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 001912383
007 cr mnu|||uuuuu
008 071008s2005 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001264
Effects of adoptive transfer of beta-amyloid sensitive immune cells in a mouse model for Alzheimer's disease
h [electronic resource] /
by Daniel Shippy.
[Tampa, Fla] :
b University of South Florida,
ABSTRACT: One major therapeutic target for preventing and treating Alzheimer's Disease (AD) is removal of excess beta-amyloid from the brain. Both active and passive immunotherapies targeting beta-amyloid have proven effective in reducing brain beta-amyloid levels and improving cognitive function in mouse transgenic models of AD. However, these approaches can induce adverse neuropathologic effects and immunologic over-activation. Indeed, clinical trials of active beta-amyloid immunotherapy in AD patients were halted due to development of meningoencephalitis, apparently resulting from wide-spread neuroinflammation. Here we show that a more restricted and specific immune re-activation through a single adoptive transfer of beta-amyloid-specific T cells can provide long-term benefits to APPsw+PS1 transgenic mice that last at least 1 1/2 months. beta-amyloid-sensitive splenocytes and lymphocytes were generated in normal mice, re-stimulated with beta-amyloid in vitro, and then adoptive ly transferred into cognitively-impaired APPsw+PS1 mice. Compared to control transgenic mice through 1 1/2 month post-infusion, those mice that received beta-amyloid-sensitive T cells exhibited a reversal of pre-infusion working memory impairment and demonstrated superior basic mnemonic processing. Step-wise forward Discriminant Function Analysis of behavioral results clearly demonstrated that T cell infused mice performed comparably to wild-type non-transgenics, further emphasizing the extent of cognitive benefit this therapeutic technique afforded. Importantly, a global inflammatory response did not accompany these benefits. Though no overall reductions in beta-amyloid deposition were noted for T cell recipient mice, a subset of T cell infused mice that benefited most in cognitive function had reduced hippocampal burdens, suggesting that hippocampal beta-amyloid burdes did play a role in determining performance capabilities of these mice. Since chronically high levels of beta-amy loid such as those found in APPsw+PS1 mice cause immune hypo-responsive/tolerance to beta-amyloid, our results indicate that adoptive transfer of beta-amyloid-sensitive T-cells can supercede such immune tolerance to beta-amyloid, and may provide a safe, long-lasting therapy for AD.
Thesis (M.S.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) 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 180 pages.
Adviser: Gary Arendash, Ph.D.
t USF Electronic Theses and Dissertations.
Effects of Adoptive Transfer of Beta-Amyloid Sensitive Immune Cells in a Mouse Model for Alzheimers Disease by Daniel Shippy A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Biology College of Arts and Sciences University of South Florida Major Professor: Gary Arendash, Ph.D. Huntington Potter, Ph.D. My Lien Dao, Ph.D. Date of Approval: June 8, 2005 Keywords: Immunotherapy, T cell, Behavior, Transgenic Mouse, Neurodegenerative Diseases Copyright 2005, Daniel Shippy
i Table of Contents List of Tables................................................................................................................. ....iv List of Figures.....................................................................................................................v Alzheimers Disease...........................................................................................................8 Alzheimers Disease Background...........................................................................8 Behavioral Character ization of AD........................................................................9 Pathological Characterization of AD....................................................................12 Genetics of AD.....................................................................................................19 Diagnosis of AD...................................................................................................21 Risk Factors for AD..............................................................................................24 Current Treatments of AD....................................................................................26 Animal Models of Alzheimers Disease...........................................................................29 PDAPP Model.......................................................................................................30 Pathological Characterizati on of the PDAPP Model................................30 Behavioral Characterizati on of the PDAPP Model..................................31 APPsw (APP23) Model........................................................................................32 Pathological Characterizati on of the Tg2576 Model................................33 Pathological Characterizati on of the APP23 Model.................................34 Behavioral Characterizati on of the Tg2576 Model..................................34 Behavioral Characterizati on of the APP23 Model....................................36 APPsw+PS1 Model...............................................................................................38 Pathological Characterization of the APPsw+PS1 Model........................38 Behavioral Characterization of the APPsw+PS1 Model..........................41 Mutant APP Mice as Incomplete Models of AD..................................................44 Inflammatory Responses in AD........................................................................................48 Microglia Activation.............................................................................................49 T Cell Activation...................................................................................................52 Microglia and T Cell Co-activation......................................................................54 Vaccination in AD Research.............................................................................................58 In Vitro A Vaccination Studies...........................................................................58 Innate Immune Responses to A ..........................................................................59 Active A Immunization in Animal Models........................................................60 Pathological Benefits of Active A Immunization...................................60
ii Deleterious Effects of Active A Immunization and Alternative Active Immunization Approaches........................................................................67 Behavioral Character ization in Active A Immunotherapy.....................72 Active A Immunization in Humans....................................................................79 Passive Anti-A Immunization in Animal Models..............................................82 Pathologic Effects of Passive Anti-A Immunization..............................82 Behavioral Effects of Passive Anti-A Immunization.............................86 Passive Anti-A Immunization in Humans..........................................................90 Other Immunotherapeutic Strategi es for Alzheimers Disease........................................92 Alternative Peripheral A Sinks...........................................................................93 T-cell Based Therapy............................................................................................94 Donor Lymphocyte Infusion For Non-CNS Diseases..............................95 Priming of the CNS Immune System With T Lymphocytes....................96 Neuroprotective Actions of T Lymphocytes.............................................99 Specific Aims..................................................................................................................107 Materials and Methods....................................................................................................109 Animals...............................................................................................................109 General Protocol.................................................................................................109 Immune Cell Infusion Protocol...........................................................................112 Behavioral Assessment.......................................................................................113 Radial Arm Water Maze.........................................................................113 Platform Recognition..............................................................................114 Y-maze....................................................................................................115 Neurological Battery...............................................................................115 Brain Collection and Sectioning.........................................................................116 Plasma Cytokine Level Measurements...............................................................116 Deposition Determinations...........................................................................117 6E10 Immunostaining.............................................................................117 Thioflavin S Staining..............................................................................118 Image Analysis........................................................................................119 Statistical Analysis..............................................................................................119 Behavioral Statistical Analysis...............................................................119 Pathological/Histochemical Statistical Analysis.....................................120 Correlation Analysis...............................................................................120 Factor Analysis.......................................................................................120 Discriminant Function Analysis.............................................................121 Results........................................................................................................................ .....122 Behavior..............................................................................................................122 RAWM....................................................................................................122 Y-Maze...................................................................................................128
iii Platform Recognition..............................................................................128 Neurologic Battery..................................................................................129 Pathology............................................................................................................132 Brain A Deposition Quantification.......................................................132 Plasma Cytokine Analysis......................................................................132 Multi-metric Statistical Analysis........................................................................135 Correlation Analysis...............................................................................135 Factor Analysis.......................................................................................142 Discriminant Function Analysis.............................................................142 Discussion.......................................................................................................................145 General Summary...............................................................................................146 Behavioral Results..............................................................................................147 Pathology............................................................................................................151 Correlations.........................................................................................................152 Factor Analysis/Discrimin ant Function Analysis...............................................155 Proposed Mechanisms of A -Sensitive T Cell Mediated Cognitive Improvement in APPsw/PS1 Mice............................................................................................156 Stimulation of the immune system ag ainst an endogenous self peptide, A ...........................................................................................................156 Overall down-regulation of the immune system, or of specific immune cells which may contribute to neuroinflammation..................................158 Neuronal protection................................................................................158 Clinical Implication of T cell Study Findings and Proposed Future Investigations......................................................................................................159 References.......................................................................................................................161
iv List of Tables Table 1. For all Tg+ animals, a correla tion matrix of RAWM error and latency measures vs. 6E10/Thioflavin S measures of A plaque burdens in the parietal cortex and hippocampus.. 136 Table 2. Correlation matrix for RAWM erro rs vs. plasma Cytokine measures for Tg+/T cell animals 141 Table 3. Factor loadings of behavioral measures... 143 Table 4. Summary of discrimi nant function analyses 145
v List of Figures Figure 1. IFNprimed Microglia cultured in the presence of A are capable of MHC II associated presentation of A to both Th1 and Th2 cell types... 57 Figure 2. Mechanisms of T cell based neuroprotection. 106 Figure 3. General protocol time lin e for the adoptive transfer study. 111 Figure 4. Pre-infusion overall wo rking memory performances. 123 Figure 5. Post-infusion Block 1 RAWM errors. 124 Figure 6. Figure 6. RAWM during post-infu sion Block 3 Trial 1 vs. Trial 4+5... 126 Figure 7. RAWM pre-infusion vs. postinfusion comparisons for Trial 4+5 combined errors by group... 127 Figure 8. RAWM Trial 4+5 performance for pre-infusion (final block) vs. postinfusion (first block)... 129 Figure 9. Post-infusion Y-maze spontaneous percent alternation performances... 130 Figure 10. Post-infusion Platform Recognition latencies by day... 131 Figure 11. Quantification of 6E10and Thioflavin S A burdens. 133 Figure 12. Standardized m ean signal intensities fo r 10 plasma cytokines... 134 Figure 13. For all Tg+ mice, a correlat ion between RAWM Block 1 Trial 4+5 Errors and Hippocampal 6E10 burden.. 138
vi Effects of Adoptive Transfer of Beta-Amy loid Sensitive Immune Cells in a Mouse Model for Alzheimers Disease Daniel Shippy ABSTRACT One major therapeutic target for preventing and treating Alzheimers Disease (AD) is removal of excess -amyloid (A ) from the brain. Both active and passive immunotherapies targeting A have proven effective in reducing brain A levels and improving cognitive function in mouse transgenic models of AD. Howe ver, these approaches can induce adverse neuropathologic effects and immunologic over-activation. In deed, clinical trials of active A immunotherapy in AD patients were halted due to development of meningoencepha litis, apparently resulting from wide-spread neuroinflammation. Here we show that a more rest ricted and specific immune re-activation through a single adoptive transfer of A -specific T cells can provide l ong-term benefits to APPsw+PS1 transgenic mice that last at least 1 months. A -sensitive splenocytes and lymphocytes were generated in normal mice, re-stimulated with A in vitro and then adoptively transferred into cognitively-impaired APPsw+PS1 mice. Compared to control transgenic mice through 1 month post-infusion, those mice that received A -sensitive T cells exhibited a reversal of pre-infusion working memory impairment and demonstrated superior basic mnemonic processing. Step-wise forward Discriminant Function Analysis of behavi oral results clearly demonstrated that T cell infused mice performed comparably to wild-type non-transgenics, fu rther emphasizing the extent of
vii cognitive benefit this therapeutic technique afforded. Importantly, a global inflammatory response did not accompany these benefits. Though no overall reductions in A deposition were noted for T cell recipient mice, a subset of T cell infused mi ce that benefited most in cognitive function had reduced hippocampal burdens, suggesting that hippocampal A burdes did play a role in determining performance capabilities of these mice. Since chronically high levels of A such as those found in APPsw+PS1 mice cause immune h ypo-responsive/tolerance to A our results indicate that adoptive transfer of A -sensitive T-cells can superced e such immune tolerance to A and may provide a safe, long-lasting therapy for AD.
8 Alzheimers Disease Alzheimers Disease Background It is estimated that by the year 2050, 13.2 million individuals in the United States will be living with a diagnosis of Alzheimers Disease, further reinforcing its position as the most common neurodegenera tive disorder and leading cause of dementia therein (Hebert et al. 2003). Alzh eimers is characterized by a progressive loss of cognition and retention accompanied by varied disrup tions in behavior and in many cases extreme psychological disturbances. These symptoms lead to eventual loss of social integration while at the same time stripping the affected individual of the ability to care for his or he r self. Death is the eventual end result of the disease, though it is often preceded by an extensive period of complete mobility loss and a comatose state. The resulting widespread social and economic impact of caring for such afflicted individuals, already devastating today for many families affected by the disease, will undoubtedly continue to grow as Alzheimers becomes more prevalent. Though treatments for Alzheimers ha ve been approved by the FDA, none provide long-term reversal or delay of the disease and none serve as a cure (Doody et al. 2001, Areosa et al. 2005). This, despite th e fact that the dis ease was characterized nearly a century ago in 1907 by Dr. Alois Alzheimer, who correlated some of the behavioral impacts of the disease to the most noteworthy aspects of its pathology,
9 including dramatic loss of brain mass and development of both neuritic plaques and neurofibrillary tangles as well as gliosis (Alz heimer et al. 1995). Even with the surge of research activity which has occurred within this field over the past several decades resulting from both its increased prevalence and from increased exposure due to diagnosis of Alzheimers Disease in such noteworthy figures as President Ronald Reagan, progress in developing treatments and determining preventative measures has been slow at best. Recent innovative studies involving varied immunotherapeutic approaches have explored out side the field of typical pharmacological treatments and have shown promise, however have at the same time been a source of major disappointment in human applications. With increased understanding of the faults of such treatments, it is possible that altern ate immunological approaches may lead to valid treatments for Alzheimers. Behavioral Charact erization of AD Due to the progressive nature of Alzheimers Disease and to the incomplete understanding of its development, it has prove n useful for clinicians to establish a staging sequence for Alzheimers behavioral impacts. Prior to development of dementia, it is thought that most individua ls enter a transition state termed Mild Cognitive Impairment (MCI). It is estimated that up to 15% of individuals diagnosed with MCI proceed to Alzheimers development each year, though many are diagnosed with AD without a prior MCI di agnosis (Feldman et al. 2004). Using mainly the extent of cognitive impairment present, Alzheimers patients can then be grouped as early, moderate, or advanced phase. Only after death can a true
10 verification of AD diagnosi s occur via correlation w ith specific Alzheimers pathology. A diagnosis of Mild Cognitive Impair ment does not always correlate to development of AD, though the term was speci fically developed to describe a state of lesser impairment preceding Alzheimers. It is thought that some limitations of MCI application may be due to variations in sy mptom sets used to de fine it and reliance on the testimonies of patients and relative s for establishing a history of memory disruption (Davis and Rockwood 2003). Still, MCI is considered a strong risk factor for AD development and is therefore clini cally relevant. MCI is characterized by reported working memory deficits, verified to be more significant than is normal for a patients age, with no loss of other func tions. Recent studies have also found that certain patterns of neuropsychiatric symptoms are present in up to 59% of subjects diagnosed with MCI, including various mood abnormalities and psychosis. In fact, there appears to be a correlation between presence of neuropsy chiatric symptoms common in AD and progression from MCI to early Alzheimers (Feldman et al 2004, Hwang et al 2004). During the initial phase of AD, mild to moderate cognitive impairment with respect to short-term memory is present a nd is definitive for diagnosis during the life of the patient (McKhann et al. 1984). Specific behavioral deficits and abnormalities are present throughout AD and may show some correlation to staging determined by working memory deficits, though it should be noted that many of these symptoms are not consistently present on an individual basis. Other major clinical features often associated with the disease include langua ge and visuospatial deficits, though the
11 former may not appear until mid or often late phase AD. Behavioral symptoms such as depression, agitation, anxiety, aggressi on, and insomnia, as well as psychological symptoms such as delusions and hallucinations have been found to accompany these deficits during early phase AD, though again these symptom sets are more prevalent in later AD phases (Moran et al. 2003). These symptoms are often coupled with functional behavioral impacts such as erro rs in judgment, untidiness, and transient confusion as the disease progres ses. Not all symptoms are present in all patients, they occur in widely varied percentages of indivi duals in this phase of the disease. Moran et al. recently described three classes of early phase AD patients based on observed symptom sets, the first with little to no symptoms outside of memory deficits, the second with additional symptoms related mo stly to anxiety and depression, and the third with additional aggression-based symptoms. Such grouping methodologies could prove useful in unders tanding the basis of these symptoms if correlation to neuropathology can be made. Entry into moderate phase AD is established based on continued loss of shortterm memory coupled with new, progressi ve loss of long-term memory. Other symptoms described for MCI and early pha se AD are common to moderate phase AD as well. Language deficits, which tend to develop later than visuospatial disruptions, become especially prevalent by the time a patient enters mo derate AD. Finally, it is not uncommon for both behavioral and psyc hological disturbances developed during MCI or early phase AD to become even more severe before being lost to immobility and stupor common to advanced phase AD (Hart et al. 2003). Advanced phase AD involves an almost complete loss of all memory forms and entry into a vegetative
12 state which continues until deat h due to illness or system ic loss of bodily functions occurs. In a retrospective study, Hart et al. showed that depression, delusions, and anxiety are the most prevalent of the psychological symptoms present throughout AD, although hallucinations, elation, and disinhibition may also occur. Numerous studies have also shown that apathy, motor distur bance, aggression, irritability, appetite changes, and sleep disturbances are all relatively common behavioral symptoms for AD (Hart et al 2003, Moran et al 2003, Lyketsos et al 2001). Of these, apathy appears to be the most prevalent form of behavioral change which occurs. From the time of diagnosis of AD, an individual may often live as long as twenty years. With further development of correlations between certain forms of MCI and progression to AD, the time with which an individual and his or her family may live with such a diagnosis is ever in creasing, further rein forcing the importance of developing a treatment for the disease. Pathological Characterization of AD Although it was Dr. Alois Alzheimer who first described the general pathological characteristics of AD, an unders tanding of the true molecular nature of these markers could not be gained until the 1960s when electron microscopy allowed for more detailed analysis of neuritic plaques and neurofib rillary tangles (NFTs), the two hallmark brain pathologies of the disease (Kidd 1964). In the decades since, numerous hypotheses regarding the cause of this pathology and its correlation to behavioral impacts as well as other AD pa thological characteristics have been formulated. Perhaps the most widely ac knowledged of these focuses on the presence
13 of a particular protein fragment, Beta-amyloid (A ), in the brain and has thus been termed the amyloid hypothesis. Other AD pathologies such as brain atrophy, neuronal dystrophy, synaptic loss, and NFT fo rmation are thought to be tied in some way to A peptide formation in the CNS under this hypothesis. A is formed from proteolytic proce ssing of the Amyloid Precursor Protein (APP), a 695-770 amino acid protein typica lly found in membranes of many cell types throughout the body. APP contains three relevant cleav age sites which are discriminately cleaved by the , and -secretase complexes. -Secretase cleaves within the A domain of the APP molecule, preventing formation of A while sequential cleavage by and secretases is needed for amyloidogenesis. The activity of these latter two enzymatic complexes appear s to be increased in later life and even more so in AD (Canevari et al. 2004). While A is produced in and can be found in fairly low concentrations throughout the body, it is its activity within the CNS which appears to be relevant to AD, both in regard s to formation of neuritic plaques as well as through its presence in a soluble form. This is not to suggest that peripheral A is not relevant to the disease, howev er, as it has been shown that A is capable of passing through the blood brain barrier (BBB) (Zlokovic 2004). Some studies in animal models have even suggested that A may be able to trigger disruptions in the BBB, which is commonly disrupted in AD patients (Farkas et al. 2003). Two isoforms of the A peptide can be found within neuritic plaques. Most contain the 42 amino acid isoform (A 1-42), though this isoform is actually less common elsewhere in the CNS than the second A1-40 peptide. The A 1-42 peptide is
14 thought to be one of the initiating factors for neuritic plaque formation, originally aggregating to form what are known as low density or diffuse plaques (Lemere et al. 1996). Inflammatory proteins such as APOE and -ACT, along with a number of other proteins produced by various cells of th e CNS, are also thought to contribute to formation of these plaques (Nilsson et al. 2004). Though the exact nature of plaque fo rmation progression is not understood, it appears that these diffuse plaques are precu rsors to the compact plaques noted by Dr. Alois Alzheimer. It is further acknowledge d that as these plaques form there is an activation of microglial cells which, via secretion of various cytokines, induce activation of adjacent astrocytic cells, resu lting in a co-stimulatory inflammatory response which may accelerate plaque maturation (DAndrea et al. 2004). Other immune cells such as B and T-cells appear to be involved in regulation of this response. Cytokines such as IL-1 and TNFalso act by inducing APP messenger RNA translation, resulting in further increase in A production (Rogers et al 1999). Inflammatory proteins such as APOE and -ACT are produced as part of the response and may be involved in further A deposition by catalyzing conversion of A to a -pleated sheet form which is prone to aggregation (Nilsson 2001). Degradation of adjacent neuronal dendritic and axonal processes further increases the local inflammatory response via release of intercellular inflamma tory factors as well as additional degrading lysosomal and cytopl asmic enzymes, resulting in even further damage to surrounding cells. This form of damage elicited by A plaques has caused this peptide to be termed neuro toxic. During this process, A 1-42 deposition
15 continues coupled with A 1-40 incorporation, finally resul ting in formation of compact (mature) neuritic plaques, each of which is composed of a dense A core surrounded by reactive microglia and astrocytes and asso ciated with inflammatory proteins such as APOE and -ACT. (Selkoe 2001, Nilsson et al. 2004). Diffuse versions of the neuritic plaque can be found in various regions of the brain, both inside and outside of the hippo campus and association cortices, which are the regions most affected with regard to neuronal loss and NFT formation as well as neuritic dystrophy associated with A pathology. Neuronal loss and other AD pathologies, including various forms of neuroarchitectural rearrangements in these regions, are thought to correlate to the losses in working memory noted in AD. It is primarily in these regions that high levels of compact neuritic plaques can be found, further reinforcing the importance of thes e plaques with regards to AD pathology. It is important to note that Cere bral Amyloid Angi opathy (CAA) is a pathological feature common in AD which is also directly caused by deposition of A this time within the vasculature. Am yloid plaques formed within blood vessels can lead to microhemmorages and in some cases instigate a stroke. Approximately 80% of individuals over the age of 65 diagnosed with AD are thought to develop CAA. Other aspects of this pathology have been linked to A presence, though not necessarily in compact, aggregated form. Aggregation is thought to begin intracellularly, with A 1-42 fragments twisting together to form first dimmers and then various oligomeric isoforms (Selkoe 2001). Such intracellular A is thought to
16 influence phosphorylation of tau proteins (Cummings 2004). This phosphorylation event is known to be linked to imbalances in cellular kinases and phosphatases. Tau proteins are microtubule com ponents that, when phosphorylated, can twist together in -helical pairs to form Paired Helical F ilaments (PHFs). These accumulate to form true NFTs, resulting in disruption of ne uronal function and eventually cell death (Canevari et al. 2004 and Sobow et al.2004). Loss of cholinergic system functions is another major hallmark of AD, and this too has been linked to A activity. A is thought to interact with nicotinic receptors of cholinergic neur ons, specifically by binding to 7 nACh Receptors (Kihara and Shimohama 2004). This interacti on is thought to result in inhibition of calcium activation and acetyl choline release in these ne urons, indicating a mechanism for loss of function in cholinergic pathways. A is also thought to enhance glutamate excitotoxicity via NMDA receptor interactions These specific in teractions between A and neurons prone to dysfunction in AD st rongly reinforce the importance of the peptide in the progression of AD, as loss of memory and cognitive functions have been linked to disruption of these path ways (Kihara and Shimohama 2004). Further A involvement via activation of caspases has also been proposed, indicating a mechanism by which cellular a poptosis, a widespread occurrence within the CNS in AD, may be induced (Eckert et al. 2003). Inte ractions between A and mitochondrial membrane transporters al ong with mitochondrial enzymes are thought to be involved in decreased mitochondrial activity. This has been proposed as one mechanism for high levels of oxidative stress found in AD (Canevari et al. 2004, Marlatt et al. 2004).
17 A is also though to disrupt Ca++ tran sport in neurons, astrocytes, and microglia. Alteration of Ca++ transport functio nality has been linked to alterations in neuronal signal processing, increased neurona l vulnerability to excitotoxicity, and activation of both microglia and astr ocytes (Cavevari et al. 2004). A can also be linked to production of oxygen free radi cal species, which occurs during oxygen dependent formation of A -pleated sheets. Free radicals are also formed by microglia upon activation by A These reactive species are thought to play a substantiative role in triggering oxidative damage prevalent in AD. This damage tends to affect mainly lipids and DNA, and is 3 times more prevalent in AD than in normal elderly individuals (Hensley et al. 1996, Selkoe 2001). Thus, many of ADs pathological features can be attributed to abnormal production and function of A within the brain. However, it should be noted that there are also several lines of evidence which suggest that the amyloid hypot hesis described above may be incorrect or at the very least incomplete. Those malcontent with the amyloid hypothesis have emphasized an inability of the hypothesis to explain early neurophysiological develo pments such as synaptic dysfunction, neuronal loss, a nd decreased LTP, which occur in varying extents before formation of A plaques in AD. It has been suggested that A accumulation, like NFT formation, is truly more a consequen ce of a separate cellular dysfunction, not yet entirely elucidated, which occurs be fore overproduction and aggregation of the peptide. This is supported by findings in tr ansgenic mice which indicate that synaptic dysfunction may occur before formation of diffuse or compact amyloid plaques (Lee et al. 2004). It has ev en been shown that A may be involved in regulation of long-
18 term potentiation, indicating th at the peptide is not simp ly a neurotoxic by-product and that its presence in AD may be an attempt at compensating for neuronal loss, not necessarily the cause of it (Koudinov and Be rezov, 2004). It should be noted that other studies have shown the opposite effect of the peptide in lowering capacity for LTP (Selkoe 2002). Other hypothesis for the actual mechanisms underlying AD focus on cell cycling and mitochondrial enzymatic disrupti ons which result in oxidative stress and neuronal loss and on abnormal levels of enzymes involved in acetylcholine synthesis and degradation (Raina et al. 2004, Zhu et al. 2004). While it is necessary to consider these possibilities in seeking an understandi ng of, and a treatment for, Alzheimers, it is also important to consider in the pr esent that previous studies have shown correlation between A levels within the CNS and extent of demetia in early stages of AD (Bussiere et al., 2002), though A load does not correlate directly to cognitive disturbances as the disease progresses (Te rry et al. 1991). Similar correlation can be made between A plaque loads and cognitive impairment in transgenic mouse models of AD (Gordon et al. 2001). Synapse loss and neurodegeneration induced by neurofibrillary tangles serve as better correlates to cognitive impairment (Terry et al. 1991, Arriagada et al. 1992). However, as has been shown above, these pathological findings may be tied to and perhaps even directly caused by the presence of A Also, as elicited above, there are numerous studies which have shown mechanisms by which A can be neurotoxic and/or can have other negative physiological impacts within the CNS even in soluble form. Th erefore, regardless of other mechanisms involved in development of AD, there rema ins definite validity in searching for a
19 means to elicit A clearance as a means of treatment for the disease or for preventing A formation/aggregation as a means of prevention. Genetics of AD Further support for the amyloid hypothesi s is provided by analysis of known genetic mutations linked to Familial Azheimers Disease (FAD). FAD typically follows a similar time course to sporadic AD, which has no known genetic basis, and differs only with regards to age of onset. FAD diagnosis typically occurs at a much younger age, as early as 40, than sporadic AD. Each mutation linked to FAD, which by definition is congenital and runs within families, has an in vivo effect of increasing A peptide formation, and is an autosomal dom inant trait. However, only as few as one percent of reported AD cases are curren tly attributed to FAD. The mutations linked to FAD thus far include those of th e APP gene itself, al ong with those of the presenilin genes (PS1 and PS2) (Zekanowski et al. 2004, Selkoe 2001). In 1984, Glenner and Wong proposed that genetic defects in Alzheimers disease would be found on chromosome 21 afte r discovering that the amyloid protein prevelant in both Alzheimers diseas e and Downs syndrome pathology were homologous (Glenner and Wong, 1984). Indeed, it has been found that mutations of the APP gene which have been linked to FAD are located at either end of the A coding domain on chromosome 21p21. There are 16 unique mutations of the gene which promote AD pathology and 4 which do not, though only 4-6% of FAD cases are attributed to APP mutations. The re sult of each mutation is increased A peptide formation, most likely through increase d processing of the APP molecule by or secretases. With regards to current resear ch in transgenic models, two of the most
20 relevant of these are the Swedish and London APP mutations. The Swedish mutation is located at positions 670 and 671 of the APP molecule and involves a double missense mutation, while the London mutati on, which is a single missense mutation, is located at position 717. These correlate to the domains cleaved by the and secretases. Both mutations result in dramatic increases in overall A production, with varying ratios of A 1-40 to A 1-42 produced. Specifically, the London mutation results in an increased ratio of A 1-42 to A 1-40 while raising levels of both isoform overall, whereas the Swedish mutation results in approximately equal increases in the levels of both isoform (Eckman et al. 1997, Suzuki et al. 1994). As with the London mutation, it is production of the A 1-42 peptide which is most increased as a result of the presenilin (PS) mutations. It is thought that both the PS1 and PS2 molecules are incorporated into the secretase complex and serve as part or all of the active site for this complex (Farmery et al. 2003). The proteins are generally found in endoplasmic reticulum of cells and most likely act on APP during its processing via this organelle. The resu lt of the presenilin mutations is increased activity of the secretase complex, resulting in increased amyloidogenic processing of APP. The PS1 gene codes for a 467 aa protein and is lo cated on chromosome 14q23.3, while the PS2 gene codes for a 448aa protein and is located on chromosome 1q31-42. The genes are highly homologous in their sequences, indicating a similar function. PS1 mutations are the most co mmon cause of FAD based on current data, and more than 100 of these mutations have been identified, most of which are missense. Only 9 PS2 mutations have been identified, all of which are missense, and
21 the prevelance of these mutations is signifi cantly lower than thos e of PS1 (Kimberly and Wolfe 2003, Gaskell and Vance 2004). Diagnosis of AD The predominant criteria used for the diagnosis of Alzheimers comes from the National Institute for Neurologic a nd Communicative Disorders and StrokeAlzheimers Disease and Related Disord ers Association (NINCDS-ADRDA), which specifies diagnosis as being either definite probable, or possibl e based on the extent of confirmation available. If a hist ological confirmation, preceded by clinical diagnosis, is obtained, then a definitive di agnosis for AD is established. Clinical diagnosis without histological confirmation is deemed as a probable AD case, while a case with clinical abnormalities but no othe r diagnostic approaches available is considered possible AD (Cummings 2004). The specific clinical features looked for in diagnosing AD are the three major behavioral hallmarks of the disease; ep isodic and working memory impairment and both language and visuospatial deficits. Ep isodic memory is based on retrieval of memories of ones specific past events, t ypically with events which have occurred recently being most easily forgotten (Cum mings 2004, Greene et al. 1996). Working memory, or short-term memory, is typically associated only with the ability to recall occurrences from within minutes of the pr esent. Loss of both episodic and working memory, which largely correlates to genera l anterograde amnesia, is perhaps the easiest of the behavioral hallmarks of AD to identify and the most important for establishing clinical diagnos is (Welsh et al. 1992, Green e et al. 1996). There are many tests which can be administered by clin icians in able to determine if such
22 memory loss is present and extent of impairment, including the most common, the Mini-Mental State Exam (MMSE), which i nvolves giving a patient number of words to remember followed by a series of ques tions which serve as a distraction. An individuals lack of ability to retain the words given to them at the beginning of the task (registration-recall) is an indicator of short-term memory impairment. Other tools for evaluating mental state of potent ial AD patients include the 7-Minute Screen and Geriatric Depression Scale (Langbart 2002). Basic histological measurements which may be correlated to memory impairment to help strengthen a diagnosis of AD include cerebro-spinal fluid (CSF) levels of A 1-42 and both tau and phospho-tau (p-tau ) proteins. Numerous studies have shown that a dramatic decrease in CSF A 1-42 is prevalent in AD (Sobow et al., 2004; Andreasen et al., 2003). This is thought to correlate to trapping and aggregation of the peptide in insoluble forms within the CNS, allowing the brain to act as an A 1-42 sink. At the same time, levels of both tau and p-tau in the CSF increase significantly in most cases of AD (Sobow et al., 2004; Andreasen et al., 2003). Unfortunately, CSF sample collecti ng is a highly invasi ve process, making widespreade clinical application of these histological markers difficult. Current research studies in the area of A and tau levels in the plasma or other more accessible fluids are ongoing (Sobow et al. 2004). Several brain imaging techniques also serves as diagnostic tool for AD. Computed tomography (CT) and magnetic re sonance imaging (MRI) have proven to be useful as multiple scans over time can show overall brain atrophy and localized atrophy of the hippocampus and entorrhinal cortex (Barnes et al., 2004; Pennanen et
23 al., 2004; de Leon et al., 1997). This form of atrophy is specific to AD, allowing for exclusion of other dementia types, and can al so be used to monitor the progress of the disease. In some cases it is possible to detect this atrophy during MCI, allowing for anticipation of an imminent diagnosis of AD. Importan tly, studies of AD patients have shown direct correlation be tween hippocampal atrophy and memory impairment, specifically short-term memo ry (Grundman et al. 2003). Similarly, overall atrophy of the cerebra l cortices has been correlated to long-term memory and memory retrieval impairment in AD (Bilgle r et al. 2004). As discussed above, this atrophy is likely caused by neuronal loss a nd restructuring common in these areas during AD progression. More recently, positron-emmision tomogrogphy (PET) and single-photon-emission CT scans have been applied to diagnosis of Alzheimers. These scans allow for analysis of regional brain activities, and can target deficits in activity in regions affected by AD (C ummings 2004, Minoshima 2003). Recently, the possibility of a technique for imagi ng of amyloid plaques using a modified version of the PET scan has been explored which may prove to be a very useful diagnostic tool in the future (Klunk et al. 2003). A diagnosis of AD, whether it be defin itive, probable, or possible, in a family with a strong history of th e disease with early onset is sufficient cause for genetic testing for FAD. Tests are cl inically available for presenilin mutations, however their application is limited and no su ch tests are available for APP mutations. Since there are numerous mutations which may be presen t in these genes which do not result in FAD, it is important that both physician and patient be well informed before genetic testing results are reviewed. Extensive Gene tic counseling is an important aspect of
24 this education, meaning that genetic testing for FAD must typically occur through a hospital with trained genetic counseling staff and should not be performed at a standard physicians office. As more F AD mutations and their resulting phenotypes become better understood, it is likely that the availability of genetic testing will spread. In the present, the results of genetic testing for FAD are not definitive for diagnosis but can confirm the presence of a FAD mutation in a family, which may prove useful for such families in preparing for the disease financially and via preventative measures (Gaskell a nd Vance 2004, Zekanowski et al. 2004). Risk Factors for AD Certain risk factors for AD such as the most significant one, aging, are in no way preventable. Other examples of these include sex (females run a higher risk of developing AD than do males), having a firs t degree relative with AD, and carrying the ApoE4 allele of the ApoE gene. The A poE allele is an extremely important risk factor for AD, with each ApoE4 allele car ried resulting in a sizable increase in AD risk and typically earlier ons et of the disease if deve loped (Selkoe 2001, Yao et al. 2004). Other risk factors for Alzheimers may be prevented or treated, making them particularly important for individua ls already at a high risk for AD. Preventable AD risk factors include head trauma (especially in young adulthood), high blood pressure in midlife, high cholesterol levels in midlife, high blood homocysteine levels, and a number of f actors related directly to diet (Fleminger et al. 2003, Wellington 2004, Ellinson et al. 2004) High levels of social interaction, mental engagement, and exercise are thought to be related to lowering risk for developing AD as well via cognitive stimulation. Diet and exercise are also indirectly
25 important for regulation of blood pressure and cholesterol levels (Wang et al. 2004). Reduction of cholesterol by use of statins has also been strongly linked to a decreased risk for AD (Miller and Chacko 2004). Blood homocysteine levels can be regulated by intake of vitamins B6, B12, and folic acid, which are involved in metabolism of the molecule (Ellinson et al. 2004). Other dietary considerations for pr eventing AD include decreasing caloric intake, increasing intake of Tocopherol (Vitamin E) and other antioxidants, and/or increasing dietary levels of Omega-3 fatty acids (Berman and Brodaty 2004, Bourre 2004). Antioxidants are important for defe nding the brain agains t oxidative damage, and some epidemiological studies looking at dietary antioxidant levels, especially Vitamin E levels, have shown a negative co rrelation between intake levels and risk for AD development, though clinical trials have elicited mixed results (Berman and Brodaty 2004, Sobow and Kloszewska 2003). The antioxidant properties of Vitamin E are important in AD largely because they prevent membrane lipid peroxidation, reducing oxidative damage in the brain. It is thought that Omeg a-3 fatty acids such as Docosahexaenoic acid (DHA) are important for neuronal health and may directly impact APP processing. Epidemiological studies indicate that decreas ed dietary levels of fish, which contain high levels of DHA and other fatty acids, can be consid ered a risk factor for AD (Bourre 2004, Kalmijn et al. 1997). A diet high in fruits and vegetables along with fish results in increased uptake of vari ous antioxidants and n-3 fatty acids, and patients at risk for Alzheimers are often encouraged by doctors to pursue such diets both because of the direct impact of th ese molecular components and the indirect
26 impact of such diets on cholesterol levels. Other diete tic factors which have been linked to decreased AD risk in clude regular intake of nicotine, caffeine, and alcohol, though for each of these it has yet to be determined if possible preventative capabilities of these drugs outweigh risks associated with th eir use or abuse (Dall'Igna et al. 2004, Teaktong et al. 2004, Letenneur et al. 2004). Recently, the importance of environment with respect to AD risk has been a major point of emphasis in Alzheimer res earch. It is becoming apparent that education level, social in tegration throughout life, and engagement in mentally stimulating activities such as writing and working puzzles has some level of protection against AD development (Fratiglioni et al. 2004). Most risk factors for AD have been determined through epidemiological studies, so th eir exact role and mechanism of protection are difficult to define and are often tied to those of other risk factors, such as with cholesterol, diet, a nd exercise. Fortunately, most actions that individuals can take to help in prevention of AD are generally encouraged as part of a standard healthy lifestyle a nd can be undertaken without a full understanding of their effectiveness. Current Treatments of AD There are currently five pharmacological treatments for AD which have been approved by the FDA. Tacrine (which is no longer used clinically due to its negative impact on the liver), donepezil, rivastigmine, and galantamine are all acetylcholinesterase inhibitors, while the fi fth drug, memantine, is an antagonist for NMDA-receptors (Cummings et al. 2004). Acetylcholinesterase inhibitors have been applied clinically in the United States for ma ny years, and are directed against lost of
27 cholinergic function in AD. By inhibiting acetylcholinesterase, which breaks down acetylcholine, cholinergic transmissions are increased due to raised levels of the neurotransmitter in the synaptic cleft. Galantamine also acts as an allosteric ligand at nicotinic acetylcholine receptors, fu rther improving cholinergic function. Rivastigmine inhibits both acetylcholineste rase and butyrylcholinesterase, which is another cholinesterase largely associated with glial cells (Sca rpini et al. 2004). Treatment with these drugs brings about improvement or stabilization in cognitive symptoms of mild to moderate AD, however these improvements are small and do not typically last for more than 1-2 years. Studies are ongoing to determine whether long-term use is at all beneficial (Sca rpini et al. 2004, Sonkusare et al. 2005). Memantine was only recently approved for treatment of AD by the FDA. By acting as a partial NMDA-receptor agonist, this drug helps to prevent excitotoxic over activation of these receptors by glutamate, which is thought to be present in elevated levels in AD (Sonkusare et al. 2005). No rmal levels of NMDA-receptor activation are desirable due to their importance in l ong-term potentiation, however, since these receptors allow long periods of Ca2+ in flux into cells upon activation, having too many receptors open for extended periods may lead to Ca2+ mediated neuronal degeneration. Proper administra tion of memantine appears to be able to prevent this type of excitotoxicity while allowing for normal levels of receptor activation. This treatment leads to improved cognitive and be havioral functions in the short term over placebo. Studies combining memantine and donepezil treatments ha ve shown great
28 promise in short term treatment of moderate to severe AD (Tariot et al. 2004). Still, as with the other approved treatments fo r AD, memantines effects are relatively small and short lived, emphasizing the importa nce of continued research in the field of alternative AD treatment s (Scarpini et al. 2004).
