Alzheimer’s disease is a debilitating, progressive neurodegenerative disorder in the elderly, characterized by severe loss of memory and higher cognitive functions. In the hundred years since its discovery, Alzheimer’s disease (AD) has traversed from the status of a ‘rare neurological oddity’ to one of the greatest challenges faced by healthcare and medicine in this millennium. A reported 44 million people currently suffer from AD but only 1 in 4 people have been diagnosed. Although AD has been an area of intense research for almost 50 years now, most studies have focused on the end stage disease. Years of study on the pathological cause underlying AD; have conclusively shown that the accumulation of the sticky peptide, Aβ, is one of the major triggers of AD pathogenesis. However, after the initial Aβ trigger, multiple processes contribute to disease progression, so that by the time a patient is diagnosed on the basis of overt behavioral phenotypes, it is difficult to understand and differentiate between the causative mechanisms and the consequential effects of the disease. It is, perhaps, because of this, that we are still struggling to find therapies for AD which will stop or at the very least slow the course of the disease. In the 2015 report on AD, issued by the Alzheimer’s association, much emphasis has been placed on the early diagnosis of AD and the revision of the diagnostic criteria for AD. According to the new guidelines proposed in 2011, AD has been divided into three stages where the first stage occurs before the appearance of overt behavioral symptoms such as memory loss, whereas by the 1984 guidelines, cognitive disabilities must have already occurred for diagnoses of AD. This proposed preclinical stage of AD has been defined, reflecting the current belief that AD pathogenesis begins almost 20 years before the occurrence of behavioral dysfunction. However, no diagnostic criteria are currently available to establish this stage. Hence, there is a need to understand the early pathogenic mechanisms of AD, which will yield early therapeutic targets as well as early diagnostic markers of AD. One of the earliest documented events in AD pathogenesis is synaptic dysfunction, which is later manifested as loss of dendritic spines. Deficits in long term potentiation (LTP) has been demonstrated in Aβ exposed hippocampal slices as well as in mouse models of AD, much before the appearance of pathological hallmarks such as plaques and tangles as well as overt behavioral phenotypes. While these and other studies indicate clearly that elevated levels of soluble Aβ peptide leads to impairment of synaptic function, the underlying molecular mechanisms are yet to be elucidated. One of the purported mediators of Aβ induced dysfunction is oxidative stress. The Aβ peptide, especially the Aβ42, is a self aggregating peptide with a propensity to form peptidyl radicals. Interaction of the peptidyl radicals with biomolecules leads to the generation of more free radical species via cascading chain reactions. Additionally, Aβ peptide has also been demonstrated to have synaptotoxic effects via its effect on NMDA receptors and calcium influx leading to deregulated reactive oxygen species (ROS) production as well as excitotoxicity. Hence, with a view to understanding Aβ mediated early synaptic dysfunction in AD, we studied early signaling changes in the synaptosomes derived from the cortex of APP/PS1 mice model of AD at various ages. The APP/PS1 model contains a mouse/human chimeric APP gene bearing the KM670/671NL Swedish mutation and the human PS1 gene with an exon 9 deletion. These mice exhibit behavioral deficits from 7 months of age while plaque deposition and gliosis become apparent by 9 months of age. We chose to study both pre-symptomatic ages (1 and 3 months old) as well as post symptomatic (9 months old) mice. Post nuclear supernatant (PNS) as well as synaptosomes were isolated from the cortex of APP/PS1 and age matched control mice. We assayed the levels of reactive oxygen species (ROS) in the PNS and the synaptosomes of post symptomatic 9 months old APP/PS1 mice and age matched controls. In contrast to reports of enhanced oxidative stress markers in the brains of AD patients, we did not find any increase in the levels of ROS in the PNS of post symptomatic APP/PS1 mice compared to age matched controls. However, synaptosomes from the cortex of these animals exhibited a significant increase in ROS levels in APP/PS1 mice compared to controls. We further found that there was significant increase in the ROS levels in synaptosomes, but not PNS, of very young asymptomatic 1 and 3 months old APP/PS1 mice. This is a first demonstration of synapse specific increase in oxidative stress in AD mice, as young as 1 month of age, indicating that disease specific mechanisms operate at the synapse much before the appearance of any overt cellular or behavioral symptoms. The increase in synaptic ROS levels correlated with a small but significant increase in the levels of Aβ42 in the brains of APP/PS1 mice compared to controls. We also found a concurrent change in the redox status of the cytoskeletal protein, actin, at the synapse. As early as 1 month of age, there was a significant decrease in the protein level of reduced actin indicating that there is an increase in the level of oxidized actin at the synapse. This loss of reduced actin was specific to the fibrillar pool of actin while no significant change was observed in the redox status of the monomeric globular pool of actin. Oxidation of actin has been demonstrated to lead to its depolymerization. Concurrently, we found a significant loss of fibrillar actin in the synaptosomes of APP/PS1 mice. Actin