Yosbida, Y., Arai, H., Sasaki, H., Shouji, M., Hirai, S,. Khise, K., Ta naka, K., Hamamoto, M., Yamamoto, H., Matsubayashi, T., Urakami, K., Adachi, Y., Nakashima, K., Toji, H., Nakamura, S. and Yoshida, H. 1998. Basic and clin ical studies on the measurement of tau proteins in cerebrospinal fluid as a biol ogical marker for Alzheimers disease and related disorders: multicenter study in Japan. Methods Find Exp Clin Parmacol 20: 227-235. 249. Nourooz-Zadeh, J., Liu, E., Yhlen, B., nggrd, E. and Halliwell, B. 1999. F4Isoprostanes as specific marker of do cosahexaenoic acid peroxidation in Alzheimers disease. Journal of Neurochemistry 72: 734-740. 250. Ohm, T., Glckner, F., Distl, R., Treibe r-Held, S., Meske, V. and Schnheit, B. 2003. Plasticity and the Spread of Al zheimers Disease-Like Changes. Neurochemical Research 28, 11: 1715-1723. 251. Olin, J. and Schneider, L. 2002. Galantamine for Alzheimers disease (Cochrane Review) In: The Cochrane Review 252. Olsen, R. 1998. Discovery of the lipoproteins their role in fat transport and their significance as risk factors. American Society for the Nutritional Sciences 439S443S. 253. Ono, M., Wilson, A., Nobrega, J., Westaw ay, D., Verhoeff, P., Zhuang, Z., Kung, M. and Kung, H. 2003. 11C-labeled stilbene derivatives as Abeta-aggregatespecific PET imaging agents for Alzheimers disease Nucl Med Biol 30: 565571.
294 254. Patterson, C., Passmore, A. and Crawford, V. 2004. A 6-month open-label study of the effectiveness and tolerability of ga lantamine in patients with Alzheimers disease. Int J Clin Pract 58: 144-148. 255. Pappolla, M., Chyan, Y., Omar, R., Hsiao, K ., Perry, G., Smith, M. and Bozner, P. 1998. Evidence of oxidative stress and in Vivo neurotoxicity of -amyloid in a transgenic mouse model of Alzheimers disease. American Journal of Pathology 152: 871-877. 256. Payet, M., Wsmail, M., Polichetti, E., Brun, G., Adjemout, L., Donnarel, G., Portugal, H. and Pierni, G. 2004. Docosahexaenoic acid-enriched egg consumption induces accretion of arachidoni c acid in erythrocytes of elderly patients. British Journal of Nutrition 91: 789-796. 257. Petursdottir, D. and Olafsottir, I. and Hardardottir, I. 2002. Dietary fish oil increases tumor necrosis factor secre tion buy decreases inte rleukin-10 secretion by murine peritoneal macrophages. Nutritional Immunology 3740-3743. 258. Pfeifer, M., Boncristiano, S., Bondolfi, L., Stalder, A., Deller, T., Staufenbiel, M., Mathews, M. and Jucker, M. 2002. Cerebral Hemorrhage after passive antiA immunotherapy. Science 298: 1379. 259. Picciotto, M. and Wickman, K. 1998. Using Knockout and Transgenic Mice to Study Neurophysiology and Behavior. Physiological Reviews 78: 1131-1163. 260. Picq, M., Dubois, M, Grynberg, A., Lagard e, M. and Prigent, A. 1996. Specific effects of n-3 fatty acids and 8-br omo-cGMP on the cyclic nucleotide phosphodiesterase activity in ne onatal rat cardiac myocytes. J Mol Cell Cardiol 28: 2151-2161. 261. Poling, J., Vicini, S., Rogawski, M. and Salem, N. 1996. Docosahexaenoic acid block of neuronal voltage-gated K+ ch annels: subunit selective antagonism by zinc. Neuropharmacology 35: 969-982. 262. Potter, H., Wefes, I. and Nilsson, L. 2001. The inflammation-induced pathological chaperones ACT and apo-E are necessary catalysts of Alzheimer amyloid formation. Neurobiology of Aging 22: 932-930. 263. Prasad, M., Lovell, M., Yatin, M., Dhill on, H. and Markesbery, W. 1998. Regional membrane phospholipid alterations in Alzheimers disease. Neurochemical Research 23: 81-88. 264. Pratico, D., uryu, K., Leight, S., Trojanowsk i, J. and Lee, V. 2001. Increased lipid peroxidation precedes amyloid plaque forma tion in an animal model of Alzheimer amyloidosis. J Neurosci 21: 4183-4187.