29 Animal Models of Alzheimers Disease Various animal models for Alzheimers Disease have been developed based on a persistent need for an easily reprodu ced and cost effective means for studying behavioral and pathological progression of the disease. These models have since become invaluable in evaluating potential th erapeutics and preventatives. The most commonly used method for producing these models in mice is through random insertion of the desired gene into a ferti lized mouse egg genome via microinjection or electroporation. Mice models are desira ble because they age quickly for a mammalian animal, and are relatively easy and cheap to maintain as colonies. The genes inserted to create these mice are mutant-type human transgenes which are either linked to FAD, such as mutant APP and presenilin alleles, or which are known risk factors for AD, such as the ApoE4 allele In order for these inserted mutations to exhibit profound effects in subsequent gene rations of transgenic mice, they are inserted with strong promoters specifically activated in the mouse brain. For FAD APP transgenes, the result is massive production of A in the brains of the resulting transgenic animals which then leads to progr essive deposition of this protein fragment to form neuritic plaques in the CNS. Models based on these mutations include the PD-APP, Psw (APP23), and APPsw transgenic mice. Other mutations work in conjunction with these in double transgenic models to further enhance A production
30 and/or deposition, but are not sufficient to induce neuritic plaque formation in single transgenic models. An important example of this type of double transgenic model is the APPsw+PS1 mouse. PDAPP Model Mice transgenic for the London mutati on human allele of the APP gene were initially described in 1995 and have since been termed PD -APP mice due to use of a platelet-derived growth factor promoter for driving expression of the gene. As a result of using this promoter, PD-APP mi ce express human APP (hAPP) at levels approximately 10 times greater than expr ession levels in human AD brains. This allows for pathology development during the relatively short lifetime of the mice. mRNA expression consists of three major vari ants due to alternative splicing of exons 7 and 8 of the APP gene construct. Ov er-expression of hAPP containing the 717 single point mutation of Valine to Leucin e in the PD-APP mouse brain leads to development of pathology and behavior whic h correlates in many ways with that of Alzheimers Disease (Games et al. 1995). Pathological Characterization of the PDAPP Model AD-type pathology begins to exhibit itself in PD-APP mice between 6 and 9 months when deposits of A can be detected in plaques formed solely in the hippocampus, corpus callosum, and cerebral cortex. Deposition continues past this timepoint until plaque levels parallel or exceed those found in AD. The plaques noted in PD-APP mice brains range from di ffuse to compact, and compact plaques consist of an A core surrounded by activated astroc ytes and microglia, just as is typical for AD neuritic plaques (Games et al. 1995). While these mice do not exhibit
31 cerebral atrophy or widespread neuronal loss, key features of Alzheimers pathology in humans, they do exhibit a loss of both synapses and dendrites in the hippocampal dentate gyrus coupled with overall hippocampa l atrophy (Irizarry et al. 1997). This correlates to a similar loss in this area of the human AD brain, which is thought to contribute to short-term memory loss. Furt her studies of synaptic transmission in the hippocampus of these mice have shown that there is a reduction of neurotransmission and LTP which precedes A deposition (Giacchino et al. 2000). This reduction is also thought to play an important role in affecting behavioral impairment of these mice. Behavioral Characterization of the PDAPP Model In 1999, Dodart et al. reported that PD-APP transgenic mice show ageindependent impairment in the radial arm maze task as early as 3 months of age despite no deficit in overall motor activity. By 6 months they were shown to be impaired in object recognition tasks (Doda rt et al. 1999). In 2000, a correlation between object recognition and A plaque loads was re ported along with data correlating deficits in both spatial and wo rking memory to changes in synaptic density and hippocampal atrophy (Dodart et al. 2000). That same year, Chen et al. (2000) reported an age relate d deficit in working memory using a modified version of the water maze which included use of a shifti ng assortment of visual cues and escape platform locations. Alterations in platform locations after several trials required that the mice tested learn each new location, forcing them to rely on working memory. For all age time points, PDAPP mice took l onger to learn the location of the hidden platform than non-transgenic controls in an age-independent manner. However,
32 based on last day performance at each time point, impairment was not present at a 6-9 month time point but became apparent at a 13-15 month time point and became even more prominent at a final 18-21 month time point (Chen et al. 2000). In 2004, Nilsson et al. tested PDAPP mice and non-tran sgenic controls in a full 6 week battery of sensorimotor and cognitive based tasks starting at 2 months and 16 months for separate groups of animals. While no differences were noted between groups at the earlier time point, aged PDAPP animals show ed impairment in the final block of Morris water maze and in overall RAWM performance (Nilsson et al. 2004). That same year, Leighty et al. reported a si gnificant statistical correlation between impairment in Morris water maze, platform recognition, and RAWM performance and deposition of A (diffuse and/or compact) in the hippocampus and parietal cortex in 15-16 month old PDAPP mice. A correla tion was also made between circular platform escape latency and diffuse A deposition in the cerebral cortex. These patterns of impairment correlated to A deposition in the hippocampus and cerebral cortex of the mice brains indicate a direct mechanism for working memory loss in this model of AD. APPsw (APP23) Model The Swedish mutation transgene has also been used to create hAPP transgenic mice commonly used in Alzheime rs research. This transgene carries a double mutation at positions 670 (Lys Asn) and 671 (Met Leu) in the APP gene, a mutation originally discovered in a Swedish family predisposed to early onset AD (Mullan et al., 1995). The most prevalent lin es of mice carrying this mutation are the
33 APPsw/Tg2576 and the Psw/APP23 mice, which carry different promoters and have different transgene insert locations resulting in some distinct pathological features. Pathological Characteriza tion of the Tg2576 Model In Tg2576 mice, the APP transgene is expressed only by neurons, predominately in the hippocampus but with a constant, lower leve l of expression in other varied cortical regions (Irizarry et al. 1997). By 6-7 months, insoluble A is present in the brains of thes e mice, with levels of both A 1-42 and A 1-40 isoforms then increasing drastically through 10 months (Kawarabayashi et al. 2001). Amyloid plaque deposition occurs by 11 months in cortical and limbic regions. Past this time point and through 23 months of age, levels of both diffuse and neuritic plaques increase to levels similar to those found in AD. As a result of plaque formation, decreases in plasma and CSF A levels can be noted progr essively through these age time points (Kawarabayashi et al. 2001). Amyloid plaques in Tg2576 mice have been shown to be associated with gliosis and neuritic dystrophy, though im portantly no loss of hippocampal CA1 neurons have been reported (Irizarry et al. 1997). A formation and aggregation has also been associated with a number of other pathological and neurophysiological features, including activation of microglia and astrocytes, induction of oxidative stress and damage including lipid peroxida tion, and impairment of LTP in CA1 and dentate gyrus neurons (Frautschy et al. 1998, Benzing et al. 1 999, Mehlhorn et al. 2000, Smith et al. 1998, Practico et al. 2001, Chapman et al. 1999). Activated
34 microglia were shown to localize with A depostits and increase expression of Interlukin-1 (IL-1 ) and Tumor Necrosis Factor(TNF) (Benzing et al. 1999). Immunoreactive astrocytes directly adjacent to plaques were shown to increase expression of IL-6 and Glial Fibrillary Acidic Protein (GFAP) (Meh lhorn et al. 2000). These activated immunoreactive cells serv e to enhance the inflammatory response elicited by A Tg2576 mice have also been shown to have an increased risk for development of ischemic brain damage, eith er as a direct result of amyloid plaque formation or as a result of the inflammatory response process (Zhang et al. 1997). Pathological Characteriza tion of the APP23 Model The Psw or APP23 model differs in pathology from the Tg2576 model in one major respect. These mice are known to develop Cerebral Amyloid Angiopathy (CAA), which results from amyloid plaque fo rmation in the cerebral vasculature. This severe pathological feat ure is not present in Tg2576 mi ce but occurs at as early as 14 months in the Psw model. CAA in thes e mice is associated with loss of neurons in areas adjacent to CAA pathology (Cal houn et al. 1999). Synaptic abnormalities, microglial activation, and microhemorrhage ar e also prevalent in these regions. Damage to the vasculature can be extrem e, beginning with loss of vascular smooth muscle cells, aneurismal vasodialation a nd microhemorrhages and eventually leading to large hemotomas and in some cases hemorrhagic stroke (Winkler et al. 2001). CAA has been noted in AD, leading some re searchers to believe that the Psw mouse provides a more realistic AD model (C alhoun et al. 1999, Winkler et al. 2001). Behavioral Characterization of the Tg2576 Model
35 Behavioral deficits ha ve been found in both Swed ish mutation transgenic mouse models. Initial studies in Tg2576 mice showed behavioral impairment in Ymaze and Morris water maze tasks by 10-11 months of age, at the same time plaques developed in these animals (H siao et al. 1996, Irizarry et al. 1997). Impaired CA1 and dentate gyrus neuronal LTP was also found to correlate to spatial working memory impairment based on T-maze alternation impairments at 16 months compared to non-transgenic controls (Chapman et al. 1999). In two separate studies, Holcomb et al. showed that Y-maze impairme nt with regards to percent alternation was present at 3 and 6 but not 9 months, that no impairment was present at any of these time points for Morris water maze, a nd that Tg2576 mice were not impaired in a number of sensorimotor baseline tasks (H olcomb et al. 1998, Holcomb et al. 1999). In 1999, it was shown that impairment was present in 7 month old Tg2576 mice not for initial application but for reversal learning of the circular platform task (Pompl et al. 1999). In 2002, King and Arendash behaviorally characterized the Tg2576 model at 3, 9, 14, and 19 months, providing a comprehens ive overview of impairment in this model. This was done via application of a 6-week battery of sensorimotor and cognitive based tasks. Over all timepo ints, it was shown that Tg2576 mice were impaired in Y-maze spontaneous alternations by nine months in visible platform recognition, and over numerous timepoints in sensorimotor tasks such as balance beam and string agility tasks. Continue d levels of performance on par with nontransgenic mice in cognitive based tasks such as Morris water maze, circular platform, and both passive and active avoi dance through 19-months indicated that
36 profound cognitive impairment is not present in this model even into old age (King et al. 2002 (1)). Westerman et al. (2002) f ound that memory loss in Tg2576 mice at as early as 6 months could actua lly be attributed to formation of small insoluble A aggregates, though at later time points a co rrelation between these aggregates and impairment could no longer be drawn based on Morris water maze task results alone (Westerman et al. 2002). Also in 20 02, however, King et al. showed that age dependent impairment in Morris water maze and platform recognition could be correlated with synaptophysin staining levels present in APPsw mice through 19 months of age, indicating a connecti on between impairment and compensatory measures taken by the mouse brain to combat loss of function (King et al. 2002 (2)). In 2004, Arendash et al. reported early impairment at 5 months in APPsw mice in balance beam but no other sensorimot or or anxiety based tasks. Between 5 and 8.5 months, these mice were signif icantly impaired in Y-maze percent alternations. Impairment was also note d between 5 and 7 months in Morris maze acquisition and retenti on, platform recognition, and RA WM, indicating that for these mice cognitive impairment is present at an earlier time point th an was previously noted. This was thought to be due to actions of soluble A oligomers present in the mouse brains at these timepoints, with increased affects due to continual crossbreeding of transgenics and a resulting increased influence of the APP transgene. (Arendash et al. 2004). Behavioral Characterization of the APP23 Model In the APP23 model, three recent studies have shown that cognitive impairment is present at various timepoint s in these animals. In 2003, Kelly et al.
37 tested APP23 mice at 3, 18, and 25 months in passive avoidance as well as small and large Morris water maze and platform recognition. Passive avoidance and small pool performance impairments were shown to be age-related when comparing APP23 mice to non-transgenic controls, while large pool performance was highly impaired across timepoints. No impairment was noted for platform recognition performance, which occurred during large Morris maze testing (K elly et al. 2003). Also in 2003, Van Dam et al. also tested APP23 mice in Morris water maze and passive avoidance at 6-8 weeks, 3, and 6 months. Impairment in Morris water maze was found to be age related in this study as well, however passi ve avoidance was not found to be impaired at these early timepoints. No significant differences between APP23 mice and nontransgenics were noted in a variety of be havioral and neuromot or measures at any timepoint except for increased total path length for APP23 mice in open field at 6 months. It was also noted that overnight cage activity appeared to be higher for APP23 mice than non-transgenics at various tim e points (Van Dam et al. 2003). In 2004, Dumont et al. tested ag ed (24 months) female APP23 mice in open field, elevated plus-maze, and Morris water maze. These mice showed hyperactivity compared to controls, decreased anxiety based on elevated plus-maze performance, and impairment in Morris water maze acqui sition, but not retention. This impairment did not extend to probe or platform recogni tion trials (Dumont et al. 2004). The three studies described above give a very general picture of sp atial learning impairment in APP23 mice and hint at some of the behavioral disturbances present in these animals, however comprehensive assessment of impa irment including application of working
38 memory tasks has not been performed. Futu re testing may provide insight into the exact progression of impairment in this model. APPsw+PS1 Model Formation of a mouse line transgenic for a FAD PS1 mutation was thought to be the next logic step in creation of an AD mouse model based on the relatively high frequency of PS1 mutations among FAD cas es. Single transgenic mice carrying human PS1 alleles, however, were found to be lacking significant A pathology at any time point, despite a drastic (approximately 51%) increase in A 1-42 to A 1-40 ratios in the brains of thes e mice as well as in overall A levels (Duff et al. 1996). It was found however, that by crossing thes e mice with APPsw transgenics, a double mutant model could be created which exhi bited dramatically accelerated AD-type A pathogenesis (Borchelt et al. 1997, Holcom b et al. 1998). These APPsw+PS1 mice have been characterized by a number of groups with respect to both pathology and behavior and have proven us eful for numerous treatment and preventative studies in AD research because of their early age of onset for A pathology and correlated cognitive impairment. Pathological Characterizati on of the APPsw+PS1 Model Early work in the APPsw+PS1 model by Borchelt et al. showed that, similar to PS1 single transgenic mice, these mice show an approximate 50% increase in soluble A 1-42 to A 1-40 ratios at an early timepoint (Borchelt et al 1996). Double transgenics were found to develop amyloid plaques prior to APPsw single
39 transgenics, measured by Borchelt et al. at 12 months, and reactive astrocytes were shown to be associated with these plaque s. Many of these plaques were found to stain mostly for A 1-40 at this timepoint, despite increased A 1-42 levels. It was also found that double transgenics carrying a wild-type human PS1 gene did not display accelerated A deposition, indicating that the mutant allele PS1-A246E used in the study was necessary to bring this acceler ation about (Borchelt et al. 1997). In 1998, Holcomb et al. confirmed that a dramatic increase in A 1-42 to A 1-40 ratios is present in double transgenics, this time in a APPsw+PS1-M146L line at 6 weeks of age. At 13-16 weeks, it was s hown that these mice also exhibit small numbers of deposits after thioflavin S staining, which stains fo r plaques containing fibrillar and -pleated amyloid, with dramatically increased staining appearing at 2432 weeks. Many of these deposits were also shown to be congophilic (Holcomb et al. 1998). A year later, the same group showed that by 6 months of age, all double transgenic exhibited amyloid deposits, though none were present at a 3 month timepoint (Holcomb et al. 1999). In contra st, Takeuchi et al. found that deposition occurs in APPsw+PS1 mice at as early as 3M in the frontal cortex and CA1 region of the hippocampus. At 12 months, they show ed that 28.3% of the superior frontal cortex and 18.4% of the CA1 was co vered by amyloid plaques, which is approximately 20-40 times greater than cove rage levels in APPsw single transgenic mice. It was also noted that neuron al counts did not di ffer significantly among double or single transgenic from 3 to 12 months, despite apparent disruption of neuronal architecture in doubles by compact plaques (Takeuchi et al. 2000).
40 Plaque formation at 3 months was corroborated by Gordon et al. in 2002. This group also found that compact plaques continue to form through 12 months in both the hippocampus and frontal cortex, fo llowed by formation of diffuse plaques in these as well as other regions. These di ffuse plaques were shown to be composed almost entirely of A 1-42, while compact plaques were indicated as being composed of both A 1-42 and A 1-40, thus providing an explanati on for findings by Borchlet et al. that plaques formed prior to 12 months showed higher ratios of A 1-40 relative to A 1-42. Vascular A was also detected by this group and was typically noted to be A 1-40. Importantly, it was shown that co mpact plaques are associated with dystrophic neurites, GFAP expressing astrocytes, and MHC-II immunoreactive microglia even at early timepoints, indicat ing that these plaques elicit a similar neuroinflammatory response to those in Alzheimers via both neuronal degradation and direct interaction with imm une cells (Gordon et al., 2002). The effects of both soluble and insoluble A in the brains of these double transgenics appear to cent er on loss of neuronal functi on, loss of LTP in memory centers and increases in inflammatory re sponses in areas surrounding deposits of A In 1999, Wong et al. found that cholinergic synapses in the frontal cortex of the brains of double transgenics showed a lowe r density than those of single transgenic and non-transgenic groups. A lowered aver age cholinergic synapse size in both the frontal cortex and hippocampus was also noted, though no changes were noted in basal forebrain cholinergic cell body number s (Wong et al. 1999). In 2003, Dickey et al. reported that aged, 17-18 monthold APPsw+PS1 mice show reduced mRNA expression of genes involved in LTP and me mory formation, specifically in regions
41 where A accumulation had already occurred (c ortex and hippocampus). Changes in synaptic structure, such as those noted by Takeuchi et al. (2000) were not shown to be associated with any altered mRNA le vels, however inflammatory genes were upregulated. These findings were thought to provide an example of how memory capacities can be affected pr ior to any loss of neurons via modulation of LTP and other neuronal functions (Dickey et al. 2003). In a recent study, Trinchese et al. (2004) delved further into the study of LTP loss in APPsw+PS1 transgenics, finding that abnormal LTP is present at as early as 3 months, the same time point at which plaque formation begins and abnormal short-term working memory was first apparent, as indexed by the RAWM task. It was also shown that by 6 months, after increases in overall A loads, basal synaptic transmi ssion is impaired, coupled with a loss of long-term spatial learning/memory be ginning at this time point in a plaque independent manner (Trinchese et al. 2004). These connections to cognitive impairment show correlations to many, but not all, studies focused on behavioral characterization of this mouse line. Behavioral Characterization of the APPsw+PS1 Model Early testing by Holcomb et al., 1998 found a similar level of performance among APPsw+PS1 double transgenics, single transgenic APPsws and PS1s, and non-transgenics in sensorimotor tasks at 3-4 months, indicating no baseline impairment in motor or sensory functions is present in these transgenics. Both the APPsw and APPsw+PS1 groups were found to exhibit decreased Y-maze spontaneous alternations compared to PS1 and non-transgenic mice at this same timepoint, indicating a decrease in cognitive performance in these groups. Only
42 double transgenic mice also showed an increase in overall activity in this task, which, when coupled with decreased percent a lternation performance, was thought to correlate to hippocampal dysf unction (Holcomb et al. 1998). In 1999, Holcomb et al. continued testing of double transgenics in the Y-maze task, as well as in Morris Water Maze, over two additional time points. The same deficits were found in Ymaze at 6 and 9 months for these animal s, though no significant differences were noted among groups for any measure in Mo rris Water Maze at any time point. Furthermore, it was noted that findings of decreased Y-maze performance at 3 months occurred prior to A deposition in these mice, i ndicating loss of cognition did not necessarily require plaque formation directly (Holcomb et al. 1999). As was already shown, however, several other st udies did find that some level of A deposition does occur at this early ti me point, including the 1998 study by Holcomb et al., perhaps confounding interpretation of these results (Holcomb et al. 1998, Takeuchi et al. 2000, Gordon et al 2002; Trinchese et al. 2004. In 2001, Arendash et al. found impairment in aged APPsw+PS1 mice in the Radial Arm Water Maze task, and found a correlation between the level of impairment in individual mice and the level of plaque formation. Arendash et al. (2001) comprehensively tested the doubl e transgenic mouse model in a full behavioral test battery, cross-sectionally at two time points. It was noted that transgenicity did not appear to affect cognitive performance in Y-maze, circular platform, standard water maze, or platform recognition at 5-7 or 15-17 months, and anxiety levels measured via the elevated plus maze task did not appear to differ among groups. However, double transgenics did exhibit progressive increases in
43 open field activity and in string agility impairment, coupled with continual impairment in the balance beam task. Mo st significant, thoug h, were findings of cognitive impairment in both water m aze and RAWM. Water maze acquisition was found to be impaired at the later timepoint, though importantly no significant impairment was noted during probe trials for this task. RAWM impairment was limited to the later time point for this st udy as well. These findings indicate that progressive impairment in spatial referen ce and working memory is present in these mice. (Arendash et al. 2001). In a more recent report involving the same investigators, Jensen et al (2005) reported much earlier impairment at 4.5-6 months in APPsw+PS1 transgenic mice in Morris water maze and RAWM (Jensen et al. 2005). This early impairment is consistent with findings of early impairment in other studies such as Trinchese et al. (2004), in which it was reported that APPsw+PS1 mice are impaired in RAWM at as early as 34 months and in Morris water maze at as early as 6-8 months (Trinche se et al. 2004). This differe nce from previous studies such as Arendash et al. 2001, in which early impairment was not reported, are most likely due to the effects of crossbreeding of APP transgene bearing mice resulting in an increase of its effects. These e ffects were attributed to soluble A constructs in the brains of these mice which resulted in their earlier impairment. Sadowski et al. (2004) reported impa irment at 8 and 22 months for APPsw+PS1 transgenics in Morris water maze as well as in Hebb-Williams maze. Impairment in spatial memory at the 22 month time point was correlated to both a loss of neurons (35.8%) in the hippocampa l CA1 region and to reduced glucose utilization in the hippocampus (Sadowski et al. 2004).
44 Based on results from studies performed w ith APP transgenics, it is clear that the resulting presence of A in both soluble and insoluble form is sufficient to elicit cognitive impairment in these mice which in many way parallels that of Alzheimers Disease (Chen et al. 2000, Are ndash et al. 2001). Progressi ve loss of short-term and spatial memory as well as loss of long-t erm memory has been reported in animal models of AD. These losses may be associated with loss of LTP and decreased basal synaptic transmission as discussed above. This loss has been attributed to A levels (soluble and/or insoluble) as well as plaque deposition. Early decreases in LTP and early memory impairment may re sult from actions of soluble A oligomers including neurotoxic protofibrils, while later de ficits may be due to progressive A plaque burdens (Dickey et al. 2003, Trinchese et al. 2004). Several studies have been conducted which have indicated that cogni tive performance can be altered via modulation of A levels and/or clearance of A plaques, as will be elicited in discussion of current immunot herapeutic research in AD. These findings generate excitement over the prospect of developing a preventative or treatment based on these results. At the same time, however, it must be kept in mind that these models lack complete development of all the pathologica l hallmarks of Alzheimers Disease, and complete characterization of behavior in su ch models is simply not possible, making them in many ways incomplete representations of the disease. Mutant APP Mice as Incomplete Models of AD Behaviorally, it is impossi ble to gauge all measures of disturbance in mouse models to the extent that this can be accomplished in humans for obvious reasons. Typical tasks for behavioral assessmen t in mice measure exploration, object
45 recognition, anxiety and fear, and cognitive im pairment with regards to spatial shortterm or long-term memory. Disparities noted during these tasks are thought to correlate to disruptions common to Alzheimers, from decreased mobility and motivation to visuospatial learning defic its to heightened anxiety (Janus and Westaway 2001). However, it is clear that these tasks, while useful for gaining a general picture of mouse behavior, ar e not specific enough to elicit findings comparable to those obtained in humans using a wide variety of interactive tests such as the MMSE, and even in humans diagnosis of AD is convoluted as a result of often insufficient and sometimes contradictor y data noted in such studies. The question also arises: are the behavioral deficits noted in animal models of AD associated with the transgene product vi a the same mechanism noted in humans? Mice have different cellular machinery involv ed in post-translati onal modification of proteins such as APP, and the exact pro cess for neuronal production of hAPP in these models, as well as any subs equent modifications which occur during processing to form A has not fully been explored. Thus, the results of microinjection of such transgenes must be elicited via path ological analysis and comparison with AD pathology. Only the results of transgenicit y can be compared, not the cause (Schwab et al. 2004). In APP transgenics, a number of distin ct differences can be noted immediately in comparison with human pathology. Spontaneous NFT formation has not been noted for any APP model (tau transgenic ity has been necessary to induce such formation) despite the fact that FAD ge nes are known to directly result in AD development with full pathology in humans including NFT formation. The cause of
46 this NFT formation in humans has not been entirely elucidated, but as described above it may be associated with A formation. A itself is thought to act differently within animal models as well. Although this peptide is known to be neurotoxic in humans, perhaps resulting in the wide spread neuronal loss in hippocampal and cortical regions noted in AD, such loss of neurons simply does not occur in many APP transgenic models, and in no model doe s it occur at levels similar to those present in AD (Janus and Westaway 2001, Sc hwab 2004). Inflammatory responses to A once it has deposited into plaques is sign ificantly lessened as well in mice, with little or no complement system activation a nd lessened microglia activation. This can lead to misinterpretation of results in studies such as t hose designed to clear plaques from the nervous system without overactivat ion of the inflammatory system. These differences may result from variations in A pathology in transgenic mice as compared to humans, such as morphologi cal plaque density in APP models, which tend to have less compact cores. This is most apparent in the APP23 model where dense, congophilic plaques have also been noted to be still SDS soluble (Ashe 2001, Schwab 2004). It is apparent that de spite similarities in A pathology between transgenic models and humans with AD, distinctions must be drawn when considering not only the biochemical nature of plaques but their effects on cognitive performance as well. Evaluation of these effects cognitively is made difficult for reasons outlined above. Thus, behavioral and pathological studies in transgenic models must be limited in application to specific and well designed treatment or prevention studies to elicit effects, and the results of such studies must be weighed against the faults intrinsic to
47 them. Nonetheless, the paradigm of repeat ed testing in various models followed by subsequent staged clinical trials is still th e most efficient protocol for gaining insights in to AD therapies in a safe fashion, and thus continued behavioral testing in such models, though found to be lacking, is still esse ntial to AD research.
48 Inflammatory Responses in AD Neuroinflammation is a term which has evolved to specifically describe chronic inflammation within the central nervous system involving glial cell activation and subsequent neurodegeneration, a proce ss for which Alzheimers Disease provides a unique but complete representation (Stre it et al. 2004). In AD, this inflammatory process appears to be induced directly by A as described above, and involves activation of glial cells, both asctocyt es and microglia. Also reactive to A activated T-cells can play a formative role in the progression and severity of neuroinflammation in AD via direct action or through co-stimulation with microglia and production of cytokines. This is one main distinction between neuroinflammation in AD and other chroni c forms of neuroinflammation, which do not typically involve T lymphoc ytes. It may be that disruption of the BBB allows these lymphocytes access to the AD brain, further influencing the inflammatory cascade which occurs progressively throughout the disease. Along with the direct affects of immunoreactive cells, cytokines a nd inflammatory proteins (and in some cases complement proteins expressed by these glial cells and lymphocytes) have widespread impact within regions of the CNS affected by AD.
49 Microglia Activation Microglia, resident phagocytes of the CN S, are thought to be derived from the same cell lineage as macrophages and monocyt es and to provide similar functions within the brain to those provided by these cells in the periphery. In neuroinflammation associated with Alzh eimers, it appears that it is the nonphagocytic capabilities of these cells which play a major role in inflammation associated with the disease, though these ma y only be induced after frustration of phagocytic attempts (Streit 2004). Speci fically, microglia are known to associate with neuritic plaques in AD brains as well as in hAPP transgenic mouse brains, and it has been shown that these cells can be activated by A especially the fibrillar, neurotoxic form of this peptide commonly f ound in neuritic plaques. Other materials associated specifically with compact neuritic plaques are also proposed to be involved in activation of these cells, as A itself does not generally elicit the exaggerated response seen therein (DAndrea et al. 2004). It is thought that these microglia are induced by aggregated A and these other factors to attempt phagocytosis of A aggregations; however, it is also apparent that they are largely unsuccessful at degrading the dense plaques th ey are typically associated with (Streit et al. 2004). In contrast to these findings, in human brains after postmortem analysis as well as in transgenic mice brains taken afte r development of neuritic plaques, it has been shown that microglia are indeed capable of clearing A in rodent brains and that human microglia can be induced to ingest A in vitro (Frautchy et al. 1992). In rodent models, such clearance is signif icantly increased by brain lesioning and by injection of either Lipopolysaccharide (LPS) or A antibody (Smith et al. 1998,
50 DiCarlo et al. 2001, Bard et al. 2000). All of these factors tend to either activate the overall immune and inflammatory systems or to interact with A and microglia directly to facilitate phagocytosis. In vitro studies have also shown that complement system proteins such as the C1q protein may facilitate microglial phagocytosis of A (Webster et al. 2000). Furthermore, postmortem analysis from an A vaccination clinical trial indicated an association between microglia activation and amyloid clearance in humans (Nicoll et al. 2003). These and othe r studies have shown that microglia are quite capable of clearing A in all configurations under the right circumstances, however such clearance is obv iously deficient for some reason in AD. Thus it is either: 1) an overall failure of these cells to maintain clearance of plaques, due perhaps to an inherent dysfunction of these cells 2) an overwhelming effect following increased A production later in life, or 3) the nature of the plaques themselves, which prevents their breakdown. If the first is true, it is possible that Alzheimers is due simply to a failure of microglia to maintain proper balances of A within the brain, while the last hypothesis would seem to indicate that the progressive nature of plaque formation may allow th is process to be too far along prior to microglia activation for these cells to be e ffective as phagocytes after such activation (Streit 2004). Future studies may elicit whic h of these is the true initial failure of microglia, thereby allowing researchers to gain insights into how to alter this process early on. In the present, however, it is the role of activated microglia in the Alzheimers neuroinflammatory cascade which has received the most study in AD research due to the neurotoxic potential of the cells therein.
51 As stated, the major role of microg lia in exacerbating the inflammatory damage noted in AD does not appear to be related to the phagocytic role of these cells. Once activated, microglia often become frustrated due to the persistence of the compact neuritic plaques with which th ey are associated, in much the same way that immune cells of the pe riphery can, during certain au toimmune diseases or in response to mitogens, be activated excessi vely (DAndrea et al. 2004). Once overstimulated and after failed attempts at A clearance, these microglia then proceed to produce a wide array of inflammatory factors including reactive oxygen species, complement proteins, cytokines (such as IL-1, IL-6, and TNF and a number of other factors which may be considered neur otoxic or which may stimulate further A aggregation (Akiyama et al. 2000). Local neuronal degradation causes release of cytoplasmic inflammatory factors leading to further activation of glial cells. Astrocytes activated by these factors, as we ll as by cytokines released from microglia, are attracted to the plaque periphery wher e they begin to produce their own set of inflammatory proteins and cytokines, resu lting in a co-stimulatory effect with neighboring microglia (DAndrea et al. 2004)Inflammatory proteins such as ApoE (produced by subsets of both glial cell types), -ACT (produced by astrocytes), complement proteins, and perhaps other fact ors all associate with neuritic plaques and may be involved not only in promoting further A aggregation, but in further stimulating microglial response via recepto r mediated activation of these cells (Nilsson et al. 2004, Uchihara et al. 1995, DAndrea et al. 200 4). The end result is a localized inflammatory cascade with potenti ally widespread repercussions due to cytokine diffusion.
52 T Cell Activation In a general sense, T cells can be divi ded based on expression of either the CD4 or CD8 co-receptor, whic h primarily segregates th e cells into helper and cytotoxic lines respectively. T helper (Th) cells produce a variety of cytokines when activated which are involved in mediati ng inflammatory and immune responses, while cytotoxic cells are largely involved in recogn ition and destruction of abnormal or aged self cells. Helper cells can be fu rther divided into Th1 and Th2 subsets, each with distinct and often anta gonistic roles in the immune system. The former tends to promote inflammation and cellular based immunity via activation of macrophages and cytotoxic T cells (TC) as well as other immune cells such as neutrophils, while the latter tends to initiate a humoral based res ponse via activation of B cells (Vallejo et al. 2004). Each subtype of T cells requires a diff erent combination of receptor mediated activation to initiate the effects they are known to elic it. Receptor activation also dictates if and when T cells will begin to differentiate and mature. All T cells begin as nave cells and remain so unless they encounter a recognizable antigen properly presented by an antigen presenting cell (APC) for Th cells or by the majority of self cells for TC cells. T cell recognition is dictated by variations in T cell receptor (TCR) subcompontents, which alters in itial ability of a cell to recognize antigens presented by major histocompatibility (MHC) II comple xes of APCs or to recognize peptides presented by MHC I complexes of self cells. Antigens presented by APCs are typically produced after brea kdown of foreign peptides extracellularly via enzymatic degradation or intracellularly after phagocytosis. The pep tides presented by self cells
53 are typically derived from vi ral or intracellular microorganisms which have infected these cells; however aged, cancerous, or otherwise altered cells may also present peptides which promote their own deat h (Vallejo et al. 2004, Crow 2004). Coupling of the TCR with the antigen/MHC II complex or peptide/MHC I complex initiates T cell activation and can result in differentiation into effector T cells capable of varying functions, howeve r further interaction of secondary cell surface proteins is also required for induction of such activities. In many cases, such proteins are not expressed unless secondary activation via cytokine receptors has occurred. Thus, activation of indivi dual T cell subtypes depends not only on recognition of the appropriate antigen / MHC complex, but also on stimulation of regulatory cells to produce appropriate cytokines which then mediate which specific subtype of T cells will be activated. Once fully activated, memory T cells may develop which proliferate and remain pr esent in the body, awa iting reactivation or continuing to act in the presence of persistent stimuli (Vallejo et al. 2004, Chandock et al. 2004). The role of TC cells in Alzheimers neuroi nflammation is thought to be primarily detrimental due to the ability of these cells to induce neuronal or glial apoptosis, and thus Th1 cells activation may also be considered disadvantageous due to subsequent activation of TC cells. At the same time, microglia can be activated by Th1 cells resulting in a positive effect if phagocytosis of A can be induced and a negative effect if phagocytic frustration results, as discussed above. Th2 cell activation may prove to be bene ficial in Alzheimers if A antibody formation by B cells is then induced, enhancing clearance of A as will be discussed later, and other
54 distinct capabilities of uni que T cell subtypes may also be either detrimental or therapeutic based on extent of activ ation and timing of such activation. Microglia and T Cell Co-activation As discussed above, both microglia and T cells are heavily reliant on activation by other cell type s via cytokine receptor activ ation or direct receptor protein interactions with surface proteins of these other cell lines. As phagocytes, microglia are among the few cell types cap able of internaliz ing, digesting, and presenting proteins in association with MH C II receptors in the CNS, thus acting as APCs (Hickey et al. 1988). Specific lines of T cells, largely among the Th1 subtype but also among the Th2 subtype, are capable of inducing microglia to carry out this activity, and in certain cases microglia will remain inactive without such input. Thus, though many other cell lines may be involved in the process, activation of microglia and T cells can occur concurrently and can have a tendency towards exponential propogation under certain conditions. As will be shown, however, the true interactions which have been shown to o ccur between microglia and T cells in the presence of A are convoluted and often difficult to interpret. In 2003, Monsonego et al. showed that mouse microglia primed with INFare capable of serving as A APCs for both A 1-40 and A 1-42, presenting these peptides to A -specific T cells. Both Th1 and Th2 subtypes were shown to be activated by these peptides when presented by microglia. Th1 cells were shown to be activated in the short term, w ith increased expression of IFN(a microglial activator and inhibitor of Th2 prolifera tion) noted. Coactivation of microglia resulted in their expression of TNF(an inflammatory cytokine) and IL-10. As an anti-
55 inflammatory cytokine, IL-10 is known to inhibit TH1 subset activati on, and to inhibit monocyte/macrophage lineage production of NO and a vari ety of cytokines involved in inflammation. In contrast to this, it was shown that microglia were actually stimulated by Th1 cells to increase their production of NO, whic h is a known general T cell inhibitor. Inhibition of iNOS to reduce this production of NO resulted in restoration of Th1 cell prolif eration. Anti-IL-10 and IFNantibody treatments also elucidated that this inhi bition could not be reversed simply by blocking these cytokines without inhibition of iNOS, indicating that the no rmal control of IL-10 over Th1 cell lines was not the pathway of T cell regulation in this case. Thus it appears that the interactions between microglia and Th1 subsets in mice are in many ways distinct from normal interactions between Th1 cells and cells of the monocyte lineage, and that microglia are capable of i nhibiting Th1 cells as a result of Th1 cells stimulating microglia to produce NO. Th is high level of NO and perhaps other mediating factors is thought to then bri ng about apoptosis of Th1 cells (Monsonego et al. 2003). It was also noted that in cultures of microglia with Th2 subset cells, there was an increase in production of IL-4 (promotes TH2 proliferation), IL-10, TNF, and granulocyte-macrophage colony-stimulating f actor (GM-CSF). These T cells did not undergo apoptosis, but were instead activated and induced to proliferate. Thus it was assumed that the Th2 repetoire of cytokines is sufficient to block NO mediated apoptosis of T cells, as has been shown in previous studies, despite an up-regulation of NO which was noted in these cu ltures (Monsonego et al 2003).
56 Based on these in vitro results coupled with those of previous studies, some of the potential interaction which occur between T cells and microglia in Alzheimers may be speculated upon and are summarized in Figure 1. The two major effectors of microglia activation noted in the study conducted by Monsonego et al. (2003) were IFNand GM-CSF, produced by the Th1 and Th2 subsets respectively, which promoted presentation of A by microglia to these cells. Other cytokines such as IL10 were shown to decrease production of NO by microglia in vitro Depending on the type of interactions which occur during and following activation, however, microglia can then act to regulate T cell activity via production of a variety of mediating factors such as NO, as discussed above. Therefor e, with development of an understanding of the interaction which occur between these cell types, it may be possible to alter activation of specific T cell types to promote clearance of by microglia while preventing inflammatory damage due to secondary activities of both T cells and microglia.
3)MHC II APresentationMicrogliaTH1 T CellTH2 T Cell2)AProcessingBy Microglia 3)MHC II APresentation 4)AReactive T Cell Activation (short term):IFN-Produced4)AReactive T Cell ActivationGM-CSF, IL-4 Produced 5)Increased MicrogliaActivation (TNF-, NO)5)Increased MicrogliaActivation (TNF-, NO)1)Stimulation of MicrogliaMHCII Expression (IFN-) 6)Inhibition of TH1 Cells by NO, IL-10 Blocked Inhibition of TH2 Cells Figure 1. IFNprimed Microglia cultured in the presence of A are capable of MHC II associated presentation of A to both Th1 and Th2 cell types. This leads to short term activation of Th1 cells with potential for subsequent inhibition following release of Nitrous Oxide (NO) and IL-10 by microglia (red). Activated Th2 cells are protected from inhibition however, perhaps via autocrine action of GM-CSF and/or IL-4 cytokines (blue). In both instances, microglia remain active and produce inflammatory cytokines such as TNF-as well as NO (green). 57
58 Vaccination in AD Research In the late 1990s, in vitro studies reported that anti bodies specific to various A epitopes can disaggregate am yloid plaques and prevent A fibrilogenesis. It was not long thereafter that re searchers began focusing on the idea of an amyoid vaccination as a treatment or even preventa tive for AD. Promising results in early work in vitro and in vivo with animal models of AD culminated in early phase clinical trials in humans using immunizations. These trials were halted after death of a small number of individuals was reported due to CNS inflammation. Since that time, intense focus has been placed on finding an immunotherapeutic strategy for fighting AD which does not instigate the type of in flammatory response initiated by active A immunization. In Vitro A Vaccination Studies Two of the earliest studies in vitro which contributed to initiating the current immunotherapeutic era of Alzheimers res earch were conducted by Solomon et al. in 1996 and 1997. In the former study, it was noted that monoclonal antibodies recognizing either aa1-28 or aa8-17 residues of the A peptide are capable of inhibiting normal in vitro aggregation of A (Solomon et al. 1996). In the latter study it was shown that the antibodies r ecognizing the aa1-28 residue of A were actually
59 capable of disaggreg ating insoluble A fibrils into soluble A (Solomon et al. 1997). In 1998, Frenkel et al. found that these same antibodies specifically bind to aa3-6 (sequence EFRH), and it was inferred from th ese results that this sequence must be involved in the A aggregation process (Frenkel et al 1998). A year later, this group published findings which indicated that not all A antibodies recognizing the region of the peptide containing this residue were as effective at binding to it, and that low binding affinitity antibodies we re not capable of preventing A aggregation (Frenkel et al. 1999). From the results of these four in vitro studies, it was deduced that carefully selected antibodies to had the potential to block or even reverse amyloid aggregation, paving the way for future in vivo studies in A vaccination. Innate Immune Responses to A On a cellular level, the respons e of the immune system to A is three-fold. B cells can be stimulated to produce A antibodies, glial cells can be stimulated to produce a variety of inflammatory molecules and in the case of microglia to attempt phagocytosis of A plaques, and T cells can be s timulated to produce an array of modulatory cytokines as disc ussed above. Prior to A vaccination, these responses are present in varying levels. Localized, high levels of glial cell activation are common near A plaques (D'Andrea 2004), and aged AD patients are significantly more likely to exhibit st rong T cell reactivity to A (Monsonego et al. 2003). Finally, A antibodies are known to be produced in AD and may be involved in triggering microglia response to plaques, thus f acilitating phagocytos is (Nath et al. 2003, Wilcock et al. 2004). These reactions to A are also evoked in varying degrees
60 following A vaccination (Nicoll et al. 2003). The current goal of immunotherapeutic research in AD is to establish a means of creating a balanced response incorporating each of these respons es only to the extent necessary to be beneficial. Active A Immunization in Animal Models A wide variety of active A -based immunization studies have been reported on since the first reports in 1999. Retrospe ctively, with the know ledge we now have of consequences of standard A immunization in humans, it is perhaps most simple to divide these studies into three main categor ies. The first consis ts of those studies which provide a general understanding of the positive pathological consequences of A immunization, which are based largely on amyloid plaque clearance and reduction in plaque associated pathology. The second consists of those studies which provide insight into how side-effect s from such administrations in humans could be anticipated by both characteriz ing the neuroinflammatory response elicited by these or which provide alternative, less dangerous modes of eliciting an anti-A immune response. Finally, studies pr oviding behavioral characteriza tion of treated animals as potentially indicative of th e cognitive protection that ma y or may not result after administration of an A vaccine. Pathological Benefits of Active A Immunization A number of studies focused on elic iting the pathological benefits of A vaccination while not reporting results showi ng the potential for negative impacts or the behavioral consequences that result from such vaccination. In 1999, Schenk et al.
61 reported results from the first attempted active A immunization study in a mouse model of AD. Injections of A 1-42 were performed monthly fo r 11 months starting at 2 months in PDAPP mice. A separate group of mice was injected with SAP peptide, a separate component common to amyloid pl aques. Initial in jections contained complete Freunds adjuvant, while each additional injection contained incomplete Freunds adjuvant, a protocol used in the majority of future active A vaccination studies. Titres of antibodies to the injected antigen were measured at higher than 1:10,000 for most A injected mice and between 1:1,000 and 1:10,000 for SAP injected mice. The mice were sacrificed at 13 months, and it was found that A immunized mice were almost completely protected against development of A plaques in the CNS, while SAP immunized mi ce showed statistically similar levels of plaques compared to non-treated controls. This latter data would seem to indicate that an immune response elicited towards any A plaque component is not necessarily capable of cleari ng these plaques. Almost complete protection against development of dystrophic neuritis and re duction of astrocytosis and microglia activation was also noted in A immunized mice (Schenk et al. 1999). This group also performed additional expe riments to determine if reduction of plaque loads could be accomplished in aged mice. 11-month-old female PDAPP mice were again immunized with A 1-42. Sacrifice of the mice occurred at either 15 or 18 months. At both time points, A burdens were significantly reduced compared to PBS treated transgenic controls. Re duction in both diffuse and compact plaque counts were attributed to cl earance of already formed de posits. Existing plaques were
62 noted to be coated in IgG antibodies to A 1-42, indicating a direct mechanism of action for these antibodies within the CNS. Lack of plaques in regions such as the entorhinal cortex were attributed to bl ocked formation of plaques typically not deposited until later time points in PDAPP mice. A containing cells thought to be microglia or monocytes were present in areas normally containing high plaque counts at aged timepoints, indicating an activati on of these cells to phagocytize existing A deposits, perhaps via antibody receptor mediat ed activation. Astrocytosis was noted to be reduced in these A treated mice, as were other neuritic plaque pathologies. This second round of results would seem to indicate that while A vaccination does show potential efficacy as a treatment due to partial clearance of plaques at an aged time point in transgenic mice and prevented formation of additional plaques, clearly the most benefit stands to be gained from earlier, prophylactic administration of A (Schenk et al. 1999). In 2000, Weiner et al. found that intranasal administration of synthetic A 1-40 peptide, as opposed to injections of A 1-42, between 5 and 12 months in PDAPP mice resulted in significantly decreased ne uritic plaque burden and overall A 1-42 levels compared to controls treated with myelin based protein or standard controls. Decreased microglia and astrocyte activati on, decreased neuritic dystrophy as seen in Schenk et al. (1999), and decreased reactiv e mononuclear cell activation was also reported. Importantly, this study demons trated the efficacy of intranasal A vaccination for the first ti me. (Weiner et al. 2000).
63 In 2001, Das et al. showed that effectiveness of A 1-42 immunization correlates inversely to am yloid deposition levels in Tg2576 mice. Mice with minimal, modest, and significant A deposition (7-8, 10-11, and 18 months) were immunized with the peptide, a nd reductions in cerebral A were noted to be highest in mice immunized at younger timepoints. It was also noted that A 1-42 was reduced at higher levels than A 1-40. This study clearly indicat ed the importance of early vaccination as a preventative/trea tment for AD (Das et al. 2001). Further expanding on the potential benefits of A vaccination, Oddo et al. (2004) reported for the first time that A immunization results in clearance of nonhyperphosphorylated (early) but not hyper phosphorylated tau aggregates in a APPsw+PS1+Tau transgenic model, along with the normal reduction in A pathology expected after such immunizati on. It was also noted that A pathology was necessary for formation of tau pathology, and clearance of the former via non-immunological methods resulted directly in clearanc e of the latter (Oddo et al. 2004). Lemere et al. (2003) reported that the typical reduction in A pathology associated with A immunization such as that note d in the above three studies was coupled with a dramatic increase in A levels in the serum of APP transgenic mice. It was suggested that the ma in mechanism of action of A immunization is thus stimulation of A antibodies largely restricted to the blood, association of these plasma antibodies with A and clearance of CNS anti body/antigen complexes into the vasculature, resu lting in reduced A levels in the CNS (Lemere et al. 2003).
64 To better understand the time course for immune system reactivity to A Wilcock et al. (2001) delved further into the specific immune responses elicited by various numbers of A 1-42 vaccinations, reporting that anti-A antibody titers remained low after only 3 innoculations in APPsw+PS1 mice but were significantly higher after nine inoculati ons. After 5 innoculations there was a high level of microglia activation which tapered off by nine injections. A correlation was found between microglia activation and reducti ons in congo red stai ning, suggesting that microglia may be involved in clearing compact plaques in these mice. Thus this study and that by Lemere et al. (2003) pr ovided two distinct a nd probable modes of action for anti-A antibodies in clearing amyloid from the CNS (Wilcock et al. 2001). In further examining the potential for microglial phagocytocis of A plaques, it was shown that A 1-42 immunization of Tg2576 mice crossed with FcR -/(Fc receptor knock-out) mice resulted in unimpaired reduction in A levels according to Das et al. (2003) compared to immunizati on of Tg2576, Fc receptor-sufficient mice. This would seem to indicate that microgl ial, FcR-mediated phagocytosis is not required for effectiveness of the A vaccination and is not the main mechanism for A clearance (Das et al. 2003a). Das et al. (2003) also showed that B and T cell responses to A 1-42 immunization vary greatly based on specific MHC class II expression patterns, indicati ng that some combinations of expression among these molecules result in better presentation of the A peptide to these cells and once again emphasized the importance of these immune cells in determining immune response to A (Das et al. 2003b).
65 Along with an understanding of the react ions of the immune system of APP transgenic mice to A inoculation and the mechanisms of A clearance, it is also necessary that the impact of expression of this human transgene on the response of these animals be understood. In compari ng the immune response of APP transgenics and non-transgenics to A inoculations, Monsonego et al. (2001) reported that APP transgenic mice immunized with either A 1-40 or A 1-42 displayed hyporesponsiveness to these peptides imm unologically. This was thought to be associated with a lack of T cell response in these animals, perhaps due to long-term presence of A peptides in transgenic mouse brains from an early time point. It was suggested that if the T cell response coul d be raised during the vaccination period, such immunotherapy would prove more successful (Monsonego et al. 2001). Further emphasizing the hyporesponsivene ss of the the APP transgenic mouse immune system to A Dickey et al. (2001) reported that three vaccinations with A 142 was necessary to elicit a significant immune response in APPsw+PS1 transgenic mice based on IgG levels. Subsequent admi nistrations of the vaccine resulted in moderate immune responses through 5 months It was also noted that the highest immune response was to the A N terminus, and it was suggested that because the main humoral response was IgG1and IgG2-based, a TH2 T cell subset activation was likely to be involved in generating this res ponse. Importantly, it was also noted that the immunized mice displayed T-cell s stimulated specifically by the A peptide, again emphasizing the potentia l for T-cell activation to re gulate the immune response to A vaccination (Dickey et al. 2001). Both of these previous two studies provided
66 insight into the importance of proper regulation of T cell activation following immunotherapy. Additional support for the impor tant regulatory role of TH2 cells following A immunization in transgenic mice was provi ded by Town et al. (2002). It was reported that a bias towards this T cell subset in both C57BL/6 non-transgenics and APP mice inoculated with A 1-42 was found. Importantly, th is bias was not noted following A 1-40 immunization, indicating an alte red immune response based on antigen selection (T own et al. 2002). In 2004, Li et al. repor ted that while anti-A levels from A immunized APP transgenic mice are masked in ELISA measurem ents if not first treated with a low-pH solution to dissociate anti-A /A complexes, these levels are still significantly lower than those of non-transgenic mice of the same background following immunization even after such treatment. Only when A was administered associated with papillomavirus virus-like particle s (VLPs) were levels of anti-A comparable in APP and non-transgenic mice, indica ting that self-tolerance to A in APP mice results in hypoimmune responsiveness in these mice after normal active A immunization (Li et al. 2004). Based on findings that aged Caribbean Vervets, which are primates, often develop cerebral A plaques, Lemere et al. (2004) vaccinated a number of these animals for 10 months with a cocktail of A 1-40 and A 1-42 in a 3:1 ratio. No plaques were found in the four surviving immunized animals, while 11 of 13 controls showed plaque development. It was also noted that rises in plasma A were associated with
67 decreases in brain A levels in immunized anim als, and antibodies to A were noted in both the plasma and CNS. Th is study was the first to report A clearance following A vaccination in a non-human, non-transgenic mammal which normally develops A plaques with aging (Lemere et al. 2004). The above studies elucidated the potential for active A immunization in AD animal models to elicit an immune re sponse at numerous timepoints in the development of amyloid pathology, in cluding production of various anti-A antibody isoforms and stimulation of cellular-based immunity mediated by T lymphocytes. It was demonstrated that such immune system activation leads to varying levels of clearance of A pathology from the CNS based on background strain interactions, antigen selection, antigen delivery, adjuvant selection, and age. This generated excitement at the possibility of develo ping a vaccination-based therapeutic for Alzheimers disease. Still left to be discovered by additional studies, however, was the potential for such immunization to also lead to devastating consequences in the AD brain, as well as the actual cogni tive benefit to be gained by A vaccination. Deleterious Effects of Active A Immunization and Alternative Active Immunization Approaches. Two major studies in A vaccination focused largely on its negative consequences. Su et al. (1999) reported that low-level infusions of soluble A 1-40 intravascularly in 3 month old male rats over the course of 2 weeks resulted in blood vessel damage including gross pulmonary hemorrhage, disruption of the BBB, and activation of microglia and astrocytes in the CNS. While these physiological
68 consequences were not coupled with impairment in water maze, they did provide evidence that even short, lo w-level administrations of A can result in damage to both the vasculature and perhaps the CNS vi a initiation of widespread inflammation. This again emphasized the importance of determining a method for vaccination which does not instigate such damage (Su et al. 1999). Furlan et al. (2003) further elucidat ed our understanding of the potential for neuroinflammatory damage resulting from A vaccination. Young C57BL/6 mice immunized with A 1-42 were shown to display infla mmatory foci in the brain and spinal cord associated with the vasculature of these areas. These foci were noted to contain microphages, T and B cells, and immunoglobulins. This response was attributed to Th1 subset me diation of inflammatory even ts coupled with high anti-A levels. Importantly, these re sults were noted only when a co-administration of A and pertussis toxin (which elicits autoimm une processes) was performed; thus the extent to which the toxin c ontributed to development of these foci is important in interpretation of these resu lts (Furlan et al. 2003). It was apparent following the Su et al. 1999 study and based on results of phase II clinical trials in humans that th ere are dangers associated with simple A vaccinations mainly associ ated with increased neuroinflammation, and numerous studies which followed focused largely on searching for more unique modes of immune system activation which do not invoke such reactions. In 2000, Frenkel et al. attempted active A immunization in both BALB/c mi ce and guinea pigs via either intranasal administration and s.c. or i. p. injection of a phage displaying the A 3-6
69 peptide residue EFRH. Seven days followi ng injection, serum levels of IgG reactive against both wt-phage and A were measured for three injections spaced 14 days apart. Mice injected i.p. with Freunds ad juvant developed large titres of antibodies which increased after each in jection. Specific IgG and IgA antibodies were noted 7 days after only one booster inje ction of phages. Guinea pigs injected s.c. were noted to develop very high titres of anti-A antibodies following the same injection protocol. The guinea pig EFRH sequence is homologous to that of humans, whereas in mice the sequence is EFGH, so the eff ectiveness of the phage administration in both species emphasized the potential of this vaccination for effectiveness in humans. It was shown that in both animal groups, sera containing antibodies produced was able to bind specifically to the A peptide, and that this sera protected against A neurotoxicity in vitro Also noted was the ability of this sera to disrupt A fibril formation in vitro (Frenkel et al. 2000) In 2003, Frenkel et al. again immunized mice with EFRH presenting phages, this time in the APP transgenic model. This immunotherapeutic approach was show n to again dramatically reduce A pathology in these mice. Importantly, it was noted that two mice receiving the phage did not develop high titers of anti-A antibodies, indicating less than perfect penetrance (Frenkel 2003). In 2001, Sigurdsson et al. were able to show that immunization with a nontoxic, soluble A homologue results in a simila r reduction in amyloid pathology to those noted after standard A administration in APP transgenic mice. Along with reductions in soluble and insoluble A these mice also exhibited decreased microglia
70 activation, thought to be associated with an overall reducti on in inflammation levels (Sigurdsson et al. 2001). In 2004, Sigurdss on et al. carried out active immunization in APP mice with this same nonfibril logenic, nontoxic derivative of A K6A 1-30, which was known to induce A clearance in similar levels to A 1-42 treatments. A primarily IgM response resulted in these mice, and A pathology was significantly reduced in inverse correlation to levels of th ese antibodies. IgM is not able to cross the BBB, so it was inferred that A clearance resulted from these antibodies acting as a peripheral sink for these peptides. In 2002, a study conducted by McLaurin et al. showed that antibodies recognizing the 4-10 residues of A 1-42 are capable of inhibiting fibrillogenesis of A in TgCRND8 transgenic mice without i nvoking activation of the inflammatory system. This antibody was shown to target protofibrillar A which is thought to be important in formation of A plaques, and provided an important mechanism for blockage of said formation. It was furt her reported that th e immunization which led to this antibody formation caused a bi ased T cell response, downregulating inflammatory events associated with T h1 cell activation while upregulating some aspects of the humoral responses elicited by Th2 cells (McLaurin et al. 2002). Koller et al. (2004) found that immunization of APP transgenic mice with A 1-42 coupled with DnaK, a heat shock pr otein 70 analogue a nd strong adjuvant, resulted in measurable levels of anti-A antibodies after only one innoculation with maximal titers of these antibodies present after the first booster inoculation. These levels were still significantly lower than those achieved in non-transgenic controls
71 receiving the same vaccination. This em phasized the importance of adjuvant in determining the immune response elicite d by active immunization, and provided a method for reducing the total number/amount of A administration necessary to elicit an appropriate immune response. Importa ntly, however, these immunizations were all administered by 3 months of age, and as a result no affect was seen on A levels and plaque burdens after sacr ifice at an aged time point suggesting that prophylactic immunization occurring too early in life, without continua tion, may prove ineffectual. (Koller et al. 2004). Hara et al. (2004) found that oral administration of a recombinant adenoassociated virus vector containing cDNA for A resulted in expression and secretion of both A 1-43 and A 1-21 within the intestine of hAPP transgenic mice. This resulted in anti-A antibody production through 6 months and significant decreases in CNS A levels. These decreases were not associ ated with any notable inflammation or T cell response (Hara et al. 2004). The work of Su et al. (1999), Furlan et al. (2003), and results from A vaccination clinical trials served as a wa rning for researchers, demanding that new understanding of the true consequences of A immunization should be gained before any further clinical trials are conducted. Many of the studies which followed thus searched to explore alternative A vaccination methodologies to reduce neuroinflammation and a cellular-based im mune response. Even these studies, however, are not fully sufficient to gain true understanding of the potential effects of A vaccination is AD models. Yet to be determined is the effects of A vaccination
72 on cognitive behavior, as most of the st udies mentioned above did not seek to correlate any pathological findings with cognitive impacts. Behavioral Characterization in Active A Immunotherapy Due to the potential for neuroinfla mmatory damage associated with A immunization, it was necessary to establish th at cognitive protection could be elicited by such treatment despite these potential ne gative consequences. Janus et al. (2000) reported that vaccination of young TgCRND8 mice with A 1-42 peptide in complete Freunds adjuvant resulted not only in plaque load reducti ons, but also in decreased impairment in Morris water maze acquisi tion, though these mice still did not perform at the same level as non-tra nsgenic controls. 3 cohorts of mice were immunized at 6, 8, 12, 16, and 20 weeks of age with either A 1-42 peptide or islet-associated polypeptide (IAPP), both in -pleated sheet formation. This control antigen was selected based on its similarity to A but also on the fact that it only causes amyloidosis outside of the CNS. Both peptides were shown to induce antibody production by 13 weeks when the first group of injected animals underwent sacrifice. A second group of animals were sacrificed at 23 weeks and showed 2-3 fold increases in antibody productions by this time point. These antibodie s were shown to strongly bind to dense A plaques but to only weakly inte ract with diffuse, non-fibrillar deposits of A Also, these antibodies did not appear to be reactive to APP, as normal neurons were not detected by the antibody. Impacts of immunization on A pathology were noted to be similar to thos e from previous studies with regards to density and size changes of dense-cored pl aques, however at both 13 and 25 weeks,
73 no changes in formic-acid-extractable (soluble) A were noted, perhaps because of the specificity of anti-A antibodies to the -pleated sheet formation of deposited A Behavioral testing of mice by Janus et al. (2000) consisted of longitudinal testing (at 11, 15, 19, and 23 weeks of age) in a reference memory version of Morris water maze, with the submerged escape platfo rm in a different location at each test age. A 1-42 immunized transgenic mice were sh own to perform significantly better than non-immunized transgenics, however pe rformance was not improved to the level of non-transgenic mice. A immunized mice were also shown to significantly improve in task performance after repeat ed reversal testing, indicating a learning affect in these mice. In probe trial te sting following Morris water maze testing, no significant differences were noted among groups. However, this was attributed to the close temporal proximaty of task administration (30 minutes) to final day Morris water maze administration. Controls studies indicated that no performance changes were noted for non-transgenic mice following any form of PBS, A or IAPP injection with or without adj uvant (Janus et al. 2000). Morgan et al. (2000) and Arendash et al. 2001 found that APPsw+PS1 mice immunized monthly with A 1-42 in Freunds adjuvant (min eral oil after 3 months) showed significant protection against impairment in the RAWM task at an aged time point compared to controls receiving a control vaccine of keyhole limpet haemocyanin. Behavioral assessment of double transgenic mice by Morgan et al. (2000) and Arendash et al. (2001) consisted of pre-imm unization administration of a 6-week battery of sensorimotor and cogniti ve based tasks between 5 and 7 months of age followed post-immunization by Radial Arm Water Maze testing at 11.5 and 15.5
74 months and balance beam, string suspension, open maze, and Y-maze tasks at 16 months. No differences were not ed among groups in pre-immunization performances, except in balance beam perfor mance in which transgenic animals were impaired overall. In RAWM testing at 11.5 months, transgenic animals performed at the same level as non-transgenic controls Results from repeated testing at 15.5 months demonstrated that A immunized mice reached sim ilar levels of performance to non-transgenics by final trials of testi ng, however it was demonstrated that these mice required more trials to reach this level of performance than these nontransgenics. Working memory was t hus shown to be greatly improved in APPsw+PS1 mice following repeated A 1-42 immunizations over an 8 month period (Morgan et al. 2000). Mice immunized with A both transgenic and wild -type, showed high antiA serum activity while non-A immunized animals showed no such activity. Modest to no reduction in A plaque burdens in the front al cortex and hippocampus were noted post-immunization. Th ough no measures of soluble A levels were noted in these studies, it was suggested that reduced levels of soluble, toxic A oligomers may have contributed to the lack of cognitive impairment present in these A immunized animals, despite their high levels of A pathology remaining in the CNS. Based on correlation analysis of beha vioral task performances and A burden measures, a number of correlations were evident in the same studys animals (Arendash et al. 2001). While ther e was no correlation between total A immunostaining (compact and diffuse pla ques) and RAWM performance, T4 error
75 measures did correlate with both A 1-40 immunostaining and congo red staining in frontal cortex as well as with immunost aining in both the hippocampal CA1 region and dentate gyrus. However, as A reduction was modest at best it is likely the impact of vaccination on soluble A constructs in the CNS (neutralization and/or clearance by anti-A antibodies) that played a larg er role in providing cognitive benefit (Morgan et al. 2000, Arendash et al. 2001). Jensen et al. (2005) vacc inated APP+PS1 mice with A 1-42 in Freunds adjuvant (complete for initial injection, incomplete for monthly booster injections) from 2-16.5 months of age and determined the effect of these injections at 4.5-6 months and 15-16.5 months longitudinally via administration of a 6 week behavioral test battery at each time point. The study c ohort was divided into control transgenic, vaccinated transgenic, and non-transgenic mice. At the 4.5-6 month time point, both transgenic groups displayed increased open field activity, whil e no group differences were noted at the 15-16.5 month time point Only young (early time point) treatment mice were impaired at the balance beam task, however this impairment was not present at the later time point and no impairment was present for string agility performance at any time point. Comparing early and late time point performances, a significant increase in percent time spent in open arms during the elevated plus maze task in control mice was coupled with a si gnificant decrease in the same measure for vaccinated mice, perhaps indicating an anxioly tic effect in the former and an anxiety increase in the former. Consistent with open field results, vaccinated mice also showed significantly higher plus maze and Y-maze arm entries at the young time point. Only Y-maze entries remained highe r at the aged time point, however. No
76 differences among groups were noted among percent alternations in Y-maze or circular platform acquisition, however significance was found for each additional cognitive based task administered. In Morris water maze acquisition, vaccinated mice were protected against impairment noted in controls at both time points. This same trend was noticed for young mice te sted in Morris water maze retention, however both transgenic groups were impaired compared to non-transgenics at the later testing time point. No impairment at the young time point was noticed in platform recognition, whereas vaccinated mi ce showed protection against impairment exhibited by control mice at th e later time point in this task. RAWM overall and last block performances were reported for thes e mice as well, and for both measures at both time points control mice showed performa nce impairment. In contrast, only in overall RAWM performance at the young time point were vaccinated mice impaired; in last block, young time point performance and in both last block and overall aged time point performance vaccinated mice s howed were protected against this impairment. Discriminant function analys is using direct entry method revealed cognitive protection in vaccinated animals at the young time point. Stepwise-forward analysis showed similar results for all measures, as well as cognitive measures, at both time points. At 17 months of age, mice in the stud y conducted by Jensen et al. (2005) were sacrificed, and A histopathology was performed on mouse brain sections. No significant reduction in A immunostaining (total A ) or Congo Red staining (compact A ) burdens were noticed in v accinated mice overall, though one vaccinated mouse was noted to have signi ficant reductions of up to 86% in A levels.
77 Correlation analysis between the 8 A deposition measures noted and nine cognitive measures taken at the aged time point in these mice revealed no correlations between A histopathology and cognitive performance, in contrast to previous findings by Arendash et al. (2001) in which such correlation was ma de for certain measures despite a lesser duration of A administration. Still, factor analysis results from Jensen et al. (2005) did suggest an existing relationship between A deposition and numerous cognitive measures, despite a lack of significant correlation therein. More importantly, perhaps, is that this study de monstrated once again th e potential for longterm, active A vaccination to result in cognitiv e protection in an AD transgenic mouse model without significant overall reductions in deposited A levels (Jensen et al. 2005). Austin et al. (2003) showed that 4 biweekly administrations of A1-42 to APPsw+PS1 transgenic mice at 16-18 mont hs of age did not result in improved cognitive performance in RAWM or other tasks between 18 and 20 months compared to cognitively intact non-transg enic mice. Serologic anti-A titers in these aged mice were measured at an approximately 10-fold lower level compared to the previous study by Morgan et al. (2000) indicating a possible mechanism for no improvement in cognition. Prior to immunization, RA WM and platform recognition tasks were administered, and it was noted that impai rment of APPsw+PS1 mice was only present in trial 5 performance of RAWM as we ll as overall in platform recognition performance. Following immunizatio ns, both immunized and non-immunized animals showed significant impairment co mpared to non-transgenics over a number
78 of RAWM measures. Furthermore, bot h groups showed no preversus posttreatment differences in RAWM performances for any trial or block combination, and in some instances these groups showed signi ficant decline in performance measures. In contrast, non-transgenic animals showed significantly better trial 4 performances compared to testing results at a younger time point as well as stabilization of trial 5 errors at a significantly lower level compar ed to the high stabi lization levels of transgenic groups. Impairment also conti nued in platform recognition performance, with significantly poorer performances disp layed by both transgenic groups compared to non-transgenics. Based on these findings, it seems that short-term immunization was not effective in elici ting a strong immune response in these aged APPsw+PS1 animals resulting in no positive impact on cogni tive performance. If these results can be translated to human applic ation, it would seem that A vaccination will most likely be most effective in AD if admini stration is begun at early time points in disease development, resulting in increased administration periods (Austin et al. 2003). Behavioral evaluation of vaccinated mice was also performed by Sigurdsson et al. (2004) following K6A 1-30 immunization. Immunized mice performed significantly better than transgenic contro ls in the radial arm maze task, in many measures performing at a similar level to wild -type controls. It is clear based on these results that cognitive protection is possi ble using a non-toxic, non-T cell activating A derivative such as K6A 1-30 as an antigen, indicating possibilities for future vaccinations with reduced negative si de-effects (Sigurdsson et al. 2004).
79 The above behavioral based A immunization studies directly tie the pathological impacts of such vaccinations with cognitive benefits in AD transgenic mice. The potential for A vaccination as a therapeutic based on beneficial cognitive impacts following reduced CNS A burdens is establishe d (Janus et al. 2000), however, the potential of long-term vaccina tion to result in such benefit without dramatic decreases in A pathology is also reported (M organ et al. 2000, Arendash et al. 2001). Thus, the positive effects of A vaccination on behavior in APP transgenic mice appear to be mediated not simply by clearance of A deposits, but by some other mechanism such as cleara nce/neutralization of soluble A constructs. Findings by Austin et al. (2003) establish the necessi ty for long-term vaccine administrations to achieve cognitive benefit, and it is apparent based on the overall results of the above studies that application of A vaccination as a preventative as opposed to a treatment may result in increased efficac y. Finally, findings reported by Sigurdsson et al. (2004) provide suppor t for the concept of alte rnative immunotherapeutic approaches to A vaccination. Such results furt her reinforce the importance of developing AD therapeutics targeting A in the CNS, and promote A immunotherapy as a field of continued importance in AD research. Active A Immunization in Humans The collective results of Phase I and Phase IIa Clinical Trials conducted by Elan Pharmaceuticals in the U.S. and U.K. represent a disappointment which could only follow the highest level of anticipation. Protocols consisted of administration of A 1-42 (An1792) to humans in initial and booster shots for Phase I and, with repeated
80 shots planned over a 12 month period for Phas e II trials. In January 2002, reports of nerve inflammation in four out of 298 patie nts in Phase II that had begun receiving active vaccination led to furt her administrations being halted. Phase II was halted, in fact, after only the initial and one booster injection (after 2 sh ots separated by one month). 11 additional patients were reported to have developed similar symptomologies a month later, and one death was attributed to affects of the vaccine. Specifically, post-mortem analysis indicated neuroinflammation associated with T cell activation as a possible cause of death. In the end, 18 individuals, or approximately 6%, were found to develop symptoms suggesting subacute meningoencephalitis (Orgogozo et al. 2003). A number of reports of findings from these studies were still published, however. In October, 2002, Hock et al. reported on the presence of anti-A antibodies in the sera of 100% of 24 patients having received one primary and one booster injection i.m. of A vaccine containing QS-21 as an adjuvant. These antibodies were not found to be cross-reactive with APP or any nonA derivatives. Sera taken from these patients were also found to bind to A deposits in post-mortem brain tissues of AD and CAA patients. CSF samples taken from 6 patients were also analyzed for A reactivity, and it was reported that 4 out of 6 samples displayed positive results. Importantly, the presence of anti-A in the CSF was noted in patients with and without BBB dysfunction. It wa s suggested that a combination of low A antibodies and intact BBBs resu lted in prevention of these antibodies from entering the CSF in the two negatively tested patients. A titers in serum samples ranged from 1:50 to 1: 10,000, while in the CSF the highest measurement was 1:1,000 (Hock et al. 2002).
81 In March 2003, the first case report from post-mortem analysis of the lone mortality of the A 1-42 vaccination trial was published by Nicoll et al. Neuropathological analysis of this womans brain revealed uncharacteristically low levels of plaques in various areas of th e neocortex compared to seven unimunized AD patient brains. It was noted that these areas still maintained levels of NFTs and CAA common to non-treated AD patients, but that dystrophic neurite levels and astrocytic activations typically associated with am yloid plaques were significantly reduced. Involvement of microglia in A clearance was suggested by co-associated with A immunoreactivity. Perhaps most important, however, were findings of T-lymphocyte meningoencephalitis, indicating th e potential mechanism by which neuroinflammation occurred in patients receiving the A vaccine (Nicoll et al. 2003). This provides insight into the major disparity between immune reactivity to A 1-42 in mice and humans. In mice, evidence suggests that hyppoimmune responses to A may be due to low levels of T cell activation; however it is clear based on the above findings that heightened T cell response is present in at least some humans receiving the vaccine. In May of 2003, Hock et al. reported on some of the positive cognitive effects of A vaccination in humans. 19 of the 28 patients administered the vaccine in Phase I trials showed decreased decline in cogni tive functions compared to non-treated individuals. These patients were also noted to be the individuals who developed antibodies against A Cognitive stability was determin ed via administration of tests such as the MMSE and Visual Paired Asso ciates Test of delay recall from the Wechsler Memory Scale. Declines in activ ities of daily living were also diminished
82 in treated individuals as dete rmined from Disability Assessment for Dimentia results. Importantly, overall results from the 300 pa tients tested in Phase II AN1792 clinical trials did not parallel these findings, w ith no significant differences in cognitive decline noted between groups. More info rmation on the time course for antibody development was also relayed in this publica tion. It was reported that both IgG and IgM levels against aggregated A 1-42 increased significantly in the first month following initial A inoculation, and that maximal titers were reached by a month after the booster injection (Hock et al. 2003). The various findings resulting from v accination studies in humans reinforce the potential for variations between mouse model systems and human systems to result in imperfect translation of results from one to the other. Important to consider, however, is the evidence shown that cognitive benefit and A clearance can result from A administration in humans. Thus, while the risks associated with standard A immunization outweigh potential benefits, it is clear that A immunotherapeutic approaches to AD warrant further investigation. Passive Anti-A Immunization in Animal Models Pathologic Effects of Passive Anti-A Immunization Due to its neurotoxic properties, it is clear that the concept of A inoculations should elicit some level of concer n. Passive administration of A antibodies allows for introduction of these molecu les without any need for A inoculation or for immune system activations associated w ith antibody production. The major methods proposed for vaccine efficacy ar e direct solublization of A peripheral sequestration
83 of A by circulating antibodies, and anti body facilitated phagocytosis of A by microlgia. Thus, the potential for A clearance based solely on the administration of A antibodies is viable. Bard et al. (2000) performed the first study involving passive immunization in a mouse model of AD. It was shown that peripherally administered antibodies to A are capable of entering the CNS of PDAPP mice to induce clearance of amyloid plaques by as much as 93%. Tw o of a set of four monoclonal antibodies along with polyclonal antibodies from A 1-42 immunized mice were shown to induce this effect in 8-10 month-old mice. No T-cell reactivity was seen to be evoked as a result of these procedures. Analysis of microglia showed numerous A containing vesicles, indicating a major phagocytic mechan ism of plaque clearance. Internalized A was shown to be degraded by these ce lls, and the internalization process was shown to be Fc receptor mediated. Additi onal work with 13 month old mice showed that modest reduction in existing small plaque counts could be seen as early as 3 days after treatment, while 35 days after treatment up to 60% of small and diffuse plaques had been cleared. Large, compact plaques a ppeared to resist clearance. Importantly, no correlation existed between binding affinity of antibodies to soluble A and plaque clearing capabilities of these antibodies, indicating that affinity to deposited A may play the larger role (Bard et al. 2000). DeMattos et al. (2001) also showed that anti-A antibodies are able to prevent A deposition in PDAPP mice, however it was shown that the antibodies used in this study were not capable of binding to existing A deposits. It was found that these antibodies sequestered large quantities of A in the serum, indicating that in this case
84 the ability of these molecules to act as a peripheral A sink was the driving force behind protection against development of A pathology. CSF levels of A were also found to increase, which was attributed to increasing levels of solublized A (DeMattos et al. 2001). Lambert et al. (2001) showed that antibodies generated in vivo against inoculated A 1-42 oligomers were capable of binding preferentially to A constructs in vitro. These antibodies were al so shown to block toxicity of A oligomers in vitro As mentioned earlier, a study conducted by McLaurin et al. in 2002 showed that antibodies recognizing the 4-10 residues of A 1-42 inhibit fibrillogenesis of A in TgCRND8 mice. Again, the antibody was shown to target protofibrillar A an important A plaque precursor. These studi es provide a clear mechanism for prevention of plaque formation via anti-A administration which may potentially occur without modulation of A levels. (McLaurin et al. 2002). In 2002, Chauhan and Siegel reported th at a single ICV in jection of anti-A antibodies led to reduced pla que loads in the cerebrum and decreased astrocytosis in 10 month-old Tg2576 mice. In 2003, this same group reported that this same treatment was associated with reduction in activated microglia levels in areas surrounding existing Congophilic plaquesthe study, however, only involved qualitative assessments of these levels. N onetheless, it was further reported that no signs of microhemorrhage were found in these mice (Chauhan and Siegel 2003). Wilcock et al. (2004) also showed that si gnificant reduction in amyloid was possible in Tg2576 mice, this time in 19 month-old animals. It was shown in this study that
85 microglia activation between 37 days after injec tion correlated to this clearance, but that other mechanisms were involved as clearance began as early as 24 hours postinfusion. Finally, a study by Lombardo et al (2003) in aged PDAPP mice indicated that, along with clearance of A as early as 4 days following a single administration of anti-A antibodies, neuritic altera tions attributed to neurotoxic properties of A were reversed as a result of this ad ministration. This again emphasized the importance of achieving A clearance from the CNS in AD. Pfeifer et al. (2002) produced the firs t results indicating the possibility for negative consequences to passive anti-A administration. 21 month-old APP23 mice showed reduced amyloid pathology followi ng such treatment, mainly due to clearance of diffuse A plaques, but also displayed increased levels of microhemorrhages. It was also shown th at CAA levels were not affected, only hemorrhages associated with this pathology were shown to increase in occurrence. CAA development appeared to be a requisite for this pathology to occur, as 6 monthold APP23 mice lacking CAA did not deve lop these hemorrhages. One suggested mechanism for triggering of th ese hemorrhages was that anti-A antibodies are capable of binding to amyloid deposits in bl ood vessel walls, potentially causing an inflammatory event that served to weaken such amyloid-bearing blood vessel walls. It was suggested that because approxim ately 80% of AD patients over 65 develop CAA, and because only APP23 mice develop CAA at levels comparable to those found in AD, that results from other passive anti-A studies may not be a perfect indicator of the respon se such treatment would elicit in humans (Pfeifer et al. 2002).
86 Behavioral Effects of Passive Anti-A Immunization Along with those studies focused sole ly on pathological impacts of passive anti-A immunotherapy, a few studies also inves tigated behavioral impacts. Dodart et al. (2002) reported that 24 month-old PDAPP mice tr eated i.p. with m266 anti-A antibodies for six weeks showed performance in object recognition tasks similar to 8 month old wild-type mice and significantly better than non-treated PDAPP mice or IgG1 isotype treated controls. No significant reduction in A pathology was noted in these mice, in contrast with previous findings from this same group in younger PDAPP mice, and task improvement was ther efore attributed to reduced levels of soluble A in the CNS. This was supported by correlation analysis showing no correlation between A burdens and behavior in m 266 treated mice and by findings that 8 month old non-treated PDAPP mice were significantly worse than aged, treated PDAPP mice at recognizing novel objects despite having lower levels of A deposition. Further work by Dodart et al. (2002) studied the impact of acute, single administrations of m266 on 11 month old PDAPP mice. Again it was shown that these mice were better at recognizing nove l objects than age-matched transgenic controls after receiving only one m266 administrations. To further explore the effect of acute, single-dose m266 therapy on beha vior in 11 month old PDAPP mice, 4 days of the holeboard task were administered. It was reported that treated mice made significantly fewer errors in this task duri ng the final two days of testing (data not reported for initial two days of testing), again emphasizi ng the potential for learning and memory benefits based on m266 therapy. Finally, statistical analysis showed a
87 dose dependent correlation between obj ect recognition performance and m266 therapy dosage, and only mice r eceiving maximal dosage (250 g) showed performance above chance levels in this task. The antibody used in this study was known to not decorate existing A plaques and was thought to not cross the BBB, however a dose-dependent correlation was made between plasma concentrations of A isoforms and m266 therapy. Based on th ese findings, it was suggested that the major mechanism of action for m266 thera py was for these antibodies to act as a peripheral A sink, binding and sequestering peripheral amyloid and causing an equilibrium shift of CNS to peripheral A Thus, according to this theory, cognitive protection can be gained via modulation of CNS soluble A levels without reductions in A plaque burdens. (Dodart et al. 2002). Kotilinek et al. (2002) reported similar re sults to those seen by Dodart et al. (2002). It was shown that admini stration of the monoclonal anti-A antibody BAM10 (which recognizes the N terminus of A ) did not lead to plaque clearance in the CNS of 9-11 month old Tg2576 mice, but that reversal of memory loss did correlate to such administration. Baseline perfo rmance was measured just prior to immunization and consisted of administration of the visible platform task followed by hidden platform coupled with three intermitte nt probe trials (a variation of Morris water maze application). Mice were divi ded into treatment groups based on mean probe trial performances. 4-5 days after baseline evalua tion, treatment mice underwent 3 separate injections of BAM10 over a period of one week. 11-12 days following baseline testing, hidde n platform testing was administered using the same
88 protocol; however a different set of visual cues were arranged around the testing pool and a unique platform location was used. Major differences among groups based on treatment revolved around probe trial results. It was shown that percent time spent in the target quadrant for this task was si gnificantly increased in BAM-10 treated mice; in specific probe trials, however this trend was not noted in overall percent time measures (averaged from all probe trials) and no significant difference was measured between BAM-10 treated mice and IgG treated control mice overall. No improvements were noted for hidden platform performances. The interpretation of this data by the authors was that BAM-10 treated mice showed a reverse of memory loss following treatment. As there were not significant alterations in A levels in these mice, the authors contributed this reve rsal of impairment to neutralization of A assemblies in the CNS resulting in decreased neurotoxicity of these constructs. It was shown that BAM-10 treatment eliminated the negative correla tion between soluble brain A and probe trial scores, i ndicating that this treatment did in fact reduce the impact of existing A in the brains of Tg2576 mice (Kotilinek et al. 2002). Wilcock et al. (2004) administered weekly doses of anti-A antibody to Tg2576 mice for either 1, 2, or 3 months, with treatment beginning so that all mice finished the treatment regimen at 22 months of age. Behavioral assessment of these mice consisted of a single, 8 minute administration of the Y-maze task 1 day before sacrifice. Mice were treat ed with IgG antibody to A 28-40 with controls receiving an antiDrosophilia amnesiac protein IgG1 antibody. Only mice treated with Anti-A for 3 months were shown to have signifi cantly higher spontaneous alternation in Ymaze performance compared to transgenic control groups. Performance for mice
89 treated for 3 months was at the same level as non-transgenics, with reduced performances following lesser treatment periods. However, after 1 or 2 months of treatment, immunized mice were not signi ficantly different in performance from either non-transgenics or controls transgen ics. Hyperactivity based on Y-maze arm entries was noted in non-treated APP cont rols but was not noted in 2 or 3 month treated animals. Antibodies administered fo r all treated mice were shown to decorate plaques within the CNS, with direct proportionality between immunostaining and plaque density. Following 1 month of anti-A treatment, transgenic animals were noted to have slight but not significant reductions in A deposition. Following both 2 and 3 months of treatment, however, significant decreases in A immunostaining were found in hippocampal and frontal cortex regions, with both compact and diffuse plaque burdens reduced. It was further shown that microglia Fc receptor expression was upregulated afte r 1 month of anti-A antibody treatment, CD 45 molecules after 2 months, and that expression levels of both marks then returned to normal after three months, indicating temporal ac tivation of these cells. Based on these findings, it was suggested that A clearance was thus mediated via a number of mechanisms, including solublizat ion of deposited A by antibodies, sequestering of A in the plasma, as well as by microglia ac tivation and phagocytosis of A Furthermore, it was said to be this reduction in A burden in the CNS whic h resulted in improved Ymaze performance for treated Tg2576 mice (Wilcock et al. 2004), although no
90 correlation between A deposition and Y-maze alternation was done for A antibody treated mice. The speed with which acute administration of anti-A antibodies result in cognitive benefit in AD transgenic mice (b ased on findings by both Kotilinek et al. 2002 and Dodart et al. 2002) and the lack of A pathology decreases in these animals would seem to support the theo ry that these antibodies ar e acting more through either alterations in CNS:sera A levels or neutra lization of toxic A constructs. The rapid behavioral benefits reported by Dodart et al (2002) and Kotilinek et al. (2002) could, however, result from alleviation of secondary effects of A (a known vasoconstrictor) in the blood resulting in in creased cerebral blood flow. In any event, findings by Wilcock et al. (2004) still support the c oncept of cognitive protection following reductions in A pathology due to passive A immunization. Passive Anti-A Immunization in Humans While no data has been reported on passive A immunization in humans, the potential exists for such studies to occur based on results noted above in assessment of such immunization in tran sgenic models of AD. In the wake of findings from active immunization trials in humans, however, the climate for research along immunotherapeutic lines in humans with AD ha s changed dramatically in the last few years. Findings such as those noted by Pf eifer et al (2002) sugge st that inflammation following anti-A administration can occur in tr ansgenic animals which develop CAA, and it is not unreasonable to assume that similar side-effects would occur in humans following such treatment. As Anti-A antibodies have been shown to
91 sequester amyloid in the blood, it is also possible that increased development of CAA or other vasculature dysfunctions may occur. Other side-effects from continual/repeated administration of antibod ies in humans may result from eventual immune response to injected antibodies, resulting in accumulation of immune complexes and potentially further disruption of the vasculature (serum sickness) (Sigurdsson et al. 2002). Su ch activation of the periphera l immune system may also lead to development of inflammation with in the circulatory system. Although many results from passive A immunization in animal models of AD have been positive, it is clear that the immune envi ronment of transgenic models is extremely variable even based solely on background differences among mice, and the major lesson to be taken from the human vaccination trials is the potential for differences between immune responses in these mice and in humans to re sult in devastating side-effects. Until a valid and long-lasting therapeutic for Alzh eimers Disease is discovered, however, it is likely that the need for such treatments will eventually outweigh other concerns, and immunotherapeutic approaches such as passive A immunization will undoubtedly be assessed in hum an subjects at some point.
92 Other Immunotherapeutic Strategies for Alzheimers Disease Based on our current understanding of immunotherapeutic approaches to Alzheimers Disease, it is clear that the most probable means of reducing A pathology in the CNS is limited to two mech anisms. The first is humoral based and involves clearance of A from the CNS via two separate routes. Solublization of A in the CNS by direct actions of anti-A antibodies can result in clearance of both soluble and insoluble forms of the pe ptide, while sequestration of the A by anti-A antibodies in circulation can lead to an equilibration shift of A into the serum and out of the CNS. The second is cellular ba sed and relies on acti vation of key immune system cells within the CNS to clear A by phagocytosis and internal degradation of the molecule or by production of extracellu lar peptidases capab le of breaking down aggregated A A immunization, both active and passive, is known to initiate reduction of A levels within the CNS, both solubl e and insoluble, via both pathways as described above. Alternative immunotherapies attempt to achieve similar results without instigating inflammatory damages associated with A vaccination. Currently, two potential immunotherapies in particular show promise. The first involves development of non-antibody A binding agents which can be administered
93 peripherally and serve to sequester Ain a similar fashion to anti-A antibodies, again resulting in increased traffic of A out of the CNS. As discussed previously, these agents thus serve as peripheral A sinks by increasing serum levels of A via this mechanism. The second involves either stimulation of an appropriate T cell response via various mechanisms or direct infusion of specifically activated T cells to elicit a desired immune response. Alternative Peripheral A Sinks In 2002, Matsuoka et al. reported that peripheral administration of two separate A binding molecules, gelsolin and GM1, in young APPsw+PS1 mice resulted in significant A clearance from the brains of these mice. These bloodlimited molecules were thought to sequester A in the periphery, causing a shift in A equilibrium away from the CNS, as described for anti-A treatments. Gelsolin in particular is unable to cross the BBB, and so its affects are attributed solely to peripheral binding of A Importantly, it was noted after treatment that plasma levels of A did not increase at anywhere near the levels seen after passive A immunization, making the exact mechanism of action for A sequestration in the periphery difficult to interpret. Admini stration of these compounds in aged PDAPP mice did not result in significant reduction of amyloid plaques, apparently indicating once again that amyloid plaque burdens are more difficult to reduce in older animals (Matsuoka et al. 2002). This resiliency of established plaques in aged mice against reduction following immunotherapy has been established in numerous studies (Morgan et al. 2000, Arendash et al. 2001, Dodart et al. 2002, Kotilinek et al. 2002).
94 In 2003, Deane et al. showed that sRAGE, derived from the RAGE A binding protein, can also act to sequester A in the periphery, providi ng another potential for A -binding protein therapy. The exact mechanism by which A is transported across the BBB to the periphery is thought to be mediated by the LRP clearance receptor for A while reentry of the peptide into the CNS appears to be mediated by RAGE receptors. Other carrier proteins such as album in and transhyretin are also thought to be involved in this process. Modification or inhibition of certain aspects of these pathways or treatments may provide a mechanism for altering sera:CNS A ratios, providing another mode of A clearance from the CN S (Zlokovic et al. 2004). T-cell Based Therapy As indicated above, immunotherapeutic treatments for neurodegenerative diseases such as AD result in varied levels of activation among separate T helper cell subtypes. The actions of these reactive T cel ls then bias the over all immune response. In the most general terms, Th1 T cells cau se a cellular based response and Th2 cells a humoral response (Vallejo et al. 2004). Eith er response, if elicited in an appropriate fashion, may have a beneficial effect in the AD brain; the former via activation of phagocytes, the latter via activation of antibody-pr oducing B cells and/or downregulation of neuroinflammation. The potential for such bias to result in negative consequences has also been we ll established, for in stance it has been indicated in vaccination studies that a Th1 biased immune response results in increased inflammation, potentially leadi ng to an aberrant auto-immunee reaction (Furlan et al. 2003). These responses are the most general aspects of T cell activities,
95 however, and, especially with regards to active immunization, it is unclear to what extent undesirable T cell activation and th e neuroinflammatory events which result are due simply to the unique reactions elicit ed by the antigen and/or adjuvant used. Modulation of the T cell response via a lterations in antigen/adjuvant vaccine components, as well as delivery routes/vehic les, could potentially result in benefit without neuroinflammation (Cribbs et al. 2003), and certain immunological treatments have been show n to elicit such benefit without any T cell activation (Schenk et al. 2004). However, it is important to consider that T cell subsets as a whole are not anathema. The potential activitie s of these cells in response to antigen based activation are wide ranging, and it rema ins to be seen whether such activation in the context of a non-immunized system is beneficial or detrimental. Infusion of A reactive T-cells has never been attempte d in any animal model of AD, however the concept of such adoptive immunotherapy has been established for some time in other fields of study. In i ndividuals with lowered immune system efficacies or hyporesponsiveness to specific antigens, T cell infusions act to prime these systems to counterbalance such deficits. Donor Lymphocyte Infusion For Non-CNS Diseases Infusion of activated T cells has been used in the treatment of a number of disorders and diseases, with benefit gained due to actions of both CD4+ and CD8+ T cells individually and in concert. There is great potential for such adoptive transfers to serve as anti-tumor treatments due to the activity of both cytotoxic cells recognizing tumor specific antigens and memo ry T helper cells capable of sustaining an immune response to malignant cells (Vonde rheide and June et al. 2003). Findings
96 that graph-versus-leukemia anti-leukemia affects were more likely to occur following transplant of non-T-cell-depl eted bone marrow transplant s resulted in a number of studies being conducted to test the effi cacy of T cell Donor Lymphocyte Infusions (DLI), or selective infusion of non-patient lymphocytes, in treating various leukemias (Kolb et al. 1995, Collins et al. 1997, Slavin et al. 1996) as well as myelodysplastic syndrome (Bressoud et al. 1996), multiple myeloma (Lokhorst et al 1997), and nonHodgkins lymphoma (Bernard et al. 1999). Complications were most prevalent following bone marrow transplant-based treatments and c onsisted mainly of marrow aplasia, or failure of the bone marrow to thrive, though this was resolved in as many as 80% of patients (Collins et al. 1997). The potential for T cell infusions to serve as a treatment for numerous cancers, especially for recurrent hematalogic cancers postbone marrow transplant, has thus been established (Szer et al. 1993). Continuing research in immune system diseases have shown that T cell infusions serve to re-establish immune functions in patients with Epstein Barr Virus (EBV), cytomegalovirus, and Human Immunodeficiency Virus (HIV) infections, though not always without some adverse affects (Heslop and Rooney 1997, Walter et al. 1995, Levine et al. 2002). This research elicits the potential for recharging of the immune system following DLI in imm une compromised individuals or in counteracting hypo-immuneresponsiveness to a toxic self peptide such as has been implicated in AD. Priming of the CNS Immune System With T Lymphocytes APP transgenic mice display a tolerance-based effect after immunization with A which results in a decreased immune system response (Monsonego et al. 2001).
97 To elicit an immune response to A in humans, it was deemed necessary to administer an adjuvant in concert with th e peptide (Hock et al 2002). Otherwise the A peptide, which is formed via prot eolytic processing of the endogenously expressed APP molecule found in all humans would not likely elicit a significant immune response until late in AD development when it is associated with neuroinflammation. This self tolera nce may affect th e ability of A vaccination to be effective as a treatment, especially as immune system activation in AD is largely restricted to local, glial based, non-systemic types (Streit 2004, DAndrea 2004). An exploration of the ab ility of T cells to prime the immune systems of AD models is thus warranted to determine the potential fo r such cells to overcome or reverse self tolerance to A inducing either microglial phagocytosis of A or B cell anti-A antibody production. As described above, APP mice display a lessened immune response following A immunization compared to wild-type contro ls, most likely due to self tolerance. As both APP and A are present under normal c onditions in humans throughout aging, it is likely that a certain level of tolerance is present in humans as well following similar vaccination. DLIs have been used to combat deficient immune systems in murine models as well as humans, whether these deficiencies be congenital or due to an infectious dis ease of the immune system. It has been suggested a deficiency in the immune syst em may be present in AD which prevents the immune cells of the CNS from taki ng appropriate action against the insults inflicted therein during the course of diseas e. Both microglia and T lymphocytes are likely to be drawn to areas of inflammation associated with A pathology in AD
98 brains (DAndrea 2004, Farkas et al. 2003). However, these cells may not be appropriately activated early enough along the course of the di sease to be affective, or may be deficient in their activity (Str eit 2004, Monsonego et al. 2001). Prior to compact neuritic plaque formation, microglia could potentially be stimulated to clear deposited A At the same time, activated T cells can both regulate the immune response to A associated damage and, as will be elicited later, potentially play a neuroprotective role against further dama ge. If activated too late, however, frustration and continual stimulation of mi croglia and T cells (via co-stimulation) may occur, resulting in exacerbation of the inflammatory cascade rather than protection. A -sensitive T cells are unlikely to be found in large numbers in APP transgenic mice or in humans due to selec tion against thymocytes reactive to self proteins which occurs during T cell de velopment. As AD progresses and A becomes more prevalent and is associated with inflammation and cellular damage, however, is likely that A -senstive T cells would be activated and proliferate. This is supported by findings by Monsonego et al. (2003) that A -sensitive T cells are common in both elderly, hea lthy individuals and AD patient s, with AD patients more likely to possess these cells. What is not clea r, however, is if this immune response to an endogenous peptide is beneficial but is simply deficient in AD patients, or is a detrimental autoimmune dysfunction. If T cells do possess the potential to combat the damage associated with amyloid pathology, it is possible that A -sensitive T cells themselves could serve as a potential therapeutic in AD. These cells may then play a role in activating other
99 immune system cells as described above, cau sing an upregulation in immune system activity, or a charging of the AD immune system. The potential drawback of such activation, of course, is that an incorrectly balanced immune reaction can lead to a strong bias towards a cellular or hum oral based response, and under certain conditions this can lead to increased neur oinflammation, as seen in human clinical trials for A vaccination (Nicoll et al. 2003). Careful study of the effects of A sensitive T cell infusions in AD models will be necessary to determine the true potential of such cells. Neuroprotective Actions of T Lymphocytes It is thought that in AD, i mmune hyporesponsiveness to A prevents effectiveness of some A immunotherapies, as would be true for any endogenously expressed peptide due to self-tolerance. To overcome this deficit, adjuvant coupled administration of A peptide was used in human clin ical trials, and it was found that this led to potentially lethal inflammation in the CNS of immunized individuals. This may have occurred in part due to activation of cells sensitive to a self peptide. Autoimmunity of this type is typically a ssociated with a negative connotation due to such devastating effects, which are not un-similar to those noted in autoimmune disorders. However, new research in this field has begun to reve al that autoreactive cells, specifically T cells, are actually capable of providing neuroprotective benefits in areas of CNS inflammation and neurona l damage. Understanding and harnessing of this potential may lead to potentia l new therapeutics fo r neurodegenerative diseases such as AD, in which neuroinflammation plays a key role.
100 Extensive research in neural injury repa ir has shown that T cells can play an essential role in modulating repair in neuroinflammatory-b ased damage. Specifically, myelin-specific Th1 cells (the same proi nflammatory, encephalitogenic cells capable of instigating onset of au toimmune diseases) seem to be capable of providing protection following neuronal injury. It appears that these autoreactive T cells, capable of causing neuronal damage and inflammation when inappropriately activated, are also capable of being drawn to an area of injury and in some way acting as neuroprotectants under the correct neurophysiol ogical paradigm (Kipnis et al 2002, Schwartz et al. 2003). One pathway of this protection appears to be regulation of microglia activity. Microglia, known to both be activated by and activators of T cells as discussed above, can contribute to continued ne uronal damage following initia l insult if left unchecked. T cells do possess the ability to prevent frustrated microglia from continuing to cause neuronal damage (Avidan et al. 2004). A s econd route of action for T cell mediation of CNS damage may be producti on of neurotrophins. Infused autoimmune T cells are capable of producing such factors after reac tivation due to antigen presentation. Both Th1 and Th2 subsets of autoimmune cells have been shown to produce these factors following injury to the CNS, resulting in ne uroprotective effects. These effects can be blocked via inhibition of NT-receptor ac tivity, indicating that it is the direct interactions of these fact ors with neurons which is providing protection, and these receptors are known to be upregulated on injured neurons (Moalem et al. 2000). Finally, the work of Michal Schwartz and associates in the past two years has displayed the ability of T cells to protect against glutamate neurotoxicity. Excessive
101 activation of neurons and/or glutamate rel ease by activated cells such as microglia can result in such toxi city, and the potential of T cells to lessen glutamate toxicity based damage is another potential mechan ism by which these cells are capable of providing protection following CNS damage The mechanism of action for this protection appears to be two fold. In re sponse to neuronal damage associated with glutamate toxicity, self-specific T cells dir ectly decrease damage following glutamate exposure. This is attributed to the neuropr otective capabilities of autoreactive T cells as discussed above. Secondly, regulatory T cells re duce cytotoxic-type microglial activation, reducing neuroinflammatory dama ge associated with such activation (Schwartz et al. 2003). These findings seem to imply an impo rtant, direct route by which T cells typically associated with inflammation in the CNS are actually important in combating damage under appropriate condi tions. The damage and associated neuroinflammation present in AD may thus al so be regulated and counterbalanced by the same T cell subsets which might otherwise be involved in propagating neuroinflammatory events. As has been illu strated earlier, microg lia associated with neuritic plaques in the AD brain are known to instigate neuronal damage if left unchecked due to phagocytic frustration. However, microglia have been shown to be capable of acting as A APCs for T cells, indicating a mechanism by which A sensitive T cells may be reactivated local to these plaques (M onsonego et al. 2003). In 2003, Farkas et al. reported that A mediated break-down of the BBB followed by infusion of activated T cells resulted in migration of these cells into the brains of male Wistar rats. Based on these results, it is reasonable to assume that in murine AD
102 models, peripherally infused A -sensitive T cells would migrate across the BBB into the brain, allowing them to associate with mi croglia in amyloid containing regions of the CNS. Microglia express MHC-II in order to in teract with T cells, allowing microglia to participate in T cell activ ation, and thus indirectly in neuroprotection. Suppresed expression of MHC-II is thought to be present in AD and may prevent early actions of microglia in clearing amyloid plaques, a nd it has been suggested that localized TH1 cells may be able to stimulate such expressi on at an earlier time poi nt (Schwartz et al. 2003). In the presence of inflammatory f actors released by damaged neurons, and following co-stimulation with microglia, thes e T cells may then be stimulated by the same mechanisms outlined above to both re duce local gliosis and counteract neuronal damage associated with A deposition. As a caveat, it s hould again be noted that TH1 subset cells have been implicated in play ing a major role in neuroinflammation in AD and other CNS disorders (Furlan et al. 2003). Interestingly, howev er, it appears that microglia and T cell co-activ ation in the CNS results in upregulated production of anti-inflammatory cytokines such as IL-10 by T cells, resulting in decreased production of neuroinflammation-inducing cyto kines and other inflammatory factors by microglia and TH1 cell lines (Chabot et al. 1999). Mallat et al (2003) produced results emphasizing the potential benefit to be gained via this mechanism. It was shown that adoptive transfer of ovalbumin-specific Tr1 subtype T cells (which are capable of downregulating Th1 cells) reduc ed development of atherosclerosis in Apolipoprotein E knockout mice. Injection of these cells along with ovalbumin in complete Freunds adjuvant resulted in activation of infused lymphocytes and a
103 significantly reduced overall Th1 response (based on cyt okine profiles). It was suggested that this bias against a cellular based immune response resulted in decreased inflammation, which is thought to play a major role in development of atherosclerosis in these A poE KO mice (Mallat et al. 2003 ). Importantly though, this study focused on overall Th1 subset activ ities, not on the potential for specific autoreactive Th1 subsets on the CNS side of the BBB. Th2 subsets may also be capable of play ing a neuroprotective role. Benner et al. (2004) demonstrated that Cop-1 (a T h2 phenotype specific stimulator) activated immune cells adoptively transferred in to MPTP-intoxicated mice (a model for Parkinsons disease) are capable of migr ating into the inflamed nigrostriatal dopaminergic regions of these mice. Once lo calized to the site of inflammation, these cells were shown to be capable of suppre ssing microglia actions and of stimulating production of neuroprotective products such as glial cell line-de rived neurotrophic factor (GDNF) by astrocytes to result in significant neur onal protection (Benner et al. 2004). Thus, without serving to reduce A pathology, infused A autoreactive T cells, including both Th1 and Th2 subsets, may be capable of protecting against A associated damage if they too are able to produce neuroprotective agents and to combat neuroinflammation. Perhaps the most persuasive study indicat ing the potential for T cell therapy in AD was conducted in 2004 by Avidan et al. This study reported th at T cells specific to neuronal self antigens are capable of reducing damage associated with neurotoxic properties of A Intraocular A 1-40 injections in C57BL/6J mice resulted in significant death of retina l ganglion cells. In mice immunized prior to A injection
104 with interphotoreceptor reti noid-binding protein (antigen specific cells in the eye), loss of neurons was significantly less. To verify that this reduction was T cell induced, interphotoreceptor retinoid-binding protein and S-antigen specific T cells were passively transferred into C57BL/6J mice immediately following A intraocular injections, resulting once agai n in retinal ganglion cell prot ection. Further findings by Avidan et al. (2004) would seem to dow nplay the potential for neuroprotection following infusion of A -sensitive T cells in APP transgenic mice. It was noted that T cells specific to non-aggregated A1-40 are capable of inducing only slight reductions in A -induced retinal ganglion cell loss. Th is was attributed to a lack of microglial capacity for expressing MH C-II in response to aggregated A 1-40 exposure based on unpublished results by Butovsky et al. However, the authors do suggest that memory T cells specific to self antigens should be capable of activation and downregulation of cytotoxic microglia actions. Th is is more in agreement with findings by Monsonego et al. (2003) which indi cate that presentation of A and direct activation of A specific T cell subsets by activated Microglia can occur. Furthermore, additional findings published by Monsenego et al. demonstrate the presence of A sensitive T cells in AD patients, and it is suggested that processing of A in the CNS by APC cells followed by migration of AP Cs to lymph nodes leads to eventual peripheral activation of A reactive T cells (Monsonego et al. 2003-2). The potential for A -sensitive T cells to be activated via MH C-II presentation is thus established in both murine systems as well as in humans.
105 There appears to be potential for ad optive T cell transfer to serve as a therapeutic for neurodegenerative diseases. The pathways by which such protection may occur in AD, as discussed above, are outlined in Figure 2. The first pathway is based on the ability of certain T cells to stimulate the immune system early in A pathology development via either a cellular (stimulation of microglial phagocytosis of A ) or humoral (stimulation of anti-A antibody production) based response. In addition, there is the potential for T cells to provide protection against neuronal damage resulting from development of A pathology later along the course of the disease via both production of neurotrophi ns and down-regulation of glial-based neuroinflammation. This form of protecti on may also extend towards prevention of glutamate based neurotoxicity. No studies to date have established the behavioral effects of T cell infusions into a neurode generative disease model, nor have the pathological effects of such infusions b een explored in a model of Alzheimers Disease. Elucidation of the actions of va rious T cell subsets once within the CNS of AD models may provide insight into the potential for T cell therapy in AD.
1)Co-Activation:MicrogliaTH1 T CellTH2 T Cell4)Decreased Neuroinflammation:a. Decreased secretion of inflammatory proteins/cytokines(also reduces glutamate toxicity)b. Decreased Adepostion(due to decreases in ApoE/ACT etc.)1)Co-Activation: 2)AReactive T Cell Re-Activation2)AReactive T Cell Re-Activation Reduction of MicrogliaInduced NeuroinflammationMicrogliaDe-ActivationMicrogliaMHCII Expression, APresentation to T Cell 5)NeurotrophinProduction a. Increased neuronal health/growthb. Decreased neuronal lossc. Decreased glutamate toxicity Activation of Cellular Immunity (Microglia, Neutrophils, CytotoxicT Cells, etc.) Activation of HumoralImmunity (B Cells, HumoralBased Complement System, etc.)3)Recharging of the CNS Immune System, Induction of Anti-AActivity Figure 2. Mechanisms of T cell based neuroprotection. Following re-activation in the CNS via interactions with APCs such as microglia, infused auto-reactive T cells sensitive to A may act therapeutically in the AD brain via three mechanisms: 1) Early activation of the immune system to promote both humoral and cellular based immune responses to soluble and deposited A(red). 2) Suppression of frustrated cell (microglia, astrocytes, etc.) actions in the region of A mediated neuronal damage to prevent excessive neuroinflammation and associated glutamate toxicity (blue). 3) Production of neurotrophic factors which promote neuronal health and help protect against neuronal loss/dysfunction while also protecting against glutamate neurotoxicty (green). 106
107 Specific Aims The specific aims of this thesis are: 1) To isolate, culture, and prime with human A 1-42 peptide, spleenocytes isolated from young, A 1-42 immunized wild-type mice, and to inject these activated, multi-subset, T cell enriched spleenocytes into AD transgenic mice. 2) To determine the effects of A -sensitive T cell enriched spleenocyte infusion on cognitive performance in Radial Arm Water Maze, Platform Recognition, and Y-maze tasks compared to placebo-infused transgenic and wild-type mice. 3) To determine post-mortem the pathologic effects of A -sensitive T cell enriched spleenocytes within the CNS of AD transgenic mice through analysis of A burdens and distribution/activa tion levels of microglia and T cells therein.
108 4) To determine effects of A -sensitive T cell enriched spleenocyte infusions on blood cytokines levels, specifically w ith regards to levels of pro and anti-inflammatory cytokines. 5) To perform correlation analysis to determine if behavioral task performance measures correlate to any specific pathologic or biochemical measures.
109 Materials and Methods Animals To obtain the APPsw+PS1 mice used in this investigation, male heterozygous APPK670N,M671L transgenic (APPsw) mice were cr ossed with female, homozygous PS1 mice of the 6.2 transgenic line. Female he terozygous PS1 transgenic offspring were then crossed with APPsw males of th e parental generation to obtain the nontransgenic and heterozygous APPsw+PS1 mice used for th is investigation. Mouse backgrounds for this generation consist of a 56.25% C57, 12.5% B6, 18.75% SJL and 12.5% Swiss Webster heterogeneous b ackground. Mice were genotyped after weaning, and then singly housed throughout the course of the study in standard 12hour light-dark cycle conditions, with free access to water and rodent chow. Behavioral assessment and immune cell infu sions occurred during the light phase of the circadian cycle. General Protocol As indicated in the Timeline in Fig. 3, baseline cognitive performance for 10 non-transgenic and 15 APPsw+PS1 mice was established via a 12 day administration of the RAWM task for working memory at eight months of age. Transgenic mice were then divided into two groups balan ced in RAWM performance based on overall latency and error averages so that prevs. postinfusion comparson could be made
110 between groups. Seven transgenic mice were designated as recipients for A sensitive immune cells, and 8 transgenics (as well as all 10 non-transgenics) were designated as controls, to be given PBS. To generate A -sensitive immune cells, 3 additional congenic non-transgenics we re repeatedly immunized with A 1-42 at 3-4 months of age. Spleenocytes and lymphocytes isolated from these animals were then cultured for 4 days prior to infusion into immune cell recipient mice. Infusion of immune cells into these mice along with in jection of PBS into both control groups occurred two days following completion of RAWM testing (Fig. 3). After a one month delay, mice underwent additional be havioral assessment consisting of RAWM, Platform Recognition, and Y-maze task administrations (9, 4, and 1 days respectively). A neurological assessment battery was also used to test for motor and/or neurological deficits pr ior to sacrifice. Plasma sa mples and brains were taken at sacrifice. Plasma cytokine levels we re evaluated via immunoassay, and half brains (bisected mid-sagitally) were coronally sectioned and used for quantification of brain A deposition levels via direct Thiofl avin S staining, as well as by 6E10 immunostaining. Finally, statistical an alysis was conducted to determine group differences in cognitive task performa nces, plasma cytokine levels, and A load determinations. Correlation analyses, fact or analysis, and discriminant function analysis were also be used to evalua te overall group differences across these measures.
Neurological Assessment 8RAWMPlatform RecognitionRAWM Injection of ASensitive Immune CellsY-Maze Euthanasia 91011 MonthsStudy Timeline Neurological Assessment 8RAWMPlatform RecognitionRAWM Injection of ASensitive Immune CellsY-Maze Euthanasia 91011 MonthsStudy Timeline Figure 3. General protocol time line for the adoptive transfer study. 111
112 Immune Cell Infusion Protocol Three non-transgenic mice from the same generation as the study cohort were vaccinated with human A 1-42 to elicit an immune response. One day prior to injection, the site of injection on the backs of these mice was shaved. The following day, 3 separate endodermal injections of human A 1-42 peptide (Bachem) in complete Freunds adjucant (CFA) were administered to each mouse following isofluorane anesthesia. Each mouse received a total of 250 g of non-fibrillar, soluble A in 100 l of adjuvant divided equally ove r the three injections to induce proliferation of A sensitive immune cells. Each mouse then received additional injections of 200 ng pertussis toxin 1 a nd 3 days post-immuni zation. Mice were monitored for signs of distress and paralysis fo r these 3 days and for 7 days thereafter. Following this 10 day period, these mice were s acrificed for harvest of spleen tissues and lymph node fluids. Tissues from all 3 mice were pooled and homogenized with a loose fitting 15 mL dounce, followed by filtration through a 70 m sieve filter, centrifugal pelleting, and resuspension in re d blood cell lysis buffer. After five minutes, cells were again gently pelleted, then resuspended in full media and counted. Cells were then cultured in full media containing 8-10 g/mL A 1-42 and for 4 days. After culturing, cells were once again pe lleted, resuspended in phosphate buffered saline (PBS), and viable cells were count ed by trypan blue exclusion. Immune cell recipient APPsw+PS1 mice were then administered approximately 2x107 viable immune cells in 0.5 mL PBS by tail vein injection. Control transgenic and nontransgenic mice received tail vein injecti ons of 0.5 mL PBS. All animals were then monitored daily for signs of discomfort or paralysis.
113 Behavioral Assessment Radial Arm Water Maze The RAWM task for working (short-term) memory was administered for all mice at two time points as indicated on the study timeline. Testing occurred in a 100cm diameter pool of water divided by an aluminum insert which establishes 6 evenly spaced radial arms surrounding a centr al open area. Each arm was 30.5 cm in length and 19 cm wide, while the central ci rcular swim area was 40 cm in diameter. A transparent 9 cm diameter platform at a 1.5 cm depth was used as an escape platform for mice placed in the maze, and was placed approximately 15 cm away from the end of the randomly assigned goal arm of the maze each day of the task. Numerous visual cues su rrounded the pool to provide spatial references for navigation of the maze. For each trial of the task, mice were placed in the water at the entrance of a novel start arm of the m aze for that day, facing the central swim area. The start arm was never the arm cont aining the submerged escape platform, and both start arm sequences and goal arm lo cation were semi-randomly changed each day. The mouse was allowed to navigate the maze freely until fully entering an arm of the maze. If the arm selected by the anim al did not contain the escape platform, an error was recorded and the animal was pulled gently back to the start location for that trial. This procedure, including counting of an error, was also used if no arm selection was made within 20 seconds. If after 60 seconds, an animal did not find the escape platform, it was guided gently to it. Latency to find the platform up to 60 seconds and number of errors were reco rded for each trial. The RAWM task
114 consisted of five daily trials the first four of which were administered with only a 30 second delay between them during which the mouse was left on the escape platform. The fifth trial was administered 30 minutes la ter, and the animal was returned to its cage during this delay interval. For animal s which did not make at least 3 choices during a trial and did not find the escape pl atform, a penalty error was recorded for that trial. The penalty error was calculated by averaging trial one errors for the first three days of testing for animals which did not find the escape platform but did make at least three arm choices during these trials. Platform Recognition The platform recognition task was ad ministered following completion of the RAWM task to determine th e ability of the mice to recognize a visible platform placed at various locations in an open, 100cm diameter pool. This required a strategy switch, as the RAWM task required that the animals search for a submerged platform hidden in a divided pool based on orientati on of visual cues around the pool, whereas the platform recognition task only require s identification of the visible escape platform. Thus, this task also tested the ability of these animals to adapt to a new escape strategy. For this task, a 9 cm circ ular platform with a conspicuous 10 x 40 cm black and white conical visual cue att ached was used, with the surface of the platform raised .8 cm above the waters surface. Mice were placed in the same location in the pool for each of 4 daily trials For each trial of a given day, the visible platform was moved to a central part of one of four established quadrants in the pool. The mice were allowed 60 seconds to find th e platform before being guided to it
115 during each trial, and a 30 second stay on the platform was allowed between each trail. Latency up to 60 seconds for each tr ial was recorded for 4 days of testing. Y-maze Mice were given 5 minutes to explore a black, three-armed Y-maze in order to test both general activity (based on total number of arm entries) and basic mnemonic processing (based on spontaneous percent alte rnation). Mice were placed facing the center of the maze at the opening of the maze arm designated arm two. During the five minute test period, the number and se quence of are choices was recorded. Percent alternation was calculated by subt racting two from the total number of arm entries and dividing this valu e into the total number of s pontaneous alternation events (consecutive choice of each of the 3 Y-maze ar ms, without re-entry into a previously chosen arm) for each animal. Neurological Battery 19 sensorimotor measures were taken from a neurological battery derived from the Irwin Test (Irwin 1968). Qualitit ative assessment ratings of the following were obtained for each mouse: transfer arousal after placement in a new environment, ataxia, hypotenic gait, pelvic elevation (while walking), tail elevation (while walking), palpebral closure, appro ach to Q-tip (novel object), withdrawl from invasive Q-tip, response to touch, response to audible startle stimulus, corneal reflex, righting reflex, visual placing (reaching for paw hold while hanging), grip strength (while hanging), recognition of visual cliff, response to toe pinch, and response to tail pinch. Quantitative measures were also obtai ned for the following: latency to move
116 to the edge of an elevated platform and number of head pokes over the edge of an elevated platform in one minute. Brain Collection and Sectioning Two days following behavioral testing all study mice were anesthetized with Nembutal (1 mg/10 gm body weight). Bl ood (.2 ml) was collected from the heart, mixed with .5M EDTA, and centrifuged. Pl asma and red blood cells were then separated, flash frozen, and stored at 80C. Following blood collection, all mice were pericardially perfused with .9% saline. Brains we re excised and halved midsagitally. The left half was fixed in 4% paraformaldehyde for 24 hours at 4C, followed by graded sucrose solutions ( 10, 20, and 30% (w/v) sucrose in 0.1 Sorenson's phosphate buffer) prior to sectioning. Frozen 25 m coronal sections were then collected on a sliding microtome and stored in PB S. These sections were immunostained with 6E10 antibody (for diffuse A ) or stained with Thioflavin S (for compact A deposition). Plasma Cytokine Level Measurements Relative cytokine level determinations were obtained using a custom RayBio Mouse Cytokine Antibody Array. Briefly, pr ovided membranes bearing ten separate anti-mouse-cytokine antibodies were treated with blocking buffer and then incubated with 1 ml of 1:10 diluted plasma samples for one hour. The membranes were then incubated with 1x secondary biotinylated an tibodies diluted in blocking buffer for one hour, followed by a two hour incubation with 1,000 fold diluted labeled-strepavidin in blocking buffer to complete the conjug ated secondary antibody complex. The membranes were washed following each incubation with provided 1x wash buffer
117 solutions. Detection signals for each membrane were detected using Fujifilm AR xray film following a seven second incubation with provided detection buffers. Backlit photographs were taken of the developed film with a Kodak DC290 digital camera. Mean signal intensities minus background signa l intensity for duplicate readings were determined using Kodak 1D Image Analysis Software and standardized to a zero to one scale based on minimum and maximum mean intensity readings for each cytokine. This was necessary because of the naturally occurring large variability (100x-1000x) in levels (based on signal intensities) among various cytokines. Resultant standardized signal intensities were then used for relative cytokine level comparisons among animal groups. Deposition Determinations 6E10 Immunostaining 6E10 immunostaining was perf ormed on mouse brain sect ions at the level of the dorsal hippocampus. Three 25 m coronal sections spaced approximately 600700 m apart were used to analyze diffuse A deposition in sections of the dorsal hippocampus and overlaying parietal cortex. Sections were floated in 10% ethanol and mounted on pre-treated slides. For th e following steps, all incubations were carried out at room temperat ure on a rotating platform shaker unless otherwise noted. In order to induce epitope retrieval, slides were first placed in a 85C 25mM citrate buffer (pH 7.3) bath for five minutes, washed at room temperature in PBS, placed in 88% formic acid for five minutes, and th en washed under low pressure running deionized water for 10 minutes. Slides were then incubated at room temperature with pre-mixed DAKO blocking solution [49.5% DAKO (catalog #X0909; Dako,
118 Carpinteria, CA), 49.5% PBS, .3% H2O2 by volume] for 15 minutes to block endogenous peroxidase activity, with M.O.M. anti-mouse IgG antibody for 60 minutes to block non-specific binding, with a Triton X solution (.4% by volume) in two 2.5-minute PBS baths to induce perm eability, and with protein concentrate solution for 5 minutes to further block non-sp ecific binding. Slides were placed in PBS baths for 5 minutes following both DAKO and Triton X applications. Slides were then incubated with 6E10 primary an tibody overnight at 4C. Following this incubation, slides were placed in three su ccessive PBS baths at 60C for 3.5 minutes each. Slides were then incubated with s econdary IgG antibody in PBS with Triton X and protein concentrate. Slides were then incubated for 30 minutes with ABC complex solution from a NovaRed (Vector) substrate kit, which was pre-mixed and allowed to incubate for approximately 30 minutes prior to application. Three 3minute PBS baths were done following both incubation steps. Slides were then incubated with NovaRed detection reagent mixture for three minutes, followed by a five minute water bath. Slides were then washed in 45%, 60%, 80%, 95%, and 100% ethanol baths for approximately 20 seconds each or until no colors from Nova Red staining ran. A final, four second xylene bath followed by immediate application of cytoseal and coverslipping completed the staining procedure. Thioflavin S Staining Thioflavin S staining was used to detect compact A plaques in the same brain regions used for 6E10 immunostaining. Slides were immersed for five minutes in 1% Thioflavin S in 50% ethanol. Th e same graded alcohol washes and xylene
119 wash scheme used for 6E10 protocol was a pplied, followed by application of cytoseal and coverslipping. Image Analysis All 6E10 and Thioflavin-s stained sec tions were analyzed on a Nikon Eclipse E1000 microscope at either 4x (Thioflavin S) or 10x (6E 10) magnification with Plan Four objective lenses. A Retiga 1300 CCD with QImaging RGB LCD-slider was used to capture images of these sections A Nikon BV-2B fluorescence fliter cube was used for thioflavin S staining. Image Analysis was performed using customized software written in Visual Basic 6.0 (Microsoft) that used Auto-Pro function calls to segment and quantify images according to the established protocols used by Costa et al (2004). A deposition was quantified as a pe rcent of area of interest (=Area stainedtotal/Area Measuredtotal). Statistical Analysis Behavioral Statistical Analysis Interand Intra-group comparisons in pre/post infusion RAWM performance (errors and latency) and post infusion Pl atform Recognition (latency) and Y-maze (% alternation and arm entries) performances were obtained using standard one-way ANOVAs. Two-way repeated measure ANOVAs were also used to compare groups in multi-day tasks (RAWM and Platform Rec ognition). Daily averages over 4 days in Platform Recognition were used for this measure, and three 3-day block averages were used for RAWM. Using the Fisher LSD test, pos-hoc pair-by-pair differences between groups (planned comparisons) were then obtained. Swim speed in RAWM was approximated for each animal by dividing overall Trial 4 + Trial 5 average errors
120 into overall Trial 4 + Trial 5 average latencie s. Non-performers (e.g. repeated circlers etc.) for a given task were not used for sta tistical analysis in that task. 16 of the 19 sensorimotor tasks comprising the Neur ologic Screen we re analyzed nonparametrically using the KruskalWallis test and MannWhitney U -test. The remaining 3 measures (visual cliff, head poke latency, and number of head pokes) were analyzed using ANOVA. Pathological/Histochemical Statistical A nalysis Standard one-way ANOVAs were used to determine group differences for A load determinations (6E10 and Thioflavin S) and relative cytokine signal intensities. Using the Fisher LSD test, pos-hoc pair-by-pair differences between groups (planned comparisons) were then obtained for each measure. Correlation Analysis To determine if a direct relations hip existed between behavioral and pathological measures, correlation analysis was performed using Systat software. A correlation matrix was established between 18 RAWM behavioral measures and 4 A histochemical measures, as well as between these same behavioral measures and 10 plasma cytokine measures. Finally, Y-m aze % alternation and arm entry measures were correlated to A histochemical measures. Factor Analysis Factor Analysis (FA) was performe d using Systat software to group behavioral measures into common, distinct factors. Each fact or corresponds to a discrete component of behavior or cogn ition such as sensory motor function or working memory. This allows for the relati onship between behavior al measures to be
121 determined, and also indicates how individua l task performances may correlate with other task performance levels. FA wa s performed for 13 behavioral measures obtained. Discriminant Function Analysis Discriminant function analysis (DFA ) was performed using 11 similarly loading measures from FA analysis as well as with all 13 behavioral measures used for FA analysis to determine if experime ntal animal groups (Tg+/T cell, Tg+/PBS, and NT) could be distinguished from one another behaviorally. All DFA analyses were performed using Systat software usi ng both direct entry and stepwise forward DFAs. The direct entry method attempts to discriminate between groups using all behavior measures available, while the stepwise-forward method sequentially selects measures based on their variance contribution to best discriminate between the three groups.
122 Results Behavior RAWM Pre-Infusion Testing. In pre-infusion RAWM testing (Fig. 4), transgenic (Tg+) mice were significantly im paired in working memory, as evidenced by an overall group effect for average combined Trial 4 and Trial 5 errors over 4 blocks of testing [F(1,23)=6.6823, p<.02]. Af ter completion of pre-infusion testing, animals in the Tg+ group were divided into two groups for infusion of A -sensitive immune cells (Tg+/T cell group) or PB S (Tg+/PBS group). Pre-infusion RAWM performance of these two sub-groups were identical prior to infusion (data not shown). Post-Infusion Testing. Looking at overall post-infusion RAWM performance, Tg+/PBS mice made significantly more Trial 4+5 errors than Tgmice ( p<.05), while Tg+/T cell errors were not signi ficantly different from either Tgor Tg+/PBS mice (p=n.s.). Thus, overall Tg +/T cell animals performed at a level between that of Tg+ and Tgcontrol groups in RAWM working memory. Looking at individual blocks, Tg+/PBS mice maintained high Trial 4+5 errors during block 1 of post-infusion testing (Fig. 5), whereas Tg+/ T cell mice made significantly fewer Trial 4+5 errors (p<.05) and performed identic ally to Tgmice (p=n.s.). No group
T1 T4+T5 Errors 2345 *Tg+TgRAWM Pre-Infusion Figure 4. Pre-infusion overall working memory performances. Tg+ mice averaged significantly more errors in overall Trial 4+5 testing than Tgmice prior to their division into treatment groups, indicating a clear transgenic impairment in working memory. = significant difference between Tgand Tg+ groups (p<.02). 123
RAWM Block 1Post-Infusion T1 T4+T5 Errors 123456 Tg+/PBSTg+/T cellTg-* Figure 5. Post-infusion Block 1 RAWM errors. In contrast to Tg+/T cell mice, which performed similarly to Tgmice, Tg+/PBS mice averaged significantly more errors in Block 1 Trial 4+5 post-infusion testing. = significant difference between Tg+/PBS and both Tg+/T cell (p<.05) and Tgmice (p<.02). 124
125 differences were evident on the semi-random Trial 1 (Fig. 5). Additionally, no group differences were noted for Tria l 1 or Trial 4+5 errors in e ither block 2 or 3 of testing (data not shown). For Block 3 post-infusion testing (Fig 6) however, both Tg+/T cell mice (p<.05) and Tgmice (p<.05) were able to significantly reduce their combined Trial 4+5 errors from Trial 1 naive performance levels, whereas Tg+/PBS mice were not (p=n.s.). Neither Tg+/T cell nor Tg+/PBS mice were able to improve Trial 1 vs. Trial 4+5 performance in Block 1 and Block 2, whereas Tgmice showed improved Trial 1 vs. Trial 4+5 performance for all 3 blocks of testing (data not shown). To ensure that post-infusion performance differences were not attributable to differences in time taken by animals to na vigate the RAWM apparatus, an average number of seconds per arm choice in overall Trial 4+5 was calculated for each group by dividing overall Trial 4+ 5 latency by overall Trial 4+5 errors. No group differences in number of seconds per choice were noted for this measure (Tg+/T cell: 10.9 1.2, Tg+/PBS: 9.8 1.1, Tg-: 11.2 1.0 sec.). Pre-Infusion vs. Post-Infusion Testing. To further determine the effects of immune cell infusion on worki ng memory performances in Tg+ mice, a block by block comparison of pre-infusion vs. post-infusion performances in the RAWM task was performed for all groups (Fig 7). It was determined that Tg+/PBS mice did not significantly lower their number of errors for any individual block of testing (p=n.s.). In contrast, Tg+/T cell mice did significantly lowered their number of errors from Block 1 pre-infusion testing to Block 1 post-infusion testing (p<.01) as
T1 T4+5 T1 T4+5 Errors 0123456 T1 T4+5 *Tg+/PBSTg+/T cellTg-RAWM Block 3Post-Infusion Figure 6. RAWM during post-infusion Block 3 Trial 1 vs. Trial 4+5. Both Tg+/T cell and Tgmice were able to significantly reduce their number of Trial 4+5 errors compared to Trial 1 naive performances. In contrast, Tg+/PBS mice failed to obtain a significantly lower number of errors for Trial 4+5 compared to Trial 1. = significantly lower Trial 4+5 errors compared to Trial 1 errors (p<.05). 126
Pre-Infusion Post-Infusion RAWM: Pre vs. Post-InfusionTrial 4+5 Combined B1 B2 B3 Tg+/PBS Errors 0123456 B1 B2 B3 Tg+/T cell B1 B2 B3 Tg-** Figure 7. RAWM pre-infusion vs. post-infusion comparisons for Trial 4+5 combined errors by group. Tg+/PBS mice did not significantly lower their errors for any block of post-infusion testing. In contrast, Tg+/T cell mice were able to significantly lower their errors for both Block 1 and Block 2 of post-infusion testing compared with Trial 4+5 error averages for these blocks from pre-infusion testing. Tgmice maintained relatively low errors throughout and thus also lacked significant improvement post-infusion. = significantly lower Trial 4+5 error in post-infusion testing vs. pre-infusion testing (p<.05 or higher level of significance). 127
128 well as from Block 2 pre-infusion testing to Block 2 post-infusion testing (p<.05] but did not show this error reduction for Block 3 performance (p=n.s.). Tgmice did not show significantly decreased errors for any block comparison, most likely due to relatively good pre-infusion performance leve ls. Comparisons of Trial 4+5 errors during the final block of pre-infusion testing vs. the first block of post-infusion testing showed that Tg+/PBS mice became significan tly worse (p<0.05) during initial postinfusion testing (Fig. 8). By contrast, Tg+ mice that received T cell infusions maintained their final pre-infusion perfor mance level, as did Tgcontrols, on the initial post-infusion block (Fig.8). Y-Maze Post-infusion performance in the Y-m aze task revealed a significant overall effect indicating a direct treatmen t effect on basic pneumonic processes [F(2,18)=6.8962; p<.01] (Fig. 9). Tg+/PBS mice demonstrated significantly lower percent spontaneous alterna tions than Tg+/T cell mice (p<.01) or Tgmice (p<.005). In sharp contrast, Tg+/T cell mice perfor med identically to Tgmice. No group differences were noted in number of Y-maze arm entries (data not shown). Platform Recognition Analysis of average latencies for in dividual days and over all 4 days of Platform Recognition testing found no group diffe rences in performance for this task (Fig. 10). There was a highly signifi cant day effect [F(3,66)=7.02; p<.0005), indicating that all animals co llectively improved their perfor mances across the 4 days of testing.
RAWM: Last Block Prevs. First Block Post-Infusion Trial 4+5 Combined PrePostInjection Errors 12345 *Tg/PBSTg/T cellnon-Tg Figure 8. RAWM Trial 4+5 performance for pre-infusion (final block) vs. post-infusion (first block). Tg+/PBS mice significantly worsened their working memory performance from preto post-infusion testing, whereas Tg+/T cell and Tganimals maintained lower levels of errors in post-infusion testing. = significantly higher post-infusion RAWM errors compared to pre-infusion performances (p<.05, paired t-test). 129
Y-Maze % Alternation 020406080 Tg+/T cellTg-Tg+/PBS* Figure 9. Post-infusion Y-maze spontaneous percent alternation performances. Tg+/PBS mice displayed significantly lower percent alternations compared to Tganimals, whereas Tg+/T cell mice performed identically to Tgcontrols. = significantly lower percent alternation compared to Tgand Tg+/T cell mice (p<.01 or higher level of significance). 130
Platform RecognitionDa y s 1234Latency (sec) 102030405060 Tg+/PBS Tg+/T cell TgFigure 10. Post-infusion Platform Recognition latencies by day. No group differences were noted for any day or overall in Platform Recognition performance. 131
132 Neurologic Battery While no group differences were note d between Tg+/T cell and Tg+/PBS groups for the 19 measures of our Neurol ogic Battery (data not shown), direct comparison between these groups and Tganimals did elicit group differences. Tg+/PBS animals showed significantly d ecreased transfer arousal upon placement into a novel environment compared to Tg mice [p<.05]. Tg+/Tcell mice showed decreased transfer arousal as well [p<.05] and also scored significantly lower in evaluation of pelvis [p<.01] and tail eleva tion [p<.05] compared to Tganimals. Finally, Tg+/T cell mice showed an increased score for hypotonic gait qualifications compared to the Tggroup [p<.01]. The l ack of differences between Tg+/T cell and Tg+/PBS groups may indicated that infu sion of T cells did not result in any deleterious neurologic effects in Tg+ mice. Pathology Brain A Deposition Quantification 6E10 Immunohistochemi stry/Thioflavin S Histochemistry. Tg+/T cell and Tg+/PBS mice showed no group differences in % area A plaque deposition, staining either for diffuse deposits (6E10) or compact deposits (Thioflavin S) (Fig. 11). Comparisons were made between groups for both hippocampal and parietal cortex A burdens; no significant differences were noted for either brain region (p=n.s). Plasma Cytokine Analysis For the ten cytokine measures taken (F ig. 12), there were no group differences between Tg+/Tcell and Tg+/PBS mice or between Tg+/PBS and Tgmice (p=n.s).
Hippocampus % 6E10Amyloid Load 0.00.40.81.21.62.0 Parietal Cortex Tg+/PBS Tg+/T cell Hippocampus % Thioflavin SAmyloid Load 0.00.40.81.21.62.0 Parietal Cortex Figure 11. Quantification of 6E10and Thioflavin S A burdens. No group differences between Tg+ groups were noted for either A burden measure in either hippocampus or parietal cortex. 133
GM-CSF IL-12 IFNIL-10 IL-1 Standardized Mean Signal Intensity 0.00.20.40.60.8 IL1IL-2 IL-4 IL-6 TNF***(p70)TgTg+/PBS Tg+/T-Cell Figure 12. Standardized mean signal intensities for 10 plasma cytokines. Tg+/T cell mice showed significantly lower levels of plasma IL-10, TNF-, and GM-CSF compared to Tgmice. = significantly lower mean signal intensity compared to Tganimals (p.05). 134
135 However, the Tg+/Tcell group did display sign ificantly lower plasma cytokine levels than Tgmice for GM-CSF ( p .05), IL-10 ( p<.05), and TNF(p .05). Thus, T cell infusion into Tg+ mice did not result in a sustained, global increase in plasma pro-inflammatory cytokines. Multi-metric Statistical Analysis Correlation Analysis RAWM vs. A Burdens. A correlation matrix comparing A deposition quantifications in hippocam pus and parietal cortex vs. RAWM performances uncovered numerous correla tions between these measures. All correlations were positive unless otherwis e noted (e.g. higher RAWM errors/latency associated with higher A deposition). For all 15 Tg+ animals combined (Table 1), these correlations were found exclusivel y between 6E10 hippocampal measures and the following 8 RAWM working memory measures: overall Trial 4 errors (p .05), Block 1 Trial 4 errors (p<.01), Block 1 Tr ial 4+5 errors (p<.01), overall Trial 4 latency (p<.05), overall Trial 4+5 latency (p .05), Block 1 Trial 4 latency (p .001), Block 1 Trial 5 Latency (p<.05), and Block 1 Trial 4+5 Latency (p .005). No correlations were found between Thioflavin S burdens in either brain area and working memory. Thus, lower hippocampal levels of diffuse A deposits correlated with improved working memory. A strong correlation was noted between RAWM Trial 4+5 errors for Block 1 of post-infusion testing and hippocampal 6E10 burdens when considering all Tg+ mice (p<.01). This correlation is of partic ular interest because Block 1 performance for Tg+/T cell mice was identical to that of Tgmice, while Tg+/PBS mice made
Table 1. For all Tg+ animals, a correlation matrix of RAWM error and latency measures vs. 6E10/Thioflavin S measures of A plaque burdens in the parietal cortex and hippocampus. 136 RAWM Errors Overall Trial 4 Overall Trial 5 Overall Trial 4+5 Block 3 Trial 4 Block 3 Trial 5 Block 3 Trial 4+5 Block 1 Trial 4 Block 1 Trial 5 Block 1 Trial 4+5 6E10: Cortex r p 0.092 0.745 0.166 0.554 0.134 0.635 -0.199 0.477 0.261 0.347 0.022 0.939 0.218 0.436 0.243 0.382 0.248 0.373 6E10: Hippocampus r p 0.512 0.051 0.422 0.118 0.502 0.057 0.135 0.631 0.272 0.326 0.236 0.398 0.713 0.003 0.497 0.059 0.665 0.007 Thioflavin S: Cortex r p -0.133 0.637 -0.001 0.997 -0.078 0.783 -0.091 0.748 0.089 0.751 -0.006 0.982 -0.226 0.418 -0.042 0.882 -0.154 0.583 Thioflavin S: Hippocampus r p 0.238 0.393 0.076 0.789 0.175 0.533 0.11 0.698 0.45 0.092 0.319 0.247 -0.017 0.951 -0.028 0.922 -0.024 0.933 RAWM Latency Overall Trial 4 Overall Trial 5 Overall Trial 4+5 Block 3 Trial 4 Block 3 Trial 5 Block 3 Trial 4+5 Block 1 Trial 4 Block 1 Trial 5 Block 1 Trial 4+5 6E10: Cortex r p 0.19 0.497 0.094 0.738 0.15 0.594 0.048 0.864 0.24 0.389 0.162 0.565 0.309 0.263 0.232 0.405 0.287 0.299 6E10: Hippocampus r p 0.589 0.021 0.384 0.158 0.51 0.052 0.379 0.163 0.337 0.219 0.394 0.146 0.772 0.001 0.515 0.049 0.687 0.005 Thioflavin S: Cortex r p -0.027 0.924 -0.156 0.578 -0.093 0.743 -0.024 0.932 0.025 0.929 0.001 0.996 0.028 0.92 -0.093 0.741 -0.029 0.919 Thioflavin S: Hippocampus r p 0.287 0.299 0.182 0.517 0.246 0.376 0.151 0.591 0.396 0.144 0.305 0.269 0.169 0.546 0.238 0.394 0.211 0.451 Correlations (bold font) were noted between 6E10 diffuse plaque burdens in the hippocampus and several overall and Block 1 RAWM measures of working memory. Most of these correlations were also noted when looking at only Tg+/T cell animals, however none were found when considering Tg+/PBS animals (data not shown), indicating the Tg+/T cell animals as the driving force behind correlations between working memory and diffuse A depostion. r = Pearson product-moment correlation coefficient. p = probability.
137 significantly more e rrors than either group. L ooking at the graph for this comparison (Fig. 13), it is apparent that Tg +/T cell mice are driving the correlation between working memory performance and hippocampal diffuse A burdens. Interestingly, it appears that this correlati on occurs as a result of segregation of the Tg+/T cell group into two subgroups. Of th e seven total Tg+/T cell mice, a group of four mice maintained low errors in Block 1 and displayed low diffuse A burdens in the hippocampus, whereas the remaining three Tg+/T cell mice made a high number of errors in testing and displayed high hippocampal burdens. In contrast, seven of eight Tg+/PBS mice displayed high errors and diffuse A burdens. This stark difference between Tg+/T cell and Tg+/PBS animals clearly illustrates not only how Tg+/T cell animals are driving correlati ons between RAWM and 6E10 measures, but would also seem to indicate that a treatment effect is apparent in the majority of Tg+/T cell animals. It should also be noted that even the rema ining three Tg+/T cell animals with higher Block 1 RAWM errors we re able to reduce their errors by Block 3 of testing, indicating that good working memory performance was not limited to only a portion of the Tg+/T cell animals. When considering only T cell recipient mice (n=7), correlations were also exclusive to hippocampal 6E10 burden when compared to the following measures: overall Trial 4 errors (p<.05) Block 1 Trial 4 errors (p< .05), Block 1 Trial 5 errors (p<.05), Block 1 Trial 4+5 errors (p<.05), Bl ock 1 Trial 4 latency (p<.05), and Block 1 Trial 4+5 latency (p<.05). The lack of correlations between Tg+/PBS mouse RAWM measures and hippocampal 6E10 measures once again indicates that it is the Tg+/T cell mice which are driving these co rrelations for all Tg+ mice combined.
Correlation: RAWM vs. 6E10 0.20.40.60.81.01.21.41.61.8 Hippocampal 6E10 Burden (% Area) 0123456Errors Tg+/PBS Tg+/T Cellr = 0.665p = 0.007 Figure 13. For all Tg+ mice, a correlation between RAWM Block 1 Trial 4+5 Errors and Hippocampal 6E10 burden. A strong correlation between diffuse hippocampal A burden and working memory impairment in Tg+ mice is apparent. This correlation is clearly driven by the Tg+/T cell group, which show a distinct segregation into two subgroups, one with good working memory performance and low A burdens (left side of the graph), and one with poor working memory performance and high A burdens (right side of the graph). 138
139 This is especially true for 6E10 vs. overa ll and Block 1 measures where correlations were found when considering all Tg+ mice as well as Tg+/T cell mice alone, but not when considering Tg+/PBS mice alone. When considering PBS recipient transgenic mice only (n=8), correlations noted were between hippocampal Thioflavin S burdens and the following behavioral measures: Bloc k 3 Trial 5 errors (p <.05), Block 3 Trial 5 latency (p<.01), and Block 3 Trial 4+5 latency (p<.05). Y-maze and Platform Recognition vs. A Burdens. No correlations were observed between A burdens in the hippocampus and parietal cortex vs. Ymaze percent alternations for Tg+/T cell mice or for all Tg+ mice combined (p=n.s.). Similarly, no correlations were noted between A burdens and Day 4 or overall Platform Recognition Latencies fo r these groups of animals. RAWM vs. Plasma Cytokines. A correlation matrix between plasma cytokine levels and RAWM performance m easures was established for all mice, for Tg+ mice combined, and for both PBS recipi ent Tg+ mice and T cell recipient mice alone. All correlations noted were negativ e (e.g. higher plasma cytokines associated with less RAWM errors). For all mice comb ined, 9 negative correlations were found, including one between Block 3 Trial 5 errors and plasma GM-CSF levels (p .05). Negative correlations were also found between IL-1 levels and the following measures: overall Trial 4 errors (p<.05), overall Trial 4+5 errors (p<.05), Block 1 Trial 4 errors (p<.05), and Block 1 Trial 4+5 errors. As for IL-1 correlations, the same negative correlations were also noted for plasma TNFlevels: overall Trial 4 errors (p<.01), overall Trial 4+5 errors (p <.05), Block 1 Trial 4 errors (p<.005), and Block 1 Trial 4+5 errors (p<.05).
140 For all Tg+ mice combined, only 2 negativ e correlations were noted: between IL-4 levels and overall Trial 5 errors (p<.05) and between IL-1 and Block 1 Trial 5 errors (p<.05). When considering only PBS recipient Tg+ mice, the only (negative) correlation noted was between IFNand Block 1 Trial 5 erro rs (p<.05). In sharp contrast, numerous negative correlations (13) were noted between levels of plasma cytokines in T cell recipient Tg+ mice and RAWM measures (Table 2). IL-1 was found to correlate negatively with Bl ock 3 Trial 5 errors (p<.05). IL-1 was found to correlate negatively with overall Trial 4 errors (p .05) and Block 3 Trial 5 errors (p<.05). IL-2 was found to correlate ne gatively with Block 3 Trial 4 errors (p .001). IL-4 was found to correlate negatively with overall Trial 4 errors (p<.05), as well as with Block 1 Trial 5 (p<.05) and Block 1 Trial 4+5 (p<.05) errors. IL-12 was found to correlate negatively with both overall Trial 4 (p<.05) and Block 1 Trial 5 errors (p<.05). Finally, TNFwas found to correlate negativel y with overall Trial 5 errors (p<.005), overall Trial 4+5 errors (p<.05) and Block 1 Trial 4+5 errors (p<.05). The strong negative correlations noted between RAWM measures and both TNFand IL-1 when considering all animals as we ll as T cell animals alone were not noted when looking at all Tg+ animals alone. This would seem to indicate that both T cell and Tgmice are driving these nega tive correlations, implicating plasma IL1and TNFlevels as being particularly importa nt in terms of RAWM performance. Other cytokines such as IL-4, IL-1 and IL-12, which we re only negatively correlated to RAWM performance when c onsidering Tg+/T cell animals, may also have an impact. However, the lack of su ch correlations for all animals combined
Negative correlations (bold font) were noted for numerous measures when considering Tg+/T cell animals alone; correlations whic h did not exist when considering Tg+/PBS mice alone, or both groups of Tg+ mice together. (data not shown). Overal Trial 4 Overall Trial 5 Overall Trial 4+5 Block 1 Trial 4 Block 1 Trial 5 Block 1 Trial 4+5 Block 3 Trial 4 Block 3 Trial 5 Block 3 Trial 4+5 GM-CSF r p 0.057 0.915 0.345 0.503 0.203 0.699 0.456 0.363 0.239 0.648 0.452 0.369 -0.337 0.514 -0.354 0.492 -0.374 0.466 IL-12 r p -0.613 0.195 -0883 0.020 -0.765 0.076 -0.147 0.781 -0.858 0.029 -0.750 0.086 -0.170 0.747 -0.533 0.276 -0.393 0.441 IFNr p -0.234 0.655 -0556 0.252 -0.402 0.429 0.369 0.471 -0.695 0.125 -0.322 0.533 0.108 0.838 -0.414 0.415 -0.183 0.729 IL-10 r p -0.48 0.335 -0.582 0.225 -0.544 0.264 0.390 0.444 -0.692 0.128 -0.307 0.554 0.450 0.37 -0.273 0.6 0.071 0.894 IL-1 r p -0.581 0.227 -0.735 0.096 -0.674 0.142 0.185 0.726 -0.732 0.098 -0.458 0.361 -0.434 0.39 -0.820 0.046 -0.690 0.129 IL-1 r p -0.806 0.053 -0.681 0.137 -0.763 0.077 -0.307 0.554 -0.798 0.057 -0.797 0.058 -0.618 0.191 -0.813 0.049 -0.780 0.067 IL-2 r p -0.515 0.296 -0.454 0.366 -0.497 0.316 -0.433 0.391 -0.411 0.418 -0.572 0.236 -0.972 0.001 -0.762 0.078 -0.929 0.007 IL-4 r p -0.688 0.131 -0854 0.030 -0.790 0.062 -0.278 0.594 -0.844 0.035 -0.815 0.048 -0.566 0.242 -0.761 0.079 -0.723 0.104 IL-6 r p -0.235 0.353 -0.439 0.384 -0.464 0.354 0.499 0.314 -0.680 0.137 -0.235 0.654 0.367 0.474 -0.377 0.461 -0.031 0.954 TNFr p -0.827 0.060 -0950 0.004 -0.892 0.017 -0.369 0.471 -0.790 0.061 -0.827 0.042 -0.437 0.386 -0.643 0.169 -0.591 0.217 Table 2. Correlation matrix for RAWM errors vs. plasma Cytokine measures for Tg+/T cell animals. 141
142 would seem to indicate that they are not universally associated with behavioral performance in this task. Factor Analysis Factor Analysis of 13 post-infusion behavioral measures from RAWM, Y-maze, and Platform Recognition was performed to determine the underlying relationships between these measures/tasks (Table 3). FA involving all 13 be havioral measures resulted in 11 of those measures loadi ng on a common first factor, which accounted for 58.2% of the total variance (a measure was considered significant for loading on a factor if its component loading value exceeded 0.65 for that factor). Factor 1 was thus considered the primary cognition-based factor, as nearly all cognitive measures contributed to it. The two Platform Recognition measures used in this analysis co-loaded in factors 1 and 2. Only one RAWM meas ure (Block 1 Trial 5 errors) and Y-maze percent alternations did not load in the cogni tion-based factor 1. Block 1 Trial 5 errors did not load in any factor, where as Y-maze pe rcent alternations loaded alone in factor 4. Discriminant Function Analysis DFA was utilized to determine if behavi oral performances of the three groups (Tg-, Tg+/PBS, and Tg+/T cell) could distinguish the groups from one another during post-immune cell infusion testing. Both d irect entry and ste pwise-forward DFA methodologies were employed. The direct entry method included all behavioral measures evaluated, while the stepwise-forwar d selects behavioral measures from the total number evaluated based on their cont ribution to the variance. For each methodology, all 13 measures used in Factor Analysis were included. Additionally, both
Table 3. Factor loadings of behavioral measures. Factor Variance (%) Post-Infustion Behavioral Measures RAWM: Overall Trial 4+5 Errors Overall Trial 4 Errors Overall Trial 5 Errors Block 1 Trial 4+5 Errors Block 1 Trial 4 Errors Block 3 Trial 4+5 Errors Block 3 Trial 4 Errors Block 3 Trial 5 Errors Y-maze: Arm Entries Platform Recognition: Overall Latency 1 (58.2) Day 4 Latency 2 (12.3) Platform Recognition: Overall Latency Day 4 Latency 3 (9.6) None Y-maze: 4 (8.7) Percent Alternations RAWM: Did not load: Block 1 Trial 5 Errors Eight of nine RAWM measures loaded with Y-maze entries and two Platform Recognition measures in factor 1the primary cognitive factor. Both Platform Recognition measures co-loaded in factors 1 and 2. Only Y-maze alternations loaded in a separate factor (factor 4), whereas Block 1 Trial 5 RAWM errors did not load in any factor. Percent of total variance explained by a given factor is indicated in bold type within parentheses. 143
144 forms of DFAs were run using only the 11 m easures that loaded on factor 1 of Factor Analysis. Results of these DFAs are summarized in Table 4. Direct entry DFAs, both for all 13 measures and factor 1 measures onl y, could not discriminate between the three groups based on their behavioral performan ce (p=.19; p=.48, respectively). In sharp contrast, the stepwise-forward DFA method (for both behavioral measures sets) was very effective in discriminating between Tg+/PB S mice and the other 2 groups. When all 13 behavioral measures were analyzed, RAWM Block 1 Trial 4+5 erro rs and Platform Recognition Day 4 latencies were retained in step-wise forward DFA. 86% of Tg+/PBS animals were correctly classified by this an alysis, whereas only 50% of Tg+/T cell mice and 60% of Tgmice were correctly classifie d. A similar pattern was noted for step-wise forward analysis of factor 1 behavioral measures, for which RAWM Block 1 Trial 4+5 errors was retained. 88% of Tg+/PBS animals were correctly classified by this analysis. In contrast, only 48% of Tg+/T cell and 50% of Tgmice were correctly classified by this procedure. Thus, while step-wise forward DF As were capable of discriminating between Tg+/PBS groups and the other 2 groups, this analysis could not effectively discriminate between Tgand Tg+/T cell groups, indicating that behavioral performances in these latter two groups were too similar for distinction between them to be made.
Table 4. Summary of discriminant function analyses. Direct Entry Stepwise Forward Measures Significance: Significance: Measures Retained: All 13 *p < .01 RAWM Block 1 Trial 4+5 Errors N.S. PR Day 4 Latency Factor 1 N.S. *p < .05 RAWM Block 1 Trial 4+5 Errors p-values are from Wilkss = significant for Tg+/PBS vs. both Tg+/T cell and Tggroups. 145
146 Discussion General Summary In the present study, we evaluated the be havioral and pathol ogical effects of adoptive transfer of A -sensitive immune cells into AD transgenic mice. These cells consisted of spleenocytes a nd lymphocytes taken from A -immunized, congenic wildtype mice. The cells were cultured to enrich for T cells prior to infusion into recipient animals. We determined that such an infu sion of T cells led to reversal of working memory impairment and also improved basic mnemonic processing at a 1-1 month post-infusion time point. While overall no reductions in brain A deposition were found, strong correlations between diffuse hi ppocampal deposition and RAWM working memory performances in T cell infused mice we re noted, indicating th at cognitive benefit was likely A dependent. It is also important to no te that cognitive benefit in immune cell infused animals was not accompanied by increases in plasma pro-inflammatory cytokines. This would seem to indicate that a systemic infla mmatory response was not evoked by adoptive transfer of immune cells into AD transgenic mice. Results of our Neurologic Battery would seem to confirm this, with no differen ces noted between transg enic controls and T cell recipient mice. Considering the lack of differences between transgenics noted in the Neuorologic Battery, as well as the presence of cognitive benefit extending over several
147 domains following only one immunotherapeutic infusion of T cells, such adoptive immunotherapy shows great promise for futu re human application in treating AD. Behavioral Results Immunotherapeutic techni ques directed against A have been tested in AD transgenic mice since 1999, however only a ha ndful of these studies incorporated comprehensive behavioral testing of treated an imals. In 2000, Morgan et al. found that repeated immunizations with A 1-42 in Freunds adjuvant over an 8 month period resulted in protection against impairment in the RAWM task for working memory, though the 15 month old APPsw+PS1 mice tested did take more trials to l earn the task than non-transgenic controls. In 2003, Austin et al. found that short term (4 biweekly injections) administration of this same vacci ne to aged APPsw+PS1 mice did not result in reversal of impairment in the RAWM task.. In a longitudinal study, Jensen et al. 2005 reported that APPsw+PS1 mice immunized with -42 from 2-16.5 months of age were protected against impairment in RAWM tes ting performed at both 5 and 16 months of age. Thus it has been established that pr otection against impairment in RAWM can be incurred via active A immunization in APPsw+PS1 mice. Ho wever, it is also clear that repeated administrations over an extended tim e period are necessary to establish this behavioral benefit. In the current stu dy, we show that a single immunotherapeutic administration of A -sensitive T cell results in improved RAWM performance at a 1 month post-infusion time point. In pre-infusion RAWM testing, Tg+ mice were clearly impaired in working memory performance compared to Tgmice, making significantly higher Trial 4+5 errors overall. In post-infusion RAWM, Tg+/PBS mice maintained a poor level of working
148 memory performance, with significantly high er Trial 4+5 errors overall compared to Tgmice. In contrast, Tg+/Tcell mice were not significantly different from either Tgor Tg+/PBS mice for this measure, indicating an overall improvement in working memory performance to a level between that of Tgand Tg+ controls. Indeed, for Block 1 of testing, Tg+/T cell mice performed identically to Tganimals in Trial 4+5 errors, while Tg+/PBS mice made significantly more errors than either group. Further demonstrating that improvement in working memory perfor mance was present in Tg+/T cell animals, it was noted that these mice showed a significant learning effect from Trial 1 to Trial 4+5 during the final block of RAWM testing. This effect was not noted for Tg+/PBS mice for any block, whereas Tgmice showed a signifi cant learning effect for each block of postinfusion testing. To further demonstrate the level to which Tg+/T cell animals were able to improve their working memory performances from pre-infusion levels, a block by block comparison of prevs. post-infusion RAWM errors was performed. This analysis revealed an inability of Tg+/PBS animals to improve their performance for any block. In contrast, Tg+/T cell animals did show signifi cant improvements in this measure for both Block 1 and Block 2 of post-in fusion testing. One final analysis of the RAWM task comparing final block prevs. first block post-infusion also demonstrated this effect. Tg+/PBS animals significantly worsened their working memory performance following PBS infusion and a 1 month hiatus from tes ting, whereas Tg+/T cell animals as well as Tganimals maintained significantly low levels in post-infusion testing. Post-infusion and pre-vs. post-infusion analysis revealed that overall and for each block of postinfusion RAWM testing, Tg+/T cell animals di splayed either an improvement in working
149 memory performance from pre-infusion performa nces, and/or an abil ity to perform at a level greater than that of Tg+/PBS controls In many cases, this performance level was identical to Tganimals. This benefit wa s most prominent during the initial phase of post-infusion testing, but was evident thr oughout the entire post-infusion treatment period. Thus, post-infusion analysis c onfirms the potential for a single A -sensitive immune cell infusion to reverse cognitive impairment in transgenic animals. Trinchese et al. (2004) found impairment in APPsw/PS1 mice in RAWM at as early as 3-4 months of age, while our laboratory has found impairment at as early as 5.5 months (Jensen et al. 2005) but has found no impairment at a 3 m onth time point (unpublished observation). These findings combined with prevs. pos t-infusion analysis from the current study confirm that the treatment effect noted a bove was a reversal of cognitive impairment, specifically in the domain of working memory. In the Y-maze task of spontaneous alternations, Tg+/PBS animals displayed significantly lower alternations than both Tg+/T cell and Tganimals. Therefore, it can be said that T cell infused animals were clearly protected against basic mnemonic processing impairment found in un-infused transgenic controls. Thus Y-maze, which is a relatively insensitive task, was still capable of eliciting a treatment effect in 10 month old mice. The results demonstrate how both working memory (RAWM) and basic mnemonic processing (Y-maze) benefited from Tcell infusion in Tg+ mice. In early Ymaze studies, Holcomb et al. (1998; 1999) found decreased YMaze alternations in 3-4 month, 6 month, and 9 month old APPsw/PS1 mi ce, whereas other studies did not find impairment at multiple time points through 16 mo nths (Arendash et al. 2001, Jensen et al. 2005). Thus, the direct treatment effect obser ved confirms that T cell infusion promotes
150 improved performance in the Y-maze task, but does not provide insight into whether this is due to prevention or reversal of impair ment. In 2004, Wilcock et al. showed that 3 monthly infusions of anti-A antibodies were required to see behavioral improvement in the Y-maze task in 22 month old Tg2576 mice. The improved performance noted for this task in the current study occurred one month post-infusion of a single dose of A sensitive T cells. This again underscores the potential of this immunotherapeutic technique to provide cognitive benefi t without numerous administrations. No differences were noted between tr eatment groups in Platform Recognition performance. The relatively poor performance of all 3 groups compared to prior studies (Arendash et al. 2001, Jensen et al. 2005) ma y have resulted from Platform Recognition testing immediately following post-infusion RAWM. All groups may have had difficulty adjusting to the new, working memory-independ ent strategy of this task. However, it should be noted that, for all groups collectively, there was an overall learning effect, with significantly improved latencies from first to last day of testing. Importantly, no differences were found wh en comparing the Neurologic Battery results of Tg+/PBS and Tg+/T cell animals. Th is would seem to indicate a lack of EAE (Experimental Allergic Encephalomyelitis) in T cell animals, which can occur following adoptive T-cell transfer (Kipni s et al. 2002). However, diffe rences noted between Tg+/T cell animals and Tganimals may indicate potential for some hind-leg motor control difficulties in T cell animals, beyond those found in control Tg+ mice. Future studies must look further into the effects of adoptive immunotherapy on neurological function to determine if these effects are consistent and, if so, to identify their cause.
151 Pathology No group differences were noted for 6E 10 or Thioflavin S burdens in the hippocampus or parietal cort ex, indicating no overall reduc tion in diffuse or compact (insoluble) A depostion following T cell infusion into APPsw/PS1 mice. Potential does exist for reductions in soluble A levels, which would go a long way in explaining the behavioral improvement noted in Tg+/T cel l animals. Previous immunotherapeutic studies in AD models have found improved beha vior associated with reduced insoluble A levels (Janus et al. 2000, Wilcock et al. 2004). Additionally, it has been noted in several studies that behavioral improvement can be independent of overall reductions in A deposition (Morgan et al. 2000, Arendash et al. 2001, Jensen et al. 2005) following immunotherapy, as noted in the current study. Th is may be due to reductions in soluble A levels, as other immunothera peutic studies have report ed reductions in soluble A levels in the CNS following immunization (Sigurdsson et al. 2001, Sigurdsson et al. 2004). Future studies must be conducted to analyze soluble fo rms of this peptide, most importantly the A 1-40 and A 1-42 isoforms, to determine if reduced soluble amyloid levels contribute to the cognitive impr ovement noted in Tg+/T cell animals. The lack of significant difference between Plasma cytokine levels in Tg+/T cell and Tg+/PBS groups suggests that T cell infusion into Tg+/T cell mice did not result in a sustained, global immune response. Thus, c ognitive benefit associated with such an infusion was found to occur independently of an enhanced systemic immune response, although strong correlations were evident between plasma cytokines and cognitive
152 performance (see below). However, it should be noted that these cytokine measurements are for plasma only, and may not be indicative of regional brain cytokine profiles. Future studies must be conducted to analyze brain spec ific cytokine levels to determine if this trend for reduced cytokine levels is carri ed over into the CNS following infusion of A sensitive T cells. Compared to Tgmice, th ere was a trend for Tg+/T cell mice to display reduced plasma cytokine levels, w ith significance noted for IL-10, TNF, and GM-CSF vs. Tgcontrols. This reduction in cytoki ne levels may indicate an overall immune system inhibition by immune suppressive regulatory T cells (T regs), or down regulation of specific immune cell types by T helper cell su bsets. The lack of significance between cytokine levels in Tg+/PBS and Tganim als suggests no overt immunological response to A in Tg+ animals. This may be indicative of an established immune tolerance to this human peptide, as would be expected fo r any peptide produced endogenously at high levels throughout the life of these mice. In 2002, Town et al. found that APPsw mice immunized with A 1-42 displayed decreased IFNand increased IL-10 plasma cytokine levels. Levels of IL-2 and IL-4 were deem ed immeasurable in this same study. Based on these findings, the authors proposed that A 1-42 immunization elic ited a Th2 biased response in APPsw mice. Microglia and as trocytes are known to display increased expression of IL-1 TNF, and IL-6 in the brains of Tg2576 animals, however this is not necessarily indicative of plasma leve ls of these cytokines (Benzing et al. 1999, Melhorn et al. 2000). Correlations Correlations to RAWM measures were limited specifically to hippocampal 6E10 diffuse brain A burdens when considering all Tg+ mice. The majority of these
153 correlations were shown to be driven by the Tg+/T cell group however, as strong correlations also exist when considering onl y these animals. Analysis with Tg+/PBS animals only showed correlations with hippocampal Thioflavin S compact brain A burdens exclusively. Thus it is clear that hippocampal A burdens do have an impact on cognition, and diffuse hippocampal burdens seems to play the most important role following infusion of A -sensitive T cells into AD transgenic mice. Without such infusion, compact plaques seem to take on a mo re important role in determining level of impairment. Perhaps the best example of the overall correlation effect involved correlation data between Tg+ animals in RAWM Block 1 Trial 4+5 errors vs. hippocampal 6E10 A burdens. Clearly, Tg+/T cell animals drove this correlation, with 4 animals displaying low errors and low burdens, and 3 animals disp lay higher errors and higher burdens. In contrast, all but 1 Tg+/PBS animal displayed high errors with high A burdens. This correlation may also demonstrate one mode of action for T cell a doptive immunotherapy. Clearly the four Tg+/T cell animals with lower RAWM errors displayed decreased A burdens, and it may be possible that these infu sed animals benefited behaviorally due to clearance of hippocampal diffuse A thus explaining why this measure was of particular importance. We have previously demonstrated that cognitive performance in a number of tasks can correlate with brain A burdens in AD transgenic mice (Leighty et al. 2004, Gordon et al. 2001, Arendash et al. 2001b), though these correlations are not always present (Jensen et al. 2005). When presen t, correlations often i nvolve the RAWM task,
154 although they can also involve Morris Water Maze and Platform Recognition (Leighty et al. 2004). RAWM is a task for working memory, thus performance in this task is heavily hippocampus dependent. This directly explai ns the cognitive bene fit noted during Block 1 of post-infusion testing in 4 out of 7 Tg+/ T cell animals. What these correlations do confirm is that both diffuse and compact A burdens in the hippo campus can contribute to cognitive impairment in Tg+ animals, a nd also that modest reductions in diffuse hippocampal A burden are associated with improve d performance in the RAWM task. Several correlations were noted between RAWM measures and plasma cytokine levels when considering all animals. Higher pl asma cytokine levels correlated with better cognitive performance. All but one of these correlations were between either IL-1 or TNFand RAWM measures. These and seve ral other correlations were also found when looking at Tg+/T cell animals alone but not when looking at Tg+ animals combined or Tg+/PBS animals alone. Clearly then, these negative correlations are being driven by both Tgand Tg+/T cell animals. These correlations were all negative (e.g. higher plasma cytokines with lower RAWM e rrors). Therefore, it would seem that despite a trend for decreased plasma cytokine levels in Tg+/T cell animals, higher levels of individual cytokines (especially IL-1 and TNF) may play a role in improving cognitive performance capabilitie s. It is interesting that the two key cytokines driving these correlations are pro-inflammatory cy tokines which have been found to be overexpressed by microglia local to A plaques in transgenic models of AD (Benzing et al. 1999, Akiyama et al. 2000). In humans with AD, CSF levels of both cytokines are increased (Cacabelos et al. 1991, Tarkowski et al. 1999). However, several lines of study suggest that while these cyt okines are involved in the neuroinflammatory cascade of AD,
155 under situations of appropriate immune activat ion they may actually act beneficially in the CNS to reduce inflammatory damage (C orreale and Villa, 2004). This may indicate that the actions of certain cytokine produc ing cells may be important in determining extent of behavioral improvement. Once again, this emphasizes the importance of analyzing brain cytokine levels, and perhaps also quantifying levels of activated immune cells such as microglia, astrocytes and T ce lls that would contri bute to production of these cytokines in the brain. Factor Analysis/Discriminant Function Analysis All but one RAWM measure loaded in Fact or 1 of FA, as did both PR measures, indicating this factor as cognitively-base d. Stepwise forward DFA results for all behavioral measures as well as Factor 1 meas ures alone indicate that Tg+/T cell animal performances could not be separated from Tgperformances, whereas Tg+/PBS animal performance levels were poor enough that they were clearly distinguished from the other two, better performing groups. Our laboratory has shown in the past that stepwise forward DFA analysis is capable of di stinguishing not only between transgenic treatment groups and wild-type mice, but also between different strains of transgenic mice (Jensen et al. 2005, Leighty et al. 2004). Thus, the lack of discrimination between Tg+/T cell animals and Tganimals when applying this methodology across multiple cognitive measures strongly emphasizes the similar perfor mance levels of these two groups. This reinforces the hypothesis that infusion of A -sensitive immune cells leads to normalization of Tg+ animal performan ces in RAWM to that of Tglevels.
156 Proposed Mechanisms of A -Sensitive T Cell Mediated Cognitive Improvement in APPsw/PS1 Mice Although the current literature is limited with regards to application of adoptive immunotherapy in neurodegenerative diseases, it is possible to speculate as to the mechanisms by which infused T cells could act therapeutically in a mouse model for AD. It has been shown that immuno-tolerance to A is present in APP transgenic mice (Monsonego et al. 2001), and it is therefore necessary to administer A vaccinations in adjuvant designed to activate the immune syst ems to elicit an effect. One mechanism of action for A -sensitive T cells infused into such an immuno-tolerant environment could be to break this tolerance and initiate an immune system-based clearance of A from the CNS. Neuroinflammation is thought to be a major contributor to the progression of AD, and is present to a large extent in AD m ouse models. A second mechanism for T cell action could be for regulatory T cells and/or Th2 cells to down-regulate this type of cellular-based neuroinflammatory response. Fi nally, it is possible for T cells and the cells they activate (astrocytes, microglia etc.), provide neuroprotection via a number of mechanisms. Stimulation of the immune system ag ainst an endogenous self peptide, A T helper cells play a central role in determining the type of immune response elicited against a given immunogen. Depending on the major sub-type of T cell involved in stimulating such an immune response against A it is likely that one of two reactions, either cellularly or humorally based, would be elicited against this peptide. If Th1 T helper cells were the major contributor to wards developing this reaction, a phagocytic immune cell activation would be anticipated. This would potentially result in clearance
157 of deposited amyloid via phagocytic mechanis ms, a mechanism of clearance found to be present in several active immuni zation studies in AD models (D as et al. 2003a, Das et al. 2003b). In the CNS, the primary phagocytic cells are the microglia and thus such a reaction would most likely be a ssociated with increased micr oglial activation in the areas of A plaques (Wilcock et al. 2001). While correlation data from the current study suggests that decreased diffuse hippocampal A may mediate RAWM improvement in some T cell infused mice, no overall reductions in A pathology were noted, indicating that this effect was not widespread in Tg+/ T cell mice. In future s studies, it will be important to determine to what extent mi croglia and other components of the cellular immune system are activated in the CNS of T cell infused mice. If following T cell infusion a Th2 biased re sponse was elicited, in contrast to the cellular based response to A promoted by Th1 cells, a humoral response would be anticipated instead. This would be mediat ed by the ability of Th2 subset cells to stimulate B cells to produce antibodies, in this case antibodies specific to A Based on the results of passive immuni zations studies in models of AD (DeMattos et al. 2001, Wilcock et al. 2004), it is likely that such a humoral based immune response would also trigger the clearance of insoluble as well as soluble A from the CNS via standard clearance mechanisms for antigen-antibody im mune complexes. A down-regulation of cellular immunity present in Th1-biased imm une reactions would also be expected in a Th2-biased response, perhaps leading to decreased neuroinflammation as discussed below. To determine if these effects were pr esent in T cell animals, it may be beneficial to look at antibody titers and regional brain cytokine le vels in future studies.
158 Overall down-regulation of the immune syst em, or of specific immune cells which may contribute to neuroinflammation. As previously mentioned, microglia a nd other cellular components of the CNS immune system may be down-regulated by the T h2 subset of T helper cells. This would be especially beneficial in the Tg+ animal br ain paradigm, in which immune cells such as microglia are perpetually activated by A and A associated neuronal damage. Without the influence of regulatory cells, this form of activation may result in a state of frustrated phagocytosis, which results in production of neuroinflammatory mediators involved both in inappropriately activating other cells (s uch as astrocytes) and in directly causing damage/dysfunction in adjacent neurons (Str eit et al. 2004, DAndrea et al. 2004). Th2 subset cells could down-regulate such ac tivation, blocking pe rpetuation of the AD neuroinflammatory cascade. Regulatory T cells may also be involved in this process, down-regulating numerous immune cell types indi rectly via de-activation of T-cells or directly via interactions with microglia and other components of the cellular immune system (Avidan et al. 2004). Again, in futu re studies, looking at regional cytokine profiles would be beneficial to determine if such an effect was present. Analysis of activated immune cell levels in the CNS would also be extremely beneficial in determining the exact actions of adoptively transferred T cells. Neuronal protection The potential for T cells to enhance neur onal health and protect against neuronal damage has been established in the current literature (Moalem et al. 2000, Benner et al. 2004, Avidan et al. 2004). This could involve a down-regulation of immune cells which instigate neuroinflammation (and resulting ne uronal damage) as described above, or a
159 direct production and/or indu ced production of neuroprotective agents by astrocytes. Whether cells in the A -sensitive T cell populations infused into Tg+ animals possess this potential is unknown, and analysis of eith er neurotrophic factor levels in the CNS would indicate if this is a viable method by which T cells provide cognitive benefit in these animals. Clinical Implication of T cell Study Findi ngs and Proposed Future Investigations As with all immunologically based protocols, special care must be taken before application of such protocols in humans. Fu rther investigation into the potential for A sensitive immune cells to promote EAE or ot her deleterious auto-immune type reactions is of the utmost importance, especially in lieu of the adverse e ffects resulting from clinical application of active A immunotherapy (Orgogozo et al. 2003). Additionally, to help elucidate the exact mechanism of cognitive improvement/protection provided by immune cell infusion, it may be beneficial for futures studies to l ook at the effects of infusion of specific T cell subsets into transg enic animals. Analysis of the effects of these subsets on behavior and pathology, as we ll as added analysis of brain cytokines, activated immune cell (such as microglia) st ains, and neurotrophic factor levels may provide insight into the mechanisms by which A -sensitive immune cells improve cognitive-based performances in these AD tran sgenic mice. In addition, a comparison of neuronal health and activity, es pecially in the hippocampus, in treated vs. untreated Tg+ animals may elucidate the exact mechanism by which cognitive performances in tasks such as the RAWM are being improved. These and additional studies would be desirable before application of this pro cedure in humans is initiated.
160 From a methodological standpoint, A -sensitive T-cells could be isolated from an AD patients blood, or even spleen, expanded in vitro and returned to the patient on a periodic basis. While invasive, this proce ss is not unprecedented a nd has been used in donor lymphocyte infusion protocols for many years (Kolb et al. 1995). Only one infusion of T cells into AD transgenic mice was necessary to observe cognitive improvement 1-1 months later in this study. Future studies would be necessary to determine if these effects would continue without additional infusions. If such followups were unnecessary or could be spaced far apart, this immunotherapeutic approach may prove to be less invasive th an other immunotherapeutic appr oaches to AD (specifically active and passive immunization therapies against A ), which require regular injections of either A in adjuvant or anti-A antibodies. Both of these forms of treatment have potential for side effects as elucidated by studies both in animal models of AD and, for active immunotherapy, in humans. Both active and passive immunotherapeutic approaches have been shown to induce me ningoencephalitis (Lee et al. 2005) due to excessive immune system activation. This form of inflammation may not be present following a single infusion of A -sensitive T cells independent of adjuvant. Findings of reduced levels of cytokines in the plasma of T cell infused mice would seem to suggest that inflammation such as that noted in human active A immunization trials is not present, though further work will confirm or refute this hypothesis. Thus, adoptive immunotherapy may provide a means of providi ng long term cognitive benefits in AD, with decreased potential for negative side-effects.
161 References Akiyama H, Barger S, Barnum S, Br adt B, Bauer J, Cole GM, Cooper NR, Eikelenboom P, Emmerling M, Fiebich BL Finch CE, Frautschy S, Griffin WS, Hampel H, Hull M, Landreth G, Lue L, Mrak R, Mackenzie IR, McGeer PL, O'Banion MK, Pachter J, Pasinetti G, Plata-Salaman C, Rogers J, Rydel R, Shen Y, Streit W, Strohmeyer R, Tooyoma I, Va n Muiswinkel FL, Veerhuis R, Walker D, Webster S, Wegrzyniak B, Wenk G, WyssCoray T. Inflammation and Alzheimer's disease. Neurobiol Ag ing. 2000 May-Jun;21(3):383-421. Alzheimer A, Stelzmann RA, Schnitzlein HN, Murtagh FR. An English translation of Alzheimer's 1907 paper, "Uber eine eige nartige Erkankung der Hirnrinde". Clin Anat. 1995;8(6):429-31. Andreasen N, Sjogren M, Blennow K. C SF markers for Alzheimer's disease: total tau, phospho-tau and Abeta42. World J Biol Psychiatry. 2003 Oct;4(4):147-55. Areosa SA, Sherriff F, McShane R. Memantine for dementia. Cochrane Database Syst Rev. 2005 Apr 18;(2):CD003154. Arendash GW, King DL, Gordon MN, Morg an D, Hatcher JM, Hope CE, Diamond DM. Progressive, age-relate d behavioral impairments in transgenic mice carrying both mutant amyloid precursor protein and presenilin-1 transgenes. Brain Res. 2001 Feb 9;891(1-2):42-53. Arendash GW, Gordon MN, Diamond DM, Austin LA, Hatcher JM, Jantzen P, DiCarlo G, Wilcock D, Morgan D. Behavi oral assessment of Alzheimer's transgenic mice following long-term Abeta vaccination: task specificity and correlations between Abeta deposition and spat ial memory. DNA Cell Biol. 2001 Nov;20(11):737-44. Arriagada PV, Growdon JH, Hedley-Whyte ET, Hyman BT. Neurofibrillary tangles but not senile plaques parallel duration and severity of Alzheimer's disease. Neurology. 1992 Mar;42(3 Pt 1):631-9. Ashe KH. Learning and memory in transg enic mice modeling Alzheimer's disease. Learn Mem. 2001 Nov-Dec;8(6):301-8.
162 Austin L, Arendash GW, Gordon MN, Diamond DM, DiCarlo G, Dickey C, Ugen K, Morgan D. Short-term beta-amyloid vaccinations do not improve cognitive performance in cognitively impaired APP + PS1 mice. Behav Neurosci. 2003 Jun;117(3):478-84. Avidan H, Kipnis J, Butovsky O, Casp i RR, Schwartz M. Vaccination with autoantigen protects against aggregated beta-amyloid and glutamate toxicity by controlling microglia: eff ect of CD4+CD25+ T cells. Eur J Immunol. 2004 Dec;34(12):3434-45. Bard F, Cannon C, Barbour R, Burke RL, Games D, Grajeda H, Guido T, Hu K, Huang J, Johnson-Wood K, Khan K, Kholodenko D, Lee M, Lieberburg I, Motter R, Nguyen M, Soriano F, Vasquez N, Weiss K, Welch B, Seubert P, Schenk D, Yednock T. Peripherally administered anti bodies against amyloi d beta-peptide enter the central nervous system and reduce pa thology in a mouse model of Alzheimer disease. Nat Med. 2000 Aug;6(8):916-9. Barnes J, Scahill RI, Boyes RG, Frost C, Lewis EB, Rossor CL, Rossor MN, Fox NC. Differentiating AD from aging us ing semiautomated measurement of hippocampal atrophy rates. Ne uroimage. 2004 Oct;23(2):574-81. Benner EJ, Mosley RL, Destache CJ, Lewis TB, Jackson-Lewis V, Gorantla S, Nemachek C, Green SR, Przedborski S, Gendelman HE. Therapeutic immunization protects dopaminergic neurons in a mouse m odel of Parkinson's disease. Proc Natl Acad Sci U S A. 2004 Jun 22;101(25):9435-40. Epub 2004 Jun 14. Benzing WC, Wujek JR, Ward EK, Shaffe r D, Ashe KH, Younkin SG, Brunden KR. Evidence for glial-mediated inflammation in aged APP(SW) transgenic mice. Neurobiol Aging. 1999 Nov-Dec;20(6):581-9. Berman K, Brodaty H. Tocopherol (vitam in E) in Alzheimer's disease and other neurodegenerative disorders. CNS Drugs. 2004;18(12):807-25. Bernard M, Dauriac C, Drenou B, Leberre C, Branger B, Fauchet R, Le Prise PY, Lamy T. Long-term follow-up of allogeneic bone marrow transplantation in patients with poor prognosis non-Hodgkin's lympho ma. Bone Marrow Transplant. 1999 Feb;23(4):329-33. Bigler ED, Neeley ES, Miller MJ, Tate DF, Rice SA, Cleavinger H, Wolfson L, Tschanz J, Welsh-Bohmer K. Cerebral volume loss, cogn itive deficit and neuropsychological performance: compar ative measures of brain atrophy: I. Dementia. J Int Neuropsychol Soc. 2004 May;10(3):442-52.
163 Borchelt DR, Thinakaran G, Eckman CB, Lee MK, Davenport F, Ratovitsky T, Prada CM, Kim G, Seekins S, Yager D, Slunt HH, Wang R, Seeger M, Levey AI, Gandy SE, Copeland NG, Jenkins NA, Price DL, Younkin SG, Sisodia SS. Familial Alzheimer's disease-linked presenilin 1 vari ants elevate Abeta142/1-40 ratio in vitro and in vivo. Neuron. 1996 Nov;17(5):1005-13. Borchelt DR, Ratovitski T, van Lare J, Lee MK, Gonzales V, Jenkins NA, Copeland NG, Price DL, Sisodia SS. Accelerated amyloid deposition in the brains of transgenic mice coexpressing mutant presen ilin 1 and amyloid precursor proteins. Neuron. 1997 Oct;19(4):939-45. Bourre JM. Roles of unsaturated fatty aci ds (especially omega-3 fatty acids) in the brain at various ages and during agei ng. J Nutr Health Aging. 2004;8(3):163-74. Bressoud A, Chapuis B, Roux E, Cabrol C, Jeannet M, Roosnek E, Helg C. Donor lymphocyte infusion for a patient with re lapsing myelodysplastic syndrome after allogeneic bone marrow transplant ation. Blood. 1996 Sep 1;88(5):1902-3. Bussiere T, Friend PD, Sadeghi N, Wicins ki B, Lin GI, Bouras C, Giannakopoulos P, Robakis NK, Morrison JH, Perl DP, Hof PR. Stereologic assessment of the total cortical volume occupied by amyloid deposits and its relationship with cognitive status in aging and Alzheimer's dis ease. Neuroscience. 2002;112(1):75-91. Cacabelos R, Barquero M, Garcia P, Alva rez XA, Varela de Seijas E. Cerebrospinal fluid interleukin-1 beta (IL-1 beta) in Alzheimer's disease and neurological disorders. Methods Find Exp C lin Pharmacol. 1991 Sep;13(7):455-8. Calhoun ME, Burgermeister P, Phinney AL, Stalder M, Tolnay M, Wiederhold KH, Abramowski D, Sturchler-Pierrat C, Sommer B, Staufenbiel M, Jucker M. Neuronal overexpression of mutant amyloid precursor protein results in prominent deposition of cerebrovascular amyloid. Proc Natl Acad Sci U S A. 1999 Nov 23;96(24):1408893. Canevari L, Abramov AY, Duchen MR. Toxi city of amyloid beta peptide: tales of calcium, mitochondria, and oxidative stre ss. Neurochem Res. 2004 Mar;29(3):63750. Chabot S, Williams G, Hamilton M, Suth erland G, Yong VW. Mechanisms of IL10 production in human microglia-T cell interaction. J Immunol. 1999 Jun 1;162(11):6819-28. Chandok MR, Farber DL. Signaling cont rol of memory T cell generation and function. Semin Immunol. 2004 Oct;16(5):285-93.
164 Chapman PF, White GL, Jones MW, Cooper-Bla cketer D, Marshall VJ, Irizarry M, Younkin L, Good MA, Bliss TV, Hyman BT, Younkin SG, Hsiao KK. Impaired synaptic plasticity and learni ng in aged amyloid precursor protein transgenic mice. Nat Neurosci. 1999 Mar;2(3):271-6. Chauhan NB, Siegel GJ. Reversal of amyloid beta toxicity in Alzheimer's disease model Tg2576 by intraventricul ar antiamyloid beta anti body. J Neurosci Res. 2002 Jul 1;69(1):10-23. Chauhan NB, Siegel GJ. Intracerebroventricular passive immunization with antiAbeta antibody in Tg2576. J Neuros ci Res. 2003 Oct 1;74(1):142-7. Chen G, Chen KS, Knox J, Inglis J, Be rnard A, Martin SJ, Justice A, McConlogue L, Games D, Freedman SB, Morris RG. A l earning deficit related to age and betaamyloid plaques in a mouse model of Alzheimer's disease. Nature. 2000 Dec 2128;408(6815):975-9. Collins RH Jr, Shpilberg O, Drobyski WR, Porter DL, Giralt S, Champlin R, Goodman SA, Wolff SN, Hu W, Verfaillie C, List A, Dalton W, Ognoskie N, Chetrit A, Antin JH, Nemunaitis J. Donor leukocyte infusions in 140 patients with relapsed malignancy after allogeneic bone marrow transplantation. J Clin Oncol. 1997 Feb;15(2):433-44. Conte V, Uryu K, Fujimoto S, Yao Y, Rokach J, Longhi L, Trojanowski JQ, Lee VM, McIntosh TK, Pratico D. Vitamin E reduces amyloidosis and improves cognitive function in Tg2576 mice following re petitive concussive brain injury. J Neurochem. 2004 Aug;90(3):758-64. Correale J, Villa A. The neuroprotective role of inflammation in nervous system injuries. J Neurol. 2004 Nov;251(11):1304-16. Cribbs DH, Ghochikyan A, Vasilevko V, Tran M, Petrushina I, Sadzikava N, Babikyan D, Kesslak P, Kieber-Emmons T, Cotman CW, Agadjanyan MG. Adjuvant-dependent modulation of Th1 a nd Th2 responses to immunization with beta-amyloid. Int Immunol. 2003 Apr;15(4):505-14. Crow MK. Costimulatory molecules a nd T-cell-B-cell interactions. Rheum Dis Clin North Am. 2004 Feb;30(1):175-91, vii-viii. Cummings JL. Drug Therapy: Alzheime rs Disease. N Engl J Med. 2004 July; 351(1):56-67. Dall'Igna OP, Souza DO, Lara DR. Caffeine as a neuroprotective adenosine receptor antagonist. Ann Pharmacother. 2004 Apr;38(4):717-8. Epub 2004 Feb 24.
165 D'Andrea MR, Cole GM, Ard MD. The microglial phagocytic role with specific plaque types in the Alzheimer dis ease brain. Neurobiol Aging. 2004 MayJun;25(5):675-83. Davis HS, Rockwood K. Conceptualization of mild cognitive impairment: a review. Int J Geriatr Psychiat ry. 2004 Apr;19(4):313-9. Das P, Murphy MP, Younkin LH, Younkin SG, Golde TE. Reduced effectiveness of Abeta1-42 immunization in APP transgenic mice with significant amyloid deposition. Neurobiol Ag ing. 2001 Sep-Oct;22(5):721-7. Das P, Howard V, Loosbrock N, Dickson D, Murphy MP, Golde TE. Amyloid-beta immunization effectively reduces amyloi d deposition in FcRgamma-/knock-out mice. J Neurosci. 2003 Sep 17;23(24):8532-8. Das P, Chapoval S, Howard V, David CS Golde TE. Immune responses against Abeta1-42 in HLA class II transgenic mi ce: implications for Abeta1-42 immunemediated therapies. Neurobiol Aging. 2003 Nov;24(7):969-76. Deane R, Du Yan S, Submamaryan RK, LaRue B, Jovanovic S, Hogg E, Welch D, Manness L, Lin C, Yu J, Zhu H, Ghiso J, Frangione B, Stern A, Schmidt AM, Armstrong DL, Arnold B, Liliensiek B, Nawr oth P, Hofman F, Kindy M, Stern D, Zlokovic B. RAGE mediates amyloid-beta peptide transport across the blood-brain barrier and accumulation in brai n. Nat Med. 2003 Jul;9(7):907-13. de Leon MJ, Convit A, DeSanti S, Bobi nski M, George AE, Wisniewski HM, Rusinek H, Carroll R, Saint Louis LA. C ontribution of structural neuroimaging to the early diagnosis of Alzheimer's dis ease. Int Psychogeriatr. 1997;9 Suppl 1:18390; discussion 247-52. DeMattos RB, Bales KR, Cummins DJ, Dodart JC, Paul SM, Holtzman DM. Peripheral anti-A beta antibody alters CN S and plasma A beta clearance and decreases brain A beta burden in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A. 2001 Jul 17;98(15):8850-5. Epub 2001 Jul 03. DiCarlo G, Wilcock D, Henderson D, Gor don M, Morgan D. Intrahippocampal LPS injections reduce Abeta load in APP+PS1 transgenic mice. Neurobiol Aging. 2001 Nov-Dec;22(6):1007-12. Dickey CA, Morgan DG, Kudchodkar S, Weiner DB, Bai Y, Cao C, Gordon MN, Ugen KE. Duration and specificity of humoral immune responses in mice vaccinated with the Alzheimer's disease-a ssociated beta-amyloid 1-42 peptide. DNA Cell Biol. 2001 Nov;20(11):723-9.
166 Dickey CA, Loring JF, Montgomery J, Gordon MN, Eastman PS, Morgan D. Selectively reduced expression of synap tic plasticity-related genes in amyloid precursor protein + presenilin-1 tr ansgenic mice. J Neurosci. 2003 Jun 15;23(12):5219-26. Dodart JC, Meziane H, Mathis C, Bale s KR, Paul SM, Ungerer A. Behavioral disturbances in transgenic mice overexpressing the V717F beta -amyloid precursor protein. Behav Neurosci. 1999 Oct;113(5):982-90. Dodart JC, Mathis C, Saura J, Bales KR Paul SM, Ungerer A. Neuroanatomical abnormalities in behaviorally character ized APP(V717F) transgenic mice. Neurobiol Dis. 2000 Apr;7(2):71-85. Dodart JC, Bales KR, Gannon KS, Greene SJ, DeMattos RB, Mathis C, DeLong CA, Wu S, Wu X, Holtzman DM, Paul SM. Immunization reverses memory deficits without reducing brain Abeta burden in Alzheimer's disease model. Nat Neurosci. 2002 May;5(5):452-7. Doody RS, Stevens JC, Beck C, Dubins ky RM, Kaye JA, Gwyther L, Mohs RC, Thal LJ, Whitehouse PJ, DeKosky ST, Cummings JL. Practice parameter: management of dementia (an evidence-b ased review). Report of the Quality Standards Subcommittee of the Americ an Academy of Neurology. Neurology. 2001 May 8;56(9):1154-66. Duff K, Eckman C, Zehr C, Yu X, Prada CM, Perez-tur J, Hutton M, Buee L, Harigaya Y, Yager D, Morgan D, Gordon MN, Holcomb L, Refolo L, Zenk B, Hardy J, Younkin S. Increased amyloidbeta42(43) in brains of mice expressing mutant presenilin 1. Nature. 1996 Oct 24;383(6602):710-3. Eckert A, Marques CA, Keil U, Schussel K, Muller WE. Increased apoptotic cell death in sporadic and genetic Alzheime r's disease. Ann N Y Acad Sci. 2003 Dec;1010:604-9. Eckman CB, Mehta ND, Crook R, Perez-tur J, Prihar G, Pfeiffer E, Graff-Radford N, Hinder P, Yager D, Zenk B, Refolo LM, Prada CM, Younkin SG, Hutton M, Hardy J. A new pathogenic mutation in the APP gene (I716V) increases the relative proportion of A beta 42(43). Hum Mol Genet. 1997 Nov;6(12):2087-9. Ellinson M, Thomas J, Patterson A. A crit ical evaluation of the relationship between serum vitamin B, folate and total homocyst eine with cognitive impairment in the elderly. J Hum Nutr Diet. 2004 Aug;17(4):371-83; quiz 385-7.
167 Farkas IG, Czigner A, Farkas E, Dobo E, Soos K, Penke B, Endresz V, Mihaly A. Beta-amyloid peptide-induced blood-brain barrier disruption facilitates T-cell entry into the rat brain. Acta Histochem. 2003;105(2):115-25. Farmery MR, Tjernberg LO, Pursglove SE Bergman A, Winblad B, Naslund J. Partial purification and characterization of gamma-secretase from post-mortem human brain. J Biol Chem. 2003 Jul 4;278(27):24277-84. Epub 2003 Apr 15. Feldman H, Scheltens P, Scarpini E, Hermann N, Mesenbrink P, Mancione L, Tekin S, Lane R, Ferris S. Behavioral symp toms in mild cognitive impairment. Neurology. 2004 Apr 13;62(7):1199-201. Fleminger S, Oliver DL, Lovestone S, et al. Head injury as a risk factor for Alzheimers disease: the evidence 10 y ears on; a partial replication. J Neurol Neurosurg Psychiatry 2003;74:857. Fratiglioni L, Paillard-Borg S, Winblad B. An active and socially integrated lifestyle in late life might protect against dementia. Lancet Neurol. 2004 Jun;3(6):343-53. Frautschy SA, Yang F, Irrizarry M, Hyma n B, Saido TC, Hsiao K, Cole GM. Microglial response to amyloid plaques in APPsw transgenic mice. Am J Pathol. 1998 Jan;152(1):307-17. Frautschy SA, Cole GM, Baird A. Phagoc ytosis and deposition of vascular betaamyloid in rat brains injected with Alzheimer beta-amyloid. Am J Pathol. 1992 Jun;140(6):1389-99. Frenkel D, Balass M, Solomon B. N-term inal EFRH sequence of Alzheimer's betaamyloid peptide represents the epitope of its anti-aggregating antibodies. J Neuroimmunol. 1998 Aug 1;88(1-2):85-90. Frenkel D, Balass M, Katchalski-Katzir E, Solomon B. High affinity binding of monoclonal antibodies to the sequential epitope EFRH of beta-amyloid peptide is essential for modulation of fibrillar aggregation. J Neuroimmunol. 1999 Mar 1;95(1-2):136-42. Frenkel D, Katz O, Solomon B. Immuni zation against Alzheimer's beta -amyloid plaques via EFRH phage administration. Proc Natl Acad Sci U S A. 2000 Oct 10;97(21):11455-9. Frenkel D, Dewachter I, Van Leuven F, Solomon B. Reduction of beta-amyloid plaques in brain of transgenic mouse m odel of Alzheimer's disease by EFRH-phage immunization. Vaccine. 2003 Mar 7;21(11-12):1060-5.
168 Furlan R, Brambilla E, Sanvito F, Roccatagliata L, Olivieri S, Bergami A, Pluchino S, Uccelli A, Comi G, Martino G. Vacci nation with amyloid-beta peptide induces autoimmune encephalomyelitis in C57/BL 6 mice. Brain. 2003 Feb;126(Pt 2):28591. Games D, Adams D, Alessandrini R, Barbour R, Berthelette P, Blackwell C, Carr T, Clemens J, Donaldson T, Gillespie F, et al. Alzheimer-type neuropathology in transgenic mice overexpressing V717F beta-amyloid precursor protein. Nature. 1995 Feb 9;373(6514):523-7. Gaskell PC Jr, Vance JM. Alzheimer's dis ease genes and genetic testing in clinical practice. JAAPA. 2004 Mar;17(3):25-6, 29-30, 32. Giacchino J, Criado JR, Games D, Henrikse n S. In vivo synaptic transmission in young and aged amyloid precursor protein tr ansgenic mice. Brain Res. 2000 Sep 8;876(1-2):185-90. Glenner GG, Wong CW. Alzheimer's dis ease and Down's syndrome: sharing of a unique cerebrovascular amyloid fibril protein. Biochem Biophys Res Commun. 1984 Aug 16;122(3):1131-5. Gordon MN, King DL, Diamond DM, Jantzen PT, Boyett KV, Hope CE, Hatcher JM, DiCarlo G, Gottschall WP, Morgan D, Arendash GW. Correlation between cognitive deficits and Abeta deposits in transgenic APP+PS1 mice. Neurobiol Aging. 2001 May-Jun;22(3):377-85. Gordon MN, Holcomb LA, Jantzen PT, Di Carlo G, Wilcock D, Boyett KW, Connor K, Melachrino J, O'Callaghan JP, Morgan D. Time course of the development of Alzheimer-like pathology in the doubly tr ansgenic PS1+APP mouse. Exp Neurol. 2002 Feb;173(2):183-95. Greenberg SM. Cerebral amyloid angiopath y: prospects for clinical diagnosis and treatment. Neurology. 1998 Sep;51(3):690-4. Greene JD, Baddeley AD, Hodges JR. Analys is of the episodic memory deficit in early Alzheimer's disease: evidence from the doors and people test. Neuropsychologia. 1996 Jun;34(6):537-51. Grundman M, Jack CR Jr, Petersen RC, Ki m HT, Taylor C, Datvian M, Weiner MF, DeCarli C, DeKosky ST, van Dyck C, Darv esh S, Yaffe K, Kaye J, Ferris SH, Thomas RG, Thal LJ; Alzheimer's Di sease Cooperative Study. Hippocampal volume is associated with memory but not monmemory cognitive performance in patients with mild cognitive impairme nt. J Mol Neurosci. 2003;20(3):241-8.
169 Hara H, Monsonego A, Yuasa K, Adachi K, Xiao X, Takeda S, Takahashi K, Weiner HL, Tabira T. Development of a safe oral Abeta vacci ne using recombinant adeno-associated virus vector for Alzhei mer's disease. J Alzheimers Dis. 2004 Oct;6(5):483-8. Hart DJ, Craig D, Compton SA, Critchl ow S, Kerrigan BM, McIlroy SP, Passmore AP. A retrospective study of the behavioural and psychological symptoms of mid and late phase Alzheimer's disease. Int J Geriatr Psychiat ry. 2003 Nov;18(11):103742. Hebert LE, Scherr PA, Bienias JL, Benne tt DA, Evans DA. Alzheimer disease in the US population: prevalence estimates us ing the 2000 census. Arch Neurol. 2003; 60(8):1119-22. Heslop HE, Rooney CM. Adoptiv e cellular immunotherapy for EBV lymphoproliferative disease. Immunol Rev. 1997 Jun;157:217-22. Hensley K, Butterfield DA, Hall N, Cole P, Subramaniam R, Mark R, Mattson MP, Markesbery WR, Harris ME, Aksenov M, et al. Reactive oxygen species as causal agents in the neurotoxicity of the Alzh eimer's disease-associated amyloid beta peptide. Ann N Y Acad Sci. 1996 Jun 15;786:120-34. Hickey WF, Kimura H. Perivascular microglial cells of the CNS are bone marrowderived and present antigen in vivo. Science. 1988 Jan 15;239(4837):290-2. Hock C, Konietzko U, Papassotiropoulos A, Wollmer A, Streffer J, von Rotz RC, Davey G, Moritz E, Nitsch RM. Generation of antibodies specific for beta-amyloid by vaccination of patients with Alzheime r disease. Nat Med. 2002 Nov;8(11):12705. Hock C, Konietzko U, Streffer JR, Tracy J, Signorell A, Muller-Tillmanns B, Lemke U, Henke K, Moritz E, Garcia E, Woll mer MA, Umbricht D, de Quervain DJ, Hofmann M, Maddalena A, Papassotiropoulos A, Nitsch RM. Antibodies against beta-amyloid slow cognitiv e decline in Alzheimer's disease. Neuron. 2003 May 22;38(4):547-54. Holcomb L, Gordon MN, McGowan E, Yu X, Benkovic S, Jantzen P, Wright K, Saad I, Mueller R, Morgan D, Sanders S, Zehr C, O'Campo K, Hardy J, Prada CM, Eckman C, Younkin S, Hsiao K, Duff K. Accelerated Alzheimer-type phenotype in transgenic mice carrying both mutant amyl oid precursor protein and presenilin 1 transgenes. Nat Med. 1998 Jan;4(1):97-100.
170 Holcomb LA, Gordon MN, Jantzen P, Hsiao K, Duff K, Morgan D. Behavioral changes in transgenic mice expressing both amyloid precursor protein and presenilin-1 mutations: lack of associati on with amyloid deposits. Behav Genet. 1999 May;29(3):177-85. Hsiao K, Chapman P, Nilsen S, Eckman C, Harigaya Y, Younkin S, Yang F, Cole G. Correlative memory deficits, Abeta elevation, and amyloid pl aques in transgenic mice. Science. 1996 Oct 4;274(5284):99-102. Hwang TJ, Masterman DL, Ortiz F, Fairba nks LA, Cummings JL Mild cognitive impairment is associated with characteri stic neuropsychiatric symptoms. Alzheimer Dis Assoc Disord. 2004 Jan-Mar;18(1):17-21. Irizarry MC, Soriano F, McNamara M, Pa ge KJ, Schenk D, Games D, Hyman BT. Abeta deposition is associated with neuropil changes, but not with overt neuronal loss in the human amyloid pr ecursor protein V717F (PDAPP) transgenic mouse. J Neurosci. 1997 Sep 15;17(18):7053-9. Janus C, Westaway D. Transgenic mouse models of Alzheimer's disease. Physiol Behav. 2001 Aug;73(5):873-86. Janus C, Pearson J, McLaurin J, Math ews PM, Jiang Y, Schmidt SD, Chishti MA, Horne P, Heslin D, French J, Mount HT Nixon RA, Mercken M, Bergeron C, Fraser PE, St George-Hyslop P, Westaway D. A beta peptide immunization reduces behavioural impairment and plaques in a model of Alzheimer's disease. Nature. 2000 Dec 21-28;408(6815):979-82. Kalmijn S, Launer LJ, Ott A, Witteman JC Hofman A, Breteler MM. Dietary fat intake and the risk of incident demen tia in the Rotterdam Study. Ann Neurol. 1997 Nov;42(5):776-82. Kawarabayashi T, Younkin LH, Saido TC, Shoji M, Ashe KH, Younkin SG. Agedependent changes in brain, CSF, and plasma amyloid (beta) protein in the Tg2576 transgenic mouse model of Alzheimer's di sease. J Neurosci. 2001 Jan 15;21(2):37281. Kidd, M. Alzheimers DiseaseAn Electr on Microscopical Study. Brain; a Journal Of Neurology Volume 87, June 1964, Pages 307-320. Kihara T, Shimohama S. Alzheimer's disease and acetylcholine receptors. Acta Neurobiol Exp (Wars). 2004;64(1):99-105. Kimberly WT, Wolfe MS. Identity and f unction of gamma-secretase. J Neurosci Res. 2003 Nov 1;74(3):353-60.
171 King DL, Arendash GW, Crawford F, Sterk T, Menendez J, Mullan MJ. Progressive and gender-dependent cognitive impairment in the APP(SW) transgenic mouse model for Alzheimer's disease. Behav Brain Res. 1999 Sep;103(2):145-62. King DL, Arendash GW. Behavioral characterization of th e Tg2576 transgenic model of Alzheimer's disease throug h 19 months. Physiol Behav. 2002 Apr 15;75(5):627-42. King DL, Arendash GW. Maintained synaptophysin immunoreactivity in Tg2576 transgenic mice during aging: correlations with cognitive impairment. Brain Res. 2002 Feb 1;926(1-2):58-68. Kipnis J, Mizrahi T, Yoles E, Ben-Nun A, Schwartz M. Myelin specific Th1 cells are necessary for post-traumatic protec tive autoimmunity. J Neuroimmunol. 2002 Sep;130(1-2):78-85. Klunk WE, Engler H, Nordberg A, Bacs kai BJ, Wang Y, Price JC, Bergstrom M, Hyman BT, Langstrom B, Mathis CA. Imaging the pathology of Alzheimer's disease: amyloid-imaging with positron emission tomography. Neuroimaging Clin N Am. 2003 Nov;13(4):781-9, ix. Kolb HJ, Schattenberg A, Goldman JM, Hertenstein B, Jacobsen N, Arcese W, Ljungman P, Ferrant A, Verdonck L, Niederwi eser D, et al. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. European Group for Blood and Marrow Transplantation Wo rking Party Chronic Leukemia. Blood. 1995 Sep 1;86(5):2041-50. Kollerr MF, Mohajeri HM, Huber M, Wollm er MA, Roth Zgraggen BV, Sandmeier E, Moritz E, Tracy J, Nitsch RM, Christen P. Active Immunization of Mice with an A -Hsp70 Vaccine. Neurodegenerative Dis. 2004;1:20-28. Kotilinek LA, Bacskai B, Westerman M, Kawarabayashi T, Younkin L, Hyman BT, Younkin S, Ashe KH. Reversible memory loss in a mouse transgenic model of Alzheimer's disease. J Neur osci. 2002 Aug 1;22(15):6331-5. Koudinov AR, Berezov TT. Alzheimer's am yloid-beta (A beta) is an essential synaptic protein, not neurot oxic junk. Acta Neurobiol Exp (Wars). 2004;64(1):71-9. Lambert MP, Viola KL, Chromy BA, Chang L, Morgan TE, Yu J, Venton DL, Krafft GA, Finch CE, Klein WL. Vacci nation with soluble Abeta oligomers generates toxicity-neutra lizing antibodies. J Neur ochem. 2001 Nov;79(3):595-605. Langbart C. Diagnosing and treating Alzheimer's disease: a practitioner's overview. J Am Acad Nurse Pract. 2002 Mar;14(3):103-9; quiz 110-2.
172 Lee HG, Casadesus G, Zhu X, Takeda A, Perry G, Smith MA. Challenging the amyloid cascade hypothesis: senile plaque s and amyloid-beta as protective adaptations to Alzheimer disease. Ann N Y Acad Sc i. 2004 Jun;1019:1-4. Lemere CA, Blusztajn JK, Yamaguchi H, Wisniewski T, Saido TC, Selkoe DJ. Sequence of deposition of heterogeneous amyloid beta-peptides and APO E in Down syndrome: implications for initial events in amyloid plaque formation. Neurobiol Dis. 1996 Feb;3(1):16-32. Lemere CA, Spooner ET, LaFrancois J, Male ster B, Mori C, Leverone JF, Matsuoka Y, Taylor JW, DeMattos RB, Holtzman DM, Clements JD, Selkoe DJ, Duff KE. Evidence for peripheral clearance of cereb ral Abeta protein following chronic, active Abeta immunization in PSAPP mice. Neurobiol Dis. 2003 Oct;14(1):10-8. Letenneur L, Larrieu S, Barberger-Gateau P. Alcohol and tobacco consumption as risk factors of dementia: a review of epidemiological studies. Biomed Pharmacother. 2004 Mar;58(2):95-9. Levine BL, Bernstein WB, Aronson NE, Schlienger K, Cotte J, Perfetto S, Humphries MJ, Ratto-Kim S, Birx DL, Ste ffens C, Landay A, Carroll RG, June CH. Adoptive transfer of costimulated CD4+ T cells induces expansion of peripheral T cells and decreased CCR5 expression in HIV infection. Nat Med. 2002 Jan;8(1):4753. Liberto CM, Albrecht PJ, Herx LM, Y ong VW, Levison SW. ( 2004) Pro-regenerative properties of cytokine-activated astrocytes. J Neurochem. 89:1092-1100. Li Q, Cao C, Chackerian B, Schiller J, Gordon M, Ugen KE, Morgan D. Overcoming antigen masking of anti-amyl oidbeta antibodies reveals breaking of B cell tolerance by virus-like particles in amyloidbeta immunized amyloid precursor protein transgenic mice. BMC Neurosci. 2004 Jun 08;5(1):21. Lokhorst HM, Schattenberg A, Cornelis sen JJ, Thomas LL, Verdonck LF. Donor leukocyte infusions are effec tive in relapsed multiple my eloma after allogeneic bone marrow transplantation. Blood. 1997 Nov 15;90(10):4206-11. Lyketsos CG, Sheppard JM, Steinberg M, Tschanz JA, Norton MC, Steffens DC, Breitner JC. Neuropsychiatri c disturbance in Alzheimer's disease clusters into three groups: the Cache County study. Int J Ge riatr Psychiatry. 2001 Nov;16(11):104353. Mallat Z, Gojova A, Brun V, Esposito B, Fournier N, Cottrez F, Tedgui A, Groux H. Induction of a regulatory T cell type 1 response reduces the development of atherosclerosis in apolipoprotein E-knockout mice. Circulation. 2003 Sep 9;108(10):1232-7. Epub 2003 Aug 11.
173 Marlatt M, Lee HG, Perry G, Smith MA, Zhu X. Sources and mechanisms of cytoplasmic oxidative damage in Alzheimer' s disease. Acta Neurobiol Exp (Wars). 2004;64(1):81-7. Matsuoka Y, Saito M, LaFrancois J, Saito M, Gaynor K, Olm V, Wang L, Casey E, Lu Y, Shiratori C, Lemere C, Duff K. Novel therapeutic appro ach for the treatment of Alzheimer's disease by peripheral admini stration of agents with an affinity to beta-amyloid. J Neuros ci. 2003 Jan 1;23(1):29-33. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer's disease: re port of the NINCDS-ADRDA Work Group under the auspices of Department of H ealth and Human Services Task Force on Alzheimer's Disease. Ne urology. 1984 Jul;34(7):939-44. McLaurin J, Cecal R, Kierstead ME, Ti an X, Phinney AL, Manea M, French JE, Lambermon MH, Darabie AA, Brown ME, Janus C, Chishti MA, Horne P, Westaway D, Fraser PE, Mount HT, Pr zybylski M, St George-Hyslop P. Therapeutically effective antibodies agains t amyloid-beta peptide target amyloidbeta residues 4-10 and inhibit cytotoxici ty and fibrillogenesis. Nat Med. 2002 Nov;8(11):1263-9. Mehlhorn G, Hollborn M, Schliebs R. Induction of cytoki nes in glial cells surrounding cortical beta-amyloid pla ques in transgenic Tg2576 mice with Alzheimer pathology. Int J Dev Ne urosci. 2000 Jul-Aug;18(4-5):423-31. Miller LJ, Chacko R. The role of choles terol and statins in Alzheimer's disease. Ann Pharmacother. 2004 Jan;38(1):91-8. Minoshima S. Imaging Alzheimer's diseas e: clinical applicat ions. Neuroimaging Clin N Am. 2003 Nov;13(4):769-80. Moalem G, Gdalyahu A, Shani Y, Otten U, Lazarovici P, Cohen IR, Schwartz M. Production of neurotrophins by activated T cells: implications for neuroprotective autoimmunity. J Autoimmun. 2000 Nov;15(3):331-45. Monsonego A, Maron R, Zota V, Selkoe DJ, Weiner HL. Immune hyporesponsiveness to amyloid beta-peptide in amyloid precursor protein transgenic mice: implications for the pathogenesis and treatment of Alzheimer's disease. Proc Natl Acad Sci U S A. 2001 Aug 28;98(18):10273-8. Epub 2001 Aug 21. Monsonego A, Imitola J, Zota V, Oida T, Weiner HL. Microglia-mediated nitric oxide cytotoxicity of T ce lls following amyloid beta-p eptide presentation to Th1 cells. J Immunol. 2003 Sep 1;171(5):2216-24.
174 Monsonego A, Zota V, Karni A, Kriege r JI, Bar-Or A, Bitan G, Budson AE, Sperling R, Selkoe DJ, Weiner HL. Incr eased T cell reactivity to amyloid beta protein in older humans and patients with Alzheimer disease. J Clin Invest. 2003 Aug;112(3):415-22. Moran MC, Walsh C, Lynch A, Coen RF, Coakley D, Lawlor BA. Syndromes of behavioral and psychological symptoms in mild Alzheimers diseasae. Int J Geriatr Psychiatry. 2004 Apr;19(4):359-64. Morgan D, Diamond DM, Gottschall PE, Ugen KE, Dickey C, Hardy J, Duff K, Jantzen P, DiCarlo G, Wilcock D, Connor K, Hatcher J, Hope C, Gordon M, Arendash GW. A beta peptide vaccination prevents memory loss in an animal model of Alzheimer's disease. Nature. 2000 Dec 2128;408(6815):982-5. Mullan M, Bennett C, Figueredo C, H ughes D, Mant R, Owen M, Warren A, McInnis M, Marshall A, Lantos P, et al. Clinical features of early onset, familial Alzheimer's disease linked to chromosome 14. Am J Med Genet. 1995 Feb 27;60(1):44-52. Nath A, Hall E, Tuzova M, Dobbs M, J ons M, Anderson C, Woodward J, Guo Z, Fu W, Kryscio R, Wekstein D, Smith C, Markesbery WR, Mattson MP. Autoantibodies to amyloid beta-peptide (Abe ta) are increased in Alzheimer's disease patients and Abeta antibodies can enhan ce Abeta neurotoxicity: implications for disease pathogenesis and vaccine development. Neuromolecular Med. 2003;3(1):29-39. Nicoll JA, Wilkinson D, Holmes C, Steart P, Markham H, Weller RO. Neuropathology of human Alzheimer disease after immunization with amyloid-beta peptide: a case report. Nat Med. 2003 Apr;9(4):448-52. Nilsson LN, Bales KR, DiCarlo G, Gordon MN, Morgan D, Paul SM, Potter H. Alpha-1-antichymotrypsin promotes beta-s heet amyloid plaque deposition in a transgenic mouse model of Alzheime r's disease. J Neurosci. 2001 Mar 1;21(5):1444-51. Nilsson LN, Arendash GW, Leighty RE, Costa DA, Low MA, Garcia MF, Cracciolo JR, Rojiani A, Wu X, Bales KR, Paul SM Potter H. Cognitive impairment in PDAPP mice depends on ApoE and ACT-catalyzed amyloid formation. Neurobiol Aging. 2004 Oct;25(9):1153-67. Oddo S, Billings L, Kesslak JP, Cribbs DH, LaFerla FM. Abeta immunotherapy leads to clearance of early, but not late, hyperphosphorylated tau aggregates via the proteasome. Neuron. 2004 Aug 5;43(3):321-32.
175 Orgogozo JM, Gilman S, Dartigues JF, La urent B, Puel M, Kirby LC, Jouanny P, Dubois B, Eisner L, Flitman S, Michel BF Boada M, Frank A, Hock C. Subacute meningoencephalitis in a subset of patie nts with AD after Abeta42 immunization. Neurology. 2003 Jul 8;61(1):46-54. Pennanen C, Kivipelto M, Tuomainen S, Hartikainen P, Hanninen T, Laakso MP, Hallikainen M, Vanhanen M, Nissinen A, Helkala EL, Vainio P, Vanninen R, Partanen K, Soininen H. Hippocampus a nd entorhinal cortex in mild cognitive impairment and early AD. Neurobiol Aging. 2004 Mar;25(3):303-10. Pfeifer M, Boncristiano S, Bondolfi L, St alder A, Deller T, Staufenbiel M, Mathews PM, Jucker M. Cerebral hemorrhage af ter passive anti-Abe ta immunotherapy. Science. 2002 Nov 15;298(5597):1379. Pompl PN, Mullan MJ, Bjugstad K, Arendash GW. Adaptation of the circular platform spatial memory task for mice: us e in detecting cognitive impairment in the APP(SW) transgenic mouse model for Alzh eimer's disease. J Neurosci Methods. 1999 Feb 1;87(1):87-95. Pratico D, Uryu K, Leight S, Trojanoswk i JQ, Lee VM. Increased lipid peroxidation precedes amyloid plaque formation in an an imal model of Alzheimer amyloidosis. J Neurosci. 2001 Jun 15;21(12):4183-7. Raina AK, Zhu X, Smith MA. Alzheime r's disease and the cell cycle. Acta Neurobiol Exp (Wars). 2004;64(1):107-12. Rogers J, Webster S, Lue LF, Brachova L, Civin WH, Emmer ling M, Shivers B, Walker D, McGeer P. Inflammation a nd Alzheimer's disease pathogenesis. Neurobiol Aging. 1996 Sep-Oct;17(5):681-6. Scarpini E, Scheltens P, Feldman H. Treatment of Alzheimer's disease: current status and new perspectives. La ncet Neurol. 2003 Sep;2(9):539-47. 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, Seubert P. Immunization with amyloid-beta attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature. 1999 Jul 8;400(6740):173-7. Schenk D, Hagen M, Seubert P. Current progress in beta-amyloid immunotherapy. Curr Opin Immunol. 2004 Oct;16(5):599-606. Schwab C, Hosokawa M, McGeer PL. Transgenic mice overexpressing amyloid beta protein are an incomplete model of Alzheimer disease. Exp Neurol. 2004 Jul;188(1):52-64.
176 Selkoe DJ. Alzheimer's disease: genes, proteins, and therapy. Physiol Rev. 2001 Apr;81(2):741-66. Selkoe DJ. Alzheimer's disease is a synaptic failure. Science. 2002 Oct 25;298(5594):789-91. Schwartz M, Shaked I, Fisher J, Mizrah i T, Schori H. Protective autoimmunity against the enemy within: fighting glutamate toxicit y. Trends Neurosci. 2003 Jun;26(6):297-302. Sigurdsson EM, Wisniewski T, Frangione B. A safer vaccine for Alzheimer's disease? Neurobiol Ag ing. 2002 Nov-Dec;23(6):1001-8. Sigurdsson EM, Scholtzova H, Mehta PD, Frangione B, Wisniewski T. Immunization with a nontoxic/nonfibrill ar amyloid-beta homologous peptide reduces Alzheimer's disease-associated path ology in transgenic mice. Am J Pathol. 2001 Aug;159(2):439-47. Sigurdsson EM, Knudsen E, Asuni A, Fitze r-Attas C, Sage D, Quartermain D, Goni F, Frangione B, Wisniewski T. An at tenuated immune response is sufficient to enhance cognition in an Alzheimer's dis ease mouse model immunized with amyloidbeta derivatives. J Neurosci. 2004 Jul 14;24(28):6277-82. Slavin S, Naparstek E, Nagler A, Kapelus hnik Y, Ackerstein A, Or R. Allogeneic cell therapy: the treatment of choice for all hematologic malignancies relapsing post BMT. Blood. 1996 May 1;87(9):4011-3. Smith MA, Hirai K, Hsiao K, Pappolla MA, Harris PL, Siedlak SL, Tabaton M, Perry G. Amyloid-beta depos ition in Alzheimer transgenic mice is associated with oxidative stress. J Neur ochem. 1998 May;70(5):2212-5. Smith DH, Nakamura M, McIntosh TK Wang J, Rodriguez A, Chen XH, Raghupathi R, Saatman KE, Clemens J, Schmidt ML, Lee VM, Trojanowski JQ. Brain trauma induces massive hippocampal ne uron death linked to a surge in betaamyloid levels in mice overexpressing mu tant amyloid precursor protein. Am J Pathol. 1998 Sep;153(3):1005-10. Sobow T, Kloszewska I. Donepezil plus vitamin E as a treatment in Alzheimer disease. Alzheimer Dis Asso c Disord. 2003 Oct-Dec;17(4):244. Sobow T, Flirski M, Liberski PP. Amyloid-beta and tau prot eins as biochemical markers of Alzheimer's disease. Ac ta Neurobiol Exp (W ars). 2004;64(1):53-70. Solomon B, Koppel R, Hanan E, Katzav T. Monoclonal antibodies inhibit in vitro fibrillar aggregation of the Alzheimer beta-amyloid peptide. Proc Natl Acad Sci U S A. 1996 Jan 9;93(1):452-5.
177 Solomon B, Koppel R, Frankel D, HananAharon E. Disaggregation of Alzheimer beta-amyloid by site-directed mAb. Proc Natl Acad Sci U S A. 1997 Apr 15;94(8):4109-12. Sonkusare SK, Kaul CL, Ramarao P. Dementia of Alzheimers disease and other neurodegenerative disorders-memantin e, a new hope. Pharmacol Res. 2005 Jan;51(1):1-17. Streit WJ. Microglia and Alzheimer's disease pathogenesis. J Neurosci Res. 2004 Jul 1;77(1):1-8. Streit WJ, Mrak RE, Griffi n WS. Microglia and neuroi nflammation: a pathological perspective. J Neuroinflammation. 2004 Jul 30;1(1):14. Su GC, Arendash GW, Kalaria RN, Bj ugstad KB, Mullan M. Intravascular infusions of soluble beta-amyloid compro mise the blood-brain barrier, activate CNS glial cells and induce peri pheral hemorrhage. Brain Res. 1999 Feb 6;818(1):105-17. Suzuki N, Cheung TT, Cai XD, Odaka A, Otvos L Jr, Eckman C, Golde TE, Younkin SG. An increased percentage of long amyloid beta protein secreted by familial amyloid beta protein precursor (beta APP717) mutants. Science. 1994 May 27;264(5163):1336-40. Szer J, Grigg AP, Phillips GL, Sheridan WP. Donor leucocyte infusions after chemotherapy for patients relapsing with acute leukaemia following allogeneic BMT. Bone Marrow Transplant. 1993 Feb;11(2):109-11. Takeuchi A, Irizarry MC, Duff K, Sai do TC, Hsiao Ashe K, Hasegawa M, Mann DM, Hyman BT, Iwatsubo T. Age-related amyloid beta deposition in transgenic mice overexpressing both Alzheimer mutant pr esenilin 1 and amyloid beta precursor protein Swedish mutant is not associated with global neuronal loss. Am J Pathol. 2000 Jul;157(1):331-9. Tariot PN, Farlow MR, Grossberg GT, Graham SM, McDonald S, Gergel I; Memantine Study Group. Memantine treatment in patients with moderate to severe Alzheimer disease already receiving done pezil: a randomized c ontrolled trial. JAMA. 2004 Jan 21;291(3):317-24. Tarkowski E, Blennow K, Wallin A, Tar kowski A. Intracerebral production of tumor necrosis factor-alpha, a local neuroprotective agent, in Alzheimer disease and vascular dementia. J Clin Immunol. 1999 Jul;19(4):223-30. Teaktong T, Graham AJ, Johnson M, Cour t JA, Perry EK. Selective changes in nicotinic acetylcholine receptor subtyp es related to tobacco smoking: an immunohistochemical study. Neuropathol Appl Neurobiol. 2 004 Jun;30(3):243-54.
178 Terry RD, Masliah E, Salmon DP, Butters N, DeTeresa R, Hill R, Hansen LA, Katzman R. Physical basis of cognitive al terations in Alzheimer's disease: synapse loss is the major correlate of cognitive impairment. Ann Neurol. 1991 Oct;30(4):572-80. Town T, Vendrame M, Patel A, Poetter D, DelleDonne A, Mori T, Smeed R, Crawford F, Klein T, Tan J, Mullan M. Reduced Th1 and enhanced Th2 immunity after immunization with Alzheimer's be ta-amyloid(1-42). J Neuroimmunol. 2002 Nov;132(1-2):49-59. Trinchese F, Liu S, Battaglia F, Walter S, Mathews PM, Arancio O. Progressive age-related development of Alzheimer-l ike pathology in APP/PS1 mice. Ann Neurol. 2004 Jun;55(6):801-14. Uchihara T, Duyckaerts C, He Y, Kobayashi K, Seilhean D, Amouyel P, Hauw JJ. ApoE immunoreactivity and microglial cells in Alzheimer's disease brain. Neurosci Lett. 1995 Jul 28;195(1):5-8. Vallejo AN, Davila E, Weyand CM, Gor onzy JJ. Biology of T lymphocytes. Rheum Dis Clin North Am. 2004 Feb;30(1):135-57. Vonderheide RH, June CH. A translat ional bridge to cancer immunotherapy: exploiting costimulation and target antigens for active and passive T cell immunotherapy. Immunol Res. 2003;27(2-3):341-56. Walter EA, Greenberg PD, Gilbert MJ, Finch RJ, Watanabe KS, Thomas ED, Riddell SR. Reconstitution of cellular immunity against cytomegalovirus in recipients of allogeneic bone marrow by tr ansfer of T-cell clones from the donor. N Engl J Med. 1995 Oct 19;333(16):1038-44. Wang PN, Yang CL, Lin KN, Chen WT, Chwang LC, Liu HC. Weight loss, nutritional status and physica l activity in patients with Alzheimer's disease. A controlled study. J Neurol. 2004 Mar;251(3):314-20. Webster SD, Yang AJ, Margol L, GarzonRodriguez W, Glabe CG, Tenner AJ. Complement component C1q modulates th e phagocytosis of Abeta by microglia. Exp Neurol. 2000 Jan;161(1):127-38. Weiner HL, Lemere CA, Maron R, Spooner ET, Grenfell TJ, Mori C, Issazadeh S, Hancock WW, Selkoe DJ. Nasal administration of amyloid-beta peptide decreases cerebral amyloid burden in a mouse model of Alzheimer's disease. Ann Neurol. 2000 Oct;48(4):567-79. Wellington CL. Cholesterol at the cro ssroads: Alzheimer's disease and lipid metabolism. Clin Genet. 2004 Jul;66(1):1-16.
179 Welsh KA, Butters N, Hughes JP, Mohs RC, Heyman A. Detection and staging of dementia in Alzheimer's disease. Use of the neuropsychological measures developed for the Consortium to Establish a Registry for Alzheimer's Disease. Arch Neurol. 1992 May;49(5):448-52. Westerman MA, Cooper-Blacketer D, Mari ash A, Kotilinek L, Kawarabayashi T, Younkin LH, Carlson GA, Younkin SG, Ashe KH. The relationship between Abeta and memory in the Tg2576 mouse model of Alzheimer's disease. J Neurosci. 2002 Mar 1;22(5):1858-67. Wilcock DM, Gordon MN, Ugen KE, Gottschall PE, DiCarlo G, Dickey C, Boyett KW, Jantzen PT, Connor KE, Melachrino J, Hardy J, Morgan D. Number of Abeta inoculations in APP+PS1 transgenic mi ce influences antibody titers, microglial activation, and congophilic plaque levels DNA Cell Biol. 2001 Nov;20(11):731-6. Wilcock DM, Munireddy SK, Rosenthal A, Ugen KE, Gordon MN, Morgan D. Microglial activation facilitates Abeta pl aque removal following intracranial antiAbeta antibody administration. Neurobiol Dis. 2004 Feb;15(1):11-20. Winkler DT, Bondolfi L, Herzig MC, Jann L, Calhoun ME, Wiederhold KH, Tolnay M, Staufenbiel M, Jucker M. Spontaneous hemorrhagic stroke in a mouse model of cerebral amyloid angiopathy. J Neurosci. 2001 Mar 1;21(5):1619-27. Wong TP, Debeir T, Duff K, Cuello AC. Reorganization of cholinergic terminals in the cerebral cortex and hippocampus in transgenic mice carrying mutated presenilin1 and amyloid precursor protein transgen es. J Neurosci. 1999 Apr 1;19(7):2706-16. Yao J, Petanceska SS, Montine TJ, Holtzman DM, Schmidt SD, Parker CA, Callahan MJ, Lipinski WJ, Bisgaier CL, Turner BA, Nixon RA, Martins RN, Ouimet C, Smith JD, Davies P, Laska E, Ehrlich ME, Walker LC, Mathews PM, Gandy S. Aging, gender and APOE isotype modulate metabolism of Alzheimer's Abeta peptides and F-isoprostanes in the ab sence of detectable amyloid deposits. J Neurochem. 2004 Aug;90(4):1011-8. Zekanowski C, Religa D, Graff C, Filipek S, Kuznicki J. Genetic aspects of Alzheimer's disease. Acta Ne urobiol Exp (War s). 2004;64(1):19-31. Zhang F, Eckman C, Younkin S, Hsiao KK, Iadecola C. Increased susceptibility to ischemic brain damage in transgenic mice overexpressing the amyloid precursor protein. J Neurosci. 1997 Oct 15;17(20):7655-61. Zhu X, Raina AK, Perry G, Smith MA. Alzh eimer's disease: the two-hit hypothesis. Lancet Neurol. 2004 Apr;3(4):219-26. Zlokovic BV. Clearing amyloid through the blood-brain barrier. Neurochem. 2004 May;89(4):807-11.