is the major cytoskeletal protein at the synapse. Changes in the globular to fibrillar actin ratio at the synapse at early pre-symptomatic ages in APP/PS1 mice will likely lead to structural and consequent functional changes at the synapse. This could potentially be one of the triggers of synaptic dysfunction in AD. Furthermore, changes in the Akt-mTOR signaling pathway was also observed in the synaptosomes of 1 month old APP/PS1 mice, which is sustained at 9 months. There was a significant loss of the mTOR-pS6K-4EBP1 axis in the synaptosomes, but not PNS, of APP/PS1 mice. We found that loss of Akt signaling, as evinced by loss of Akt phosphorylation, Akt kinase activity as well as loss of phosphorylation of downstream effector GSK3β, potentially underlies the loss of mTOR signaling. Further, the loss of Akt signaling is mediated by synapse specific redox modification of Akt and consequent interaction with the protein phosphatase PP2a. Loss of the Akt-mTOR signaling at the synapse is indicative of deficits in local protein translation. Loss of this essential synaptic function, which plays critical roles in synapse maintenance as well as synaptic plasticity during learning and memory, at an early age, will have long ranging impact on synaptic function such as long term potentiation (LTP) in APP/PS1 mice. Our study is the first demonstration of oxidative stress and consequent signaling changes which occur specifically at the synapse of very young 1 month old APP/PS1 mice. These changes occur much before the appearance of overt phenotype such as plaque deposition and behavioral dysfunction but sustain till the appearance of classical pathological hallmarks. Hence, the study demonstrates that disease progression starts much before previously thought and provides us a critical time window during which therapeutic strategies designed to delay or stop these changes might change the course of AD.

Introduction: Idiopathic Parkinson's disease (PD) is a chronic and progressive neurodegenerative disorder that is characterised by motor symptoms. However, there is increasing awareness that a range of neuropsychiatric and cognitive problems also accompanys PD. The objective of this thesis was to examine the profile of neuropsychiatric and cognitive problems for patients with PD without dementia. Parkinson's disease patients who could be identified at the time of this study were invited to participate. Each patient was individually matched to a healthy control in terms of age, premorbid intelligence, and years of education. Results: Neuropsychiatric symptoms were common for this patient group, over 40% self reported symptoms consistent with depression, 40% with physical fatigue, 38% with mental fatigue, 38% with apathy and 32% with sleep problems. More than 77% of patients with PD reported symptoms associated with at least one problem and over 46% with 3 or more problems. Increased symptoms consistent with depression and anxiety and the presence of hallucinations also predicted poorer quality of life after controlling for motor symptoms. However, the of level agreement between patient report and that of a person who know them well was low: 40.9% for apathy, 28% for hallucinations, 39% for depression, 25% for sleep problems and only 7.7% agreement for the presence of anxiety. To obtain an accurate profile of cognitive impairments patients were assessed on measures of higher order language ability and a broad range of commonly used cognitive tests. Overall, PD patients were impaired on aspects of higher-order language. However, results indicated that these deficits were not a primary effect of PD, but could be explained in terms of deficits in speed of information processing associated with the disease. Compared to healthy controls, PD patients also showed deficits on measures of executive function, working memory, problem solving, and visuospatial skills. However, they were unimpaired on measures of planning, attention and memory/learning. Deficits in problem solving were only evident for tasks with a high visuospatial content and were no longer significant when visuospatial skills were controlled for. Further investigation indicated that planning in PD patients was not impaired in general and was dependent on the sensitivity of tests used. To further examine cognitive deficits, patients were divided into groups according to their cognitive performance. Three sub-groups of patients were identified that formed a continuum of cognitive impairment from none/mild to severe. Compared to controls, one subgroup showed no or minimal impairment (PD-NCI), a second group showed a more variable pattern of severe and mild impairments (PD-UCI), and a third group had evidence of severe impairment across most of the cognitive domains tested. This latter group was labelled PD-Mild Cognitive Impairment (PD-MCI). The PD-UCI and PD-MCI groups were also significantly different from their controls with respect to their ability to carry out functional activities of everyday living. The PD-MCI group had evidence of global cognitive decline, possibly reflecting a stage of pre-clinical dementia. The severity of cognitive deficits was not associated with other clinical and demographic characteristics such as motor impairments, age or disease duration. These results were confirmed when patients were retested one year later. Conclusions: Comorbid neuropsychiatric and cognitive problems are common for patients with PD prior to any overt signs of dementia. However, PD patients are heterogeneous with regard to their presentation and different subgroups of patients are identifiable based on cognitive performance. This information has both theoretical and clinical relevance.