295 265. Price, J., Ko, A., Wade, M., Tsou, S., McKeel, D. and Morris, J. 2001. Neuron number in the entorhinal cortex and CA 1 in preclinical Alzheimer disease. Arch Neurol 58: 1395-1402. 266. Puskas, L., Kitajka, K., Nyaka, C., Barcel o-Coblijn, G. and Farkas, T. 2003. Shortterm administration of omega-3 fatty acids from fish oil results in increased transthyretin transcriptio n in old rat hippocampus. PNAS 100: 1580-1585. 267. Puskas, L., Bereczki, E., Santha, M., Vigh, L., Csanadi, G., Spener, F., Ferdinandy, P., Onochy, A. and Kitajka, K. 2004. Cholesterol and cholesterol plus DHA dietinduced gene expression and fatty ac id changes in mouse eye and brain. Biochimie 86: 817-824. 268. Quayhagen, M. and Quayhagen, M. 1989. Differential effects of family-based strategies on Alzheimers disease. The Gerontologist 29: 150-155. 269. Raskind, M., Sadowsky, C., Sigmund, W., Beitle r, P. and Auster, S. 1997. Effect of tacrine on language, praxis, and noncognitive behavioral problems in Alzheimer disease. Arch Neurol 54: 836-840. 270. Raskind, M., Peskind, E., Wessel, T. and Yuan, W. 2000. Galantamine in AD: A 6month randomized, placebo-controlled tr ial with a 6-month extension. The Galantamine USA-1 Study Group. Neurology 54: 2261-2268. 271. Raskind, M., Peskind, E., Truyen, L., Kershaw, P. and Damaraju, C. 2004. The cognitive benefits of galantamine are sust ained for at least 36 months: a long-term extension trial. Arch Neurol 61: 252-256. 272. Redondo, C., Damas, A., Olofsson, A., Lundgren, E. and Saraiva, M. 2000. Search for intermediate structures in transthyre tin fibrillogenesis: soluble tetrameric Tyr78Phe TTR expresses a specific epitope present only in amyloid fibrils. J Mol Biol 304: 461-470. 273. Refolo, L., Pappolla, M., Malester, B., LaFrancois, J., Bryant-Thomas, T., Wang, R., Tint, G., Sambamurti, K. and Duff, K. 2000. Hypercholerolemia accelerates the Alzheimers amyloid pathology in a transgenic mouse model. Neurobiology of Disease 7: 321-331. 274. Reich, E., Markesbery, W., Roberts II, L., Swift, L., Morrow, J. and Montine, T. 2001. Brain regional quantif ication of F-ring and D/E-ring isoprostanes and neuroprostanes in Alzheimers disease. American Journal of Pathology 158: 293297.
296 275. Reitz, C., Tang, M., Luchsinger, J. and Ma yeux, R. 2004. Relation of plasma lipids to Alzheimer disease and vascular dementia. Archives of Neurology 61: 705-714. 276. Renier, G., Skamene, E., DeSanctis, J. and Radzioch, D. 1993. Dietary n-3 polyunsaturated acids prevent the development of atherosc lerotic lesions in mice. Modulation of macrophage secretory activities. Arterioscler Thromb 13: 15151524. 277. Robert II, L., Montine, T., Markesber y, W., Tapper, A., Hardy, P., Chemtob, S., Dettbarn, W. and Morrow, J. 1998. Formation of Isoprostane-like compounds (Neuroprostanes) in Vivo from docosahexaenoic acid. The Journal of Biological Chemistry 273: 13605-12612. 278. Rodondo, C., Damas, A. and Sariava, M. 2000. Designing transthryetin mutants affecting tetrameric structure: im plications in amyloidogenicity. Biochem J 348: 167-172. 279. Rogers, L., Kirby, L., Hempelman, S., Berry, D., McGeer, P., Kaszniak, A., Zalinski, J., Cofield, M., Mansukhani, L., W illson, P., et al. 1993. Clinical trial of indomethacin in Alzheimers disease. Neurology 43: 1609-1611. 280. Rodgers, R. and Johnson, N. 1995. Factor analysis of spatiotemporal and ethological measures in the murine elevated plus-maze test of anxiety. Pharmacol Biochem Behav 52: 297-303. 281. Rogers, E., Milhalik, S., Orthiz, D. and Shea, T. 2004. Apple juice prevents oxidative stress and impaired cognitiv e performance caused by genetic and dietary deficiencies in mice. Journal of Nutrition, Health and Aging 8: 92-97. 282. Roher, A., Baudry, J., Chaney, M., Kuo, Y., Stine, W. and Emmerling, M. 2000. Oligomerization and fibril assembly of the amyloidprotein. Biochimica et Biophysica Acta 1502: 31-43. 283. Rossler, M., Zarski, R., Bohl J. and Ohm, T. 2002. St age-dependent and sectorspecific neuronal loss in hippocampus Alzheimer disease. Acta Neuropathol 103: 363-369. 284. Salem, N., Litman, B., Kim, H. and Gawr isch, K. 2001. Mechanisms of action of docosahexaenoic acid in the nervous system. Lipids 36: 945-959. 285. Salmon, D and Fennema-Notestine, C. 1996. Implicit memory. Chap. In The Cognitive Neuropsychology of Alzheimer-type Dementia. 105-127. Oxford: Oxford University Press.
297 286. Sanan, D., Weisgraber, K., Russell, S ., Mahley, R., Huang, D., Saunders, A., Schmechel, D., Wisniewski, T., Frangione, B., Roses, A. and Strittmatter. 1994. Apolipoprotein E associates with beta am yloid peptide of Alzheimers disease to form novel monofibrils. Isoform apoE4 asso ciates more efficiently than apoE3. J Clin Invest 94: 860-869. 287. Sano, M., Ernesto, C., Thomas, R., Kla uber, M., Schafer, K., Grundman, M., Woodbury, P., Growdon, J., Cotman, C., Pf eiffer, E., Schneider, L., Thal, L. 1997. A controlled trial of selegiline, al pha-tocopherol, or both as treatment for Alzheimers disease. The Alzheimers Disease Cooperative Study. N Engl J Med 24: 1216-1222. 288. Sarsilmaz, M., Songur, A., zyurt, H., Kus, ., zen, O., zyurt, B., S t, S. and Akyol, . 2003. Potential role of dietary -3 essential fatty acids on some oxidant/antioxidant parameters in rats corpus striatum. Prostaglandins, Leukotrienes and Essential Fatty Acids 69: 253-259. 289. Sayre, L., Zagorski, M., Surewicz, W., Kr afft, G. and Perry, G. 1997. Mechansims of neurotoxicity associated with amyloid b deposition and the role of free radicals in the pathogenesis of Alzheimers disease: a critical appraisal. Chem Res Toxicol 10: 518-526. 290. Scarpini, E., Scheltens, P. and Feldman, H. 2003. Treatment of Alzheimers disease: current status and new prospectives. Lancet Neurol 2: 539-47. 291. Scheff, S. and Price, D. 2003. Synaptic pa thology in Alzheimers disease: a review of ultrastructural studies Neurobiology of Aging 24: 1029-1046. 292. Schenk, D., Barbour, R., Dunn, W., Gordon, G., Grajeda, H., Guido, T., Hu, K., Huang, J., Johnson-Wood, K., Khan, K., Kholodenko, D., lee, M., Liao, Z., Lieberburg, I., Motter, R., Mutter, L. Soriano, F., Shopp, G., Vasquez, N., Vandevert, C., Walker, S., Wogulis, M., Yednock, T., Games, D. and Seubert, P. 1999. Immunization with amyloidattenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 4000: 173-177. 293. Schmitz, G. and Langmann, T. 2005. Transcriptional regulatory networks in lipid metabolism control ABCA1 expression. Biochimica et Biophysica Acta 1735: 119. 294. Schneider, L and Dagerman, K. 2004. Psychos is of Alzheimers disease: clinical characteristics and history. Journal of Psychiatric Research 38: 105-111. 295. Schwarzman, A. and Goldgaber, D. 1996. In teraction of transthyretin with amyloid beta-protein: binding and inhibi tion of amyloid formation. Ciba Found Symp 199: 146-160.
298 296. Selhub, J. and Miller, J. 1992. The pathoge nesis of homocysteinemia: interruption of the coordinate regulation by S-adenos ylmethionine of the remethylation and trans-sulfuration of homocysteine. Am J Clin Nutr 55: 131-138. 297. Selkoe, D. 2001. Alzheimers Disease: Genes, Proteins, and Therapy. Physiological Reviews 81, 2: 741-766. 298. Serpell, L. 2000. Alzheimers amyloid fibrils: structur e and assembly. Biochimica et Biophysica Acta 1502: 16-30. 299. Shukitt-Hale, B., Smith, D., Meydani, M and Joseph, J. 1999. The effects of dietary antioxidants on psychomotor pe rformance in aged mice. Experimental Gerontology 34: 797-808. 300. Sigurdsson, E., Scholtzova, H., Mehta, P ., Frangione, B. and Wisniewski, T. 2001. Immunization with nontoxic/nonfibrill ar amyloid-beta homologous peptide reduces Alzheimers disease-associated pathology in transgenic mice. Am J Pathol 159: 439-447. 301. Sigurdsson, E., Knudsen, E., Asuni, A., Fit zer-Attas, C., Sage, D., Quartermain, D., Goni, F., Frangione, B. and Wisniews ki, T. 2004. An attenuated immune response is sufficient to enhance cognition in an Alzheimers disease mouse model immunized with amyloid-b derivatives. Neurobiology of Disease 24: 6277-6282. 302. Silbert, L., Quinn, J., Moore, M., Corbri dge, E., Ball, M., Murdoch, G., Sexton, G. and Kaye, J. 2003. Changes in premor bid brain volume predict Alzheimers disease pathology. Neurology 61: 487-492. 303. Simons, M., Keller, P., De Strooper, B., Be yreuther, K., Dotti, C. and Simons, K. 1998. Cholesterol depletion inhibits the generation of b-amyloid in hippocampal neurons. PNAS 95: 6460-6464. 304. Skinner, E., Watt, C., Besson, J. and Best, P. 1993. Differences in the fatty acid composition of the grey and white matter of different regions of the brains of patients with Alzheimers dise ase and control subjects. Brain 116: 717-725. 305. Smith, M., Hirai, K., Hsiao, K., Pappolla, M., Harris, P., Siedlak, S., Tabaton, M. and Perry, G. 1998. J Neurochem 2212-2215. 306. Sobel, B. 2001. Bingo vs. physical interven tion in stimulating short-term cognition in Alzheimers disease patients. Am J Alzheimers Dis Other Demen 16: 115-120.
299 307. Solfrizzi, V., Panza, F. and Capurso, A. 2003. The role of diet in cognitive decline. Journal of Neural Transmission 110: 95-110. 308. Solomon, B., Koppel, R., Hanan, E. and Katzav, T. 1996. Monoclonal antibodies inhibit in vitro fibrillar aggregati on of the Alzheimer -amyloid peptide. PNAS 93: 452-455. 309. Solomon, B., Koppel, R., Frankel, D. and Hanan-Aharon, E. 1997. Disaggregation of Alzheimer -amyloid by site directed mAb. PNAS 94: 4109-4112. 310. Souren, L., Franssen, E. and Reisberg, B. 1995. Contractures and loss of function in patients with Alzheimers disease. J Am Geriatr Soc 43: 650-655. 311. Stein, T., Anders, N., DeCarli, C., Ch an, S., Mattson, M. and Johnson, J. 2004. Neutralization of transthyretin reverses the neuroprotective e ffects of secreted amyloid precurosor protein (APP) in APPSW mice resulting in tau phosphorylation and loss of hippocampa l neurons: support for the amyloid hypothesis. J Neurosci 24: 7707-7717. 312. Stern, L., Iqbal, N., Seshadri, P., Chica no, K., Daily, D., McGrory, J., Williams, M., Gracely, E. and Samaha, F. 2004. The effects of low-carbohydrate versus conventional weight loss diets in severe ly obese adults: Oneyear follow-up of a randomized trial. Annals of Internal Medicine 140: 778-786. 313. Stewart, W., Kawas, C., Corrada, M. a nd Metter, E. 1997. Risk of Alzheimers disease and durati on of NSAID use. Neurology 48: 626-632. 314. Stillwell, W. and Wassall, S. 2003. Docosa hexaenoic acid: membrane properties of a unique fatty acid. Chemistry and Physics of Lipids 126: 1-27. 315. Stout, J., Wyman, M., Johnson, S., Pea vy, G. and Salmon, D. 2003. Frontal behavioral syndromes and functional stat us in probable Alzheimer disease. Am J Geriatr Psychiatry 11: 683-686. 316. Strittmatter, W. and Roses, A. 1995. A polipoprotein E and Alzheimer disease. PNAS 92: 4725-4727. 317. Strokin, M., Sergeeva, M. and Reiser G. 2003. Docosahexaenoic acid and arachidonic acid release in rat brain astr ocytes is mediated by two separate isoforms of phospholipase A2 and is di fferently regulated by cyclic AMP and Ca2+. Br J Pharmacol 139: 1014-1022. 318. Sturchler-Pierrat, C., Abramowski, D., Duke, M., Wiederhold, K., Mistl, C., Rothacher, S., Ledermann, B., Brki, K., Frey, P., Paganetti, P., Waridel, C., Calhoun, M., Jucker, M., Probst, A., Staufe nbiel, M. and Sommer, B. 1997. Two
300 amyloid precursor protein transgenic mous e models with Alzheimer disease-like pathology. Proc. Natl. Acad. Sci USA 94: 13287-13292. 319. Suemoto, T., Okamura, N., Shiomitsu, T., Suzuki, M., Shimadzu, H., Akatsu, H., Yamamoto, T., Kudo, Y. and Sawada, T. 2004. In vivo labeling of amyloid with BF-108. Neurosci Res 48: 65-74. 320. Sugimoto, Y., Taga, C., Nishiga, M., Fujiw ara, M., Konishi, F., Tanaka, K. and Kamei, C. 2002. Effect of docosahexaenoic acid-fortified Chlorella vulgaris strain CK22 on the radial m aze performance in aged mice Biol. Pharm. Bull. 25: 1090-1092. 321. Sun, Y., Minthon, L., Wallmark, A., Warkenti n, S., Blennow, K. and Janciauskiene, S. 2003. Inflammatory markers in matched plasma and cerebrospinal fluid from patients with Alzhiemers disease. Dement Geriatr Cogn Disord 16: 136-144. 322. Suzuki, H., Park, S., Tamura, M. and A ndo, S. 1998. Effect of the long-term feeding of dietary lipids on the learning ability, fatty acid composition of brain stem phospholipids and synaptic membrane fluidity in adult mice: a comparison of sardine oil diet and palm oil diet. Mechanisms of Ageing and Development 101: 119-128. 323. Takao, M., Ghetti, B., Hayakawa, I., Ik eda, E., Fukuuchi, Y., Miravalle, L., Piccardo, P., Murrell, J., Glazier, B. and Koto, A. 2002. A novel mutation (G217D) in the presenilin 1 gene (PSE N1) in a Japanese family: presenile dementia and parkinsonism are associated with cotton wool plaques in the cortex and striatum. Acta Neuropathol 104: 155-170. 324. Takeuchi, A., Irizarry, M., Duff, K., Saido, T., Hsiao Ashe, K., Hasegawa, M., Mann, D., Hyman, B. and Iwatsubo, T. 2000. Age-related amyloid deposition in transgenic mice overexpressing both Alzheimer mutant presenilin 1 and amyloid precursor protein Swedish mutant is not associated with global neuronal loss. American Journal of Pathology 157: 331-339. 325. Tales, A., Muir, J., Jones, R., Bayer, A. and Snowden, R. 2004. The effects of saliency and task difficulty on visual search performance in ageing and Alzheimers disease. Neuropsychologia 42: 335-45. 326. Tariot, P., Solomon, P., Morris, J., Kershaw, P., Lilienfeld, S. and Ding, C. 2000. A 5-month, randomized, placebo-controlled trial of galantamine in AD. The Galantamine USA-10 Study Group. Neurology 54: 2269-2276. 327. Tariot, P., Cummings, J., Katz, I., Mintzer, J., Perdomo, C., Schwam, E. and Whalen, E. 2001. A randomized, double-b lind, placebo-controlled study of the
301 efficacy and safety of donepezil in patie nts with Alzheimers disease in the nursing home setting. J Am Geriatr Soc 49: 1590-1599. 328. Terry, R., Hansen, L., DeTeresa, R., Davies, P., Tobias, H. and Katzman, R. Senile dementia of the Alzheimer type without neocortical neurofibrillary tangles. J Neuropath Exp Neurol 46: 262-268. 329. Thaker, U., McDonagh, A., Iwatsubo, T., Lendon, C., Pickering-Brown, S. and Mann, D. 2003. Neuropathol Appl Neurobiol 29: 35-44. 330. Thomas, R., Thomas, G., McLendon, C., Sutton, T. and Mullan, M. 1996. amyloid-mediated vasoactivity and vascular endothelial damage. Nature 380: 168-171. 331. Tomobe, Y., Morizawa, K., Tsuchida, M., Hibino, H., Nakano, Y. and Tanaka, Y. 2000. Dietary docosahexaenoic acid suppresses inflammatory and immunoresponses in contat hypersensitivity reaction in mice. Lipids 35: 61-69. 332. Thompson, P., Hayashi, K., de Zubicaray, G., Janke, A., Rose, S., Semple, J., Herman, D., Hong, M., Dittmer, S., Doddrell, D. and Toga, A. 2003. Dynamics of gray matter loss in Alzheimers disease. J Neurosci 23: 994-1005. 333. Trillo, L. and Gonzalo, L. 1992. Ageing of the human entorh inal cortex and subicular complex. Histol Histopathol 7: 17-22. 334. Tsukada, H., Kakiuchi, T., Fukumoto, D ., Nishiyama, S. and Koga, K. 2000. Docosahexaenoic acid (DHA) improves the age-related impairment of the coupling mechanism between neuronal ac tivation and functional cerebral blood flow response: a PET study in conscious monkeys. Brain Research 862: 180-186. 335. Tully, A., Roche, H., Doyle, R., Fallon, C., Bruce, I., Lawlor, B., Coakley, D. and Gibney, M. 2003. Low serum cholestyl es ter-docosahexaenoic acid levels in Alzheimers disease: a case-control study. British Journal of Nutrition 89: 483489. 336. Turner, N., Else, P. and Hulbert, A. 2003a. Docosahexaenoic acid (DHA) content of membranes determines molecular activity of the sodium pump: implications for the disease states and metabolism. Naturwissenschaften 90: 521-523. 337. Turner, P., OConnor, K., Tate, W. and Abraham, W. 2003b. Roles of amyloid precursor protein and its fragements in re gulating neural activ ity, plasticity and memory. Progress in Neurobiology 70: 1-32. 338. Ulmann, L., Mimouni, V., Roux, Porsolt, R. and Poisson, J. 2001. Brain and hippocampus fatty acid composition in phospholipid classes of aged-relative
302 cognitive deficit rats. Prostaglandins, Leukotrienes and Essential Fatty Acids 64: 189-195. 339. Van Dam, D., DHooge, R., Staufenbiel, M., Van Ginneken, C., Van Meir, F. and De Deyn, P. 2003. Age-dependent cognitive decline in the APP23 model precedes amyloid deposition. Eur J Neurosci 17: 388-396. 340. van Dongen, M., van Rossum, E., Kessels, A., Sielhorst, H. and Knipschild, P. 2000. The efficacy of ginkgo for elderly peop le with dementia and age-assocaited memory impairment: new results of a randomized clinical trial. J Am Geriatr Soc 48: 1183-1194. 341. van Dongen, M., van Rossum, E., Kessels, A., Sielhorst, H. and Knipschild, P. 2003. Ginkgo for elderly people with dementia and age-associated memory impairment: a randomized clinical trial. Journal of Clinical Epidemiology 56: 367-376. 342. van Gool, W., Weinstein, H., Scheltens, P ., Walstra, G., Scheltens, P. 2001. Effect of hydroxychloroquine on progression of de mentia in early Alzheimers disease: an 18-month randomized, double-blind, placebo-controlled study. Lancet 358: 455-460. 343. Veerkamp, J. and Zimmerman, A. 2001. Fatty acid-binding proteins of nervous tissue. Journal of Molecular Neuroscience 16: 133-142. 344. Velez-Pardo, C., Arellano, J., Cardona-Gom ez, P., Jimenez Del Rio, M., Lopera, F. and De Felipe, J. 2004. CA1 hippocampal neuronal loss in familial Alzheimers disease Presenilin-1 E280A mutation is related to epilepsy. Epilepsia 45: 751-756. 345. Verkkoniemi, A., Ylikoski, R., Rinne, J., So mer, M., Kietaharju, A., Eirkinjuntti, T., Viitanen, M., Kalimo, H. and Haltia, M. 2004. Neuropsychological functions in variant Alzheimers disease with spastic paraparesis. Journal of the Neurological Sciences 218: 29-37. 346. Vigo-Pelfrey, C, Seubert, P., Barbour, R., Bl omquist, C., Lee, M., Lee, D., Coria, F., Chang, L., Miller, B., Liebergurg, I ., et al. 1995. Elevation of microtubuleassociated protein tau in the cerebrospinal fluid of patients with Alzheimers disease. Neurology 45: 788-793. 347. Vinores, S., Xiao, W., Zi mmerman, R., Whitcup, S. and Wawrousek, E. 2003. Upregulation of vascular endothelial grow th factor (VEGF) in the retinas of transgenic mice overexpressing interleukin1 beta (IL-1beta) in the lens and mice undergoing retinal degeneration. Histol Histopathol 18: 797-810.
303 348. Voikar, V., Koks, S., Vasar, E. and Rauvala, H. 2001. Strain and gender differences in the behavior of mouse lines commonly used in transgenic studies. Physiol Behav 72: 271-281. 349. Volicer, L., Berman, S., Cipolloni, P. and Mandell, A. 1997. Persistent vegetative state in Alzheimers disease. Does it exist? Arch Neurol 54: 1382-1384. 350. Vreugdenhil, M., Bruehl, C., Voskuyl, R ., Kang, J., Leaf, A. and Wadman, W. 1996. Polyunsaturated fatty acids modulat e sodium and calcium currents in CA1 neurons. PNAS 93: 12559-12563. 351. Wainwright, P., Xing, H., Ward, G., Huang, Y., Bobik, E., Auestad, N and Montalto, M. 1999. Water maze performance is unaffected in artificially reared rats fed diets supplemented with arac hidonic acid and docosahexaenoic acid. Nutritional Neurosciences 1079-1089. 352. Walhund, L., Basun, H., Almkvist, O., Julin, P., Axelman, K., Shigeta, M., Jelic, V., Nordberg, A. and Lannfelt, L. 1998. A follow-up study of the family with the Swedish APP 670/671 Alzheimers disease mutation. Dementia and Geriatric Cognitive Disorders 10: 526-533. 353. Wahlund, L. and Blennow, K. 2003. Cerebr ospinal fluid biomarkers for disease stage and intensity in cognitively impaired patients. Neurosci Lett 20: 99-102. 354. Wallace, F., Miles, E. and Calder, P. 2003. Comparison of the effects of linseed oil and different doses of fish oil on mon onuclear cell function in healthy human subjects. British Journal of Nutrition 89: 679-689. 355. Wahrle, S., Das, P., Nyborg, A., McLendon, C., Shoji, M., Kawarabayashi, T., Younkin, L., Younkin, S. and Golde, T. C holesterol-dependent gamma-secretase activity in buoyant c holesterol-rich membrane microdomains. Neurobiol Dis 9: 11-23. 356. Wang, H., Hung, T., Wei, J. and Chiang, A. 2004. Fish oil increases antioxidant enzyme activities in macrophages and redu ces atherosclerotic lesions in apoEknockout mice. Cardiovascular Research 61: 169-176. 357. Watanabe, S., Onozaki, K., Yamamoto, S. and Okuyama, H. 1993. Regulation by dietary essential fatty acid balance of tu mor necrosis factor production in mouse macrophages. J Leukoc Biol 53: 151-156. 358. Weggen, S., Eriksen, J., Das, P., Sagi, S., Wang, R., Peitrzik, C., Findlay, K., Smith, T., Murphy, M., Bulter, T., Kang, D ., Marquez-Sterling, N., Golde, T. and Koo, E. 2001. A subset of NSAI Ds lower amyloidogenic A[beta]42 independently of cyclooxygenase activity. Nature 414: 212-216.
304 359. Weggen, S., Eriksen, J., Sagi, S., Pietrzik, C., Ozols, V., Faud, A., Golde, T. and Koo, E. 2003. Evidence that nonsterioda l anti-inflammatory drugs decrease amyloid beta 42 production by direct modul ation of gamma-secretase activity. J Biol Chem 278: 31831-31837. 360. Weiner, H., Lemere, C., Maron, R., Spooner, E., Grenfell, T., Mori, C., Issazadeh, S., Hancock, W. and Selkoe, D. 2000. Nasal administration of amyloidpeptide decreases cerebral amyloid burden in a mouse model of Alzheimers disease. Annals of Neurology 48: 567-579. 361. Westerman, M., Cooper-Blacketer, D., Mari ash, A., Kotilinek, L., Kawarabayashi, T., Younkin, L., Carlson, G., Younkin, S. and Ashe, K. 2002. The relationship between Abeta and memory in the Tg2576 mouse model of Alzheimers disease. J Neurosci 22: 1858-1567. 362. Wietrzych, M., Meziane, H., Sutter, A., Ghyselinck, N., Chapman, P., Chambon, P., Krezel, W. 2005. Working memory defi cits in retinoid X receptor gammadeficient mice Learning and Memory 12: 318-326. 363. Wilcock, D., Gordon, M., Ugen, K. Gottschall, P., DiCarlo, G., Dickey, C., Boyett, D., Jantzen, P., Connor, K., Melachrino, J., Hardy, J. and Morgan, D. 2001. Number of A inoculations in APP+PS1 transgenic mice influences antibody titers, microglial activation, a nd congophilic plaque levels. DNA and Cell Biology 20: 731-736. 364. Wilcock, G., Howe, I., Coles, H., Lilienfeld, S., Truyen, L., Zhu, Y., Bullock, R., Kershaw, P., GAL-GBR-2 Study Group. 2003. A long-term comparison of galantamine and donepezil in the trea tment of Alzheimers disease. Drugs Aging 20: 777-789. 365. Wilcock, D., Rojiani, A., Rosenthal, A ., Levkowitz, G., Subbarno, S., Alamed, J., Wilson, D., Wilson, N., Freeman, M., Gor don, M. and Morgan, D. 2004. Passive amyloid immunotherapy clears amyloid a nd transiently activat es microglia in a transgenic mouse model of amyloid deposition. J Neurosci 24: 6144-6151. 366. Wilkinson, D., Passmore, A., Bullock, R ., Hopker, S., Smith, R., Potocnik, F., Maud, C., Engelbrecht, I., Hock, C., Ieni, J. and Bahra, R. 2002. A multinational, randomized, 12-week, comparative study of donepezil and rivastigmine in patients with mild to moderate Alzheimers disease. Int J Clin Pract 56: 441-446. 367. Williard, D., Harmon, S., Kaduce, T., Preuss, M., Moore, S., Robbins, M. and Spector, A. 2001. Docosahexaenoic aci d synthesis from n-3 polyunsaturated fatty acids in differentiate d rat brain astrocytes. Journal of Lipid Research 42: 1368-1376.
305 368. Winblad, B., Engedal, K., Soininen, H ., verhey, F., Waldemar, G., Wimo, A., Wetterholm, A., Zhang, R., Haglund, A., Subbiah, P. and the Donepezil Nordic Study Group. 2001. A 1-year, randomized, placebo-controlled study of donepezil in patients with mild to moderate AD. Neurology 57: 489-495. 369. Winocur, G. and Greenwood, C. 2005. Studies of the effects of high fat diets on cognitive function in a rat model. Neurobiol Aging 26: 46-49. 370. Wisniewski, T., Castano, E., Golabek, A., Vogel, T. and Frangione, B. 1994. Acceleration of Alzheimers fibril form ation by apolipoprotein E in vitro. Am J Path 145: 1030-1035. 371. Wolozin, B. 2002. Cholesterol and Alzheimers disease. Biochemical Society Transactions 30: 525-29. 372. Xiao, Y., Kang, J., Morgan, J. and Leaf, A. 1995. Blocking effects of polyunsaturated fatty acids on Na+ channels of neonatal rat ventricular myocytes. PNAS 92: 11000-11004. 373. Yan, S., Stern, D., Kane, M., Kuo, Y., Lamp ert, H. and Roher, A. 1998. RAGE-AB interactions in the pathophysiology of Alzheimers disease. Restor Neurol Neurosci 12: 167-173 374. Yancy, W., Olsen, M., Guyton, J., Baks t, R. and Westman, E. 2004. A lowcarbohydrate, ketogenic diet versus a low-fat diet to treat obesity and hyperlipidemia: A randomized, controlled trial. Annals of Internal Medicine 140: 769-779. 375. Yehuda, S., Rabinovtz, S., Carasso, R. a nd Mostofsky, D. 1996. Essential fatty acids preparation (SR-3) improves Alzh eimers patients quality of life. Intern. J. Neuroscience 87: 141-149. 376. Yehuda, S., Rabinovitz, S., Carasso, R. a nd Mostofsky, D. 2002. The role of polyunsaturated fatty acids in restor ing the aging neuronal membrane. Neurobiology of Aging 23: 843-853. 377. Young, C., Gean, P. Chiou, L. and Shen, Y. 2000. Docosahexaenoic acid inhibits synaptic transmission and epileptifor m activity in the rat hippocampus. Synapse 37: 90-94. 378. Zamrini, E., McGwin, G. and Roseman, J. 2004. Association between statin use and Alzheimers disease. Neuroepidemiology 23: 94-98.
306 379. Zandi, P., Anthony, J., Khachaturian, A., St one, S., Gustafson, D., Tschanz, J., Norton, M., Welsh-Bohmer, K., Breitn er, J. and Cache County Study Group. 2004. Reduced risk of Alzheimer diseas e in users of antioxidant vitamin supplements: the Cache County Study. Arch Neurol 61: 82-88. 380. Zekanowski, C., Styczynska, M., Peplons ka, B., Gabryelewicz, T., Religa, D., Ilkowski, J., Kijanowska-Haladyna, B., Kota pka-Minc, S., Mikkelsen, S., Pfeffer, A., Barczak, A., Luczywek, E., Wasia k, B., Chodakowska-Zebrowska, M., Gustaw, K., Laczkowski, J., Sobow, T., Ku znicki, J. and Barcikowska, M. 2003. Mutations in presenilin 1, presenilin 2 and amyloid precursor protein genes in patients with early-onset Alzh eimers disease in Poland. Exp Neurol 184: 991996. 381. erovnik, E. 2002. Amyloid-fibril formation: Proposed mechanisms and relevance to conformational disease. Eur. J. Biochem. 269: 3362-3371. 382. Zubenko, G., Zubenko, W., McPherson, S., Spoor, E., Marin, D., Farlow, M., Smith, G., Geda, Y., Cummings, F., Petersen, R. and Sunderland, T. 2003. A collaborative study of the emergence and clin ical features of the major depressive syndrome of Alzheimers disease. American Journal of Psychiatry 160: 857-866.
About the Author Maren Jensen received her Bachelors Degree in Biology from the University of North Florida in 2000. She entered the Ph.D. program in Biology at the University of South Florida in 2002. Her work in the fiel d has been highlighted in a co-authored publication as well as by a platform presentati on at the Society for Neuroscience meeting. While in the Ph.D. program, Ms. Jensen worked as a researcher fo r the Johnnie B. Byrd Alzheimers Center and Research Institute.
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001788099
007 cr mnu|||uuuuu
008 070516s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001438
Jensen, Maren T.
Effects of A-beta immunotherapy and Omega-3 fatty acid administration in Alzheimer's transgenic mice
h [electronic resource] /
by Maren T. Jensen.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: Major therapeutics against Alzheimer's disease (AD) are targeted towards reducing beta-amyloid in the brain and improving cognitive performance. Transgenic mouse models of AD have become critical in the development of such therapeutics to protect against or treat AD. This dissertation examined the potential protective effects of both active A-beta immunotherapy and dietary omega-3 fatty acid administration to AD transgenic mice. First, immunization with A-beta 1-42 from 2-16 1/2 months of age provided protection against cognitive impairment in APP/PS1 transgenic mice well into older age. At both adult (4 1/2-6 month) and aged (15-16 1/2 month) test points, an extensive 6-week behavioral battery was administered that measured multiple sensorimotor and cognitive domains. A-beta immunotherapy either partially or completely protected APP/PS1 mice from impairment in reference learning/memory, working memory and/or recognition/identification at these test points. However,^ behavioral protection at the later test point occurred without any reduction in A-beta deposition within the brain. Therefore, the cognitive benefits of A-beta immunotherapy most likely involved neutralization or removal of A-beta oligomers from the brain. In addition to immunotherapy, this dissertation also examined the behavioral and neurochemical effects of a high omega-3 (n-3) or high omega-6 fatty acid (n-6) diet to NT and AD transgenic (Tg+) mice from 2 through 9 months of age. The same 6-week behavioral test battery, as described above, was administered between 7 1/2-9 months of age. In NT mice, dietary n-3 or n-6 fatty acids did not result in any beneficial effects on cognitive performance. In Tg+ mice, a high n-3 diet improved some, but not most, cognitive skills in comparison to standard-fed Tg+ mice; whereas a diet high in n-6 fatty acids did not lead to widespread deficits in learning or memory. In fact, there was no difference in overall performance on any behavioral ^task between Tg+ mice given a high n-3 or high n-6 diet. Administration of dietary fatty acids did not result in any significant changes in soluble or insoluble A-beta levels within the brains of Tg+ mice and plasma cytokine levels in Tg+ mice were largely unaffected. Notably, neither the high n-3 nor high n-6 diet increased cortical levels of n-3 or n-6 fatty acids, respectively, within Tg+ mice. However, NT mice on a high n-3 or high n-6 diet did show significant elevations in cortical n-3 or n-6 fatty acid levels, respectively, suggesting that Tg+ mice have a deficit in incorporation of dietary fatty acids in the brain. Collectively, these results show that life-long administration of active A-beta immunotherapy provides clear cognitive protection well into older age, whereas long-term dietary omega-3 fatty acid administration does not provide extensive cognitive benefit. Both studies underscore the value of using AD transgenic mice in determining the efficacy of prophylactics ^against AD.
Dissertation (Ph.D.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic dissertation) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 306 pages.
Adviser: Gary Arendash, Ph.D.
t USF Electronic Theses and Dissertations.