Thesis (MScMedSC)--Stellenbosch University, 2012. / Bibliography / ENGLISH ABSTRACT: Parkinson’s disease (PD) is a neurodegenerative movement disorder characterized by the loss of dopaminergic neurons in the substantia nigra of the midbrain. Although the aetiology of PD is still not fully understood, it is thought to involve a combination of environmental (such as exposure to pesticides and neurotoxins) and genetic factors. A number of PD-causing genes have been found including SNCA, LRRK2, EIF4G1 and VPS35 (for autosomal dominant forms of PD) and parkin, PINK1, DJ-1 and ATP13A2 (for autosomal recessive forms of PD – arPD). Mutations in the parkin gene are the predominant cause of arPD. Parkin plays a role in the ubiquitin-proteasomal system which degrades damaged and unwanted proteins in the cell and it is also thought to be involved in maintaining healthy mitochondria. Numerous studies have implicated mitochondrial function in the pathogenesis of PD. Therefore the aim of the present study was to investigate the role of mitochondrial dysfunction in PD patients with parkin-null mutations. Four South African PD patients, each harbouring two parkin-null mutations, were recruited for this study. A muscle biopsy was performed for analysis of mitochondrial morphology using histology and transmission electron microscopy (TEM). Skin biopsies were taken, from which fibroblasts were cultured. These fibroblasts were used in i) mitochondrial morphological assessments using TEM, ii) mitochondrial network analysis, iii) functional studies via ROS measurement and iv) analysis of the proteome using a LTQ Orbitrap Velos mass spectrometer. In addition, RNA was isolated from peripheral blood samples for gene expression studies using the RT² Profiler PCR Array (SABiosciences, USA) and the RT² PCR Primer Assay (SABiosciences, USA). Heterozygous family members (carriers) and wild-type controls were also recruited for this study. Results from the histological and TEM analysis from the muscle biopsy observed subtle mitochondrial changes including the presence of type II fibres, atrophic fibres, the presence of lipids, and wrinkling of the sarcolemmal membrane. Enlarged mitochondria were also observed in one patient. TEM analysis on the patient’s fibroblasts observed an increase in the number of electron dense vacuoles, speculated to be autolysosomes. The mitochondrial network in two of the patients’ fibroblasts showed fragmented and dot-like networks which are indicative of damaged mitochondria. An increase in mitochondrial ROS levels was observed in three of the four patients. Expression studies found down-regulation of 14 genes from four of the five mitochondrial complexes and a total of 688 proteins were found only in the control and not in the patient fibroblasts. Some of these proteins are known to be part of the ‘mitochondrial dysfunction’ pathway. Taken together, these results indicate that the absence of parkin results in a number of mitochondrial alterations. Based on these findings, a model of PD was proposed: It is speculated that when parkin is absent, electron transport chain complex genes are down-regulated. This results in impaired oxidative phosphorylation, causing an increase in the production of mitochondrial ROS and subsequent oxidative stress. Mitochondria are then damaged; resulting in the fragmentation of the mitochondrial network. The impaired mitochondria are thus tagged for degradation, causing the recruitment of autolysosomes which engulf the mitochondria via mitophagy. Ultimately, as the compensatory mechanisms fail, this triggers the consequential cascade of cellular apoptotic events. This study has elucidated the effect of parkin on the mitochondria, and can act as a ‘stepping stone’ towards future development of therapeutic strategies and/or biochemical markers that will benefit not only patients with PD but also other neurodegenerative disorders. / AFRIKAANSE OPSOMMING: Parkinson se siekte (PS) is ‘n neurodegeneratiewe bewegings-afwyking gedefineer deur die verlies van dopaminergiese neurone in die substantia nigra van die midde brein. Alhoewel die spesifieke oorsprong van die afwyking nog nie ten volle begryp is nie, word bydraes van beide omgewings faktore (bv. blootstelling aan plaagdoders en neurotoksienes) asook genetiese faktore gespekuleer. Vanuit ‘n genetiese aspek is ‘n aantal gene al geassosieer met PS. Hierdie gene sluit in SNCA, LRRK2, EIF4G1 en VPS35 (vir outosomale dominante vorms van PS) en parkin, PINK1, DJ-1, en ATP13A2 (vir outosomale resessiewe vorms van PS - orPS). Mutasies in die parkin geen is aangedui as die hoof oorsaak van orPS. Parkin speel ‘n rol in die ubiquitine-proteasomale sisteem wat beskadige en ongewensde proteïne binne in die sel verwyder en is verdink om by te dra tot die instandhouding van gesonde mitokondria. Mitokondriese wanfunksionering is ook deur talle studies gewys as ‘n bydraende faktor in die patologie van PS. Die doel van die studie is om ondersoek in te stel tot die spesifieke rol wat mitokondriese wanfunsionering speel in PS pasiënte met parkin-nul mutasies. Vier Suid-Afrikaanse PS-pasiënte, elk met twee parkin-nul mutasies, is gebruik vir die studie. Deur middel van spierbiopsies is monsters verkry vir mitokondriese morfologiese analises met behulp van histologiese en elektron-oordrag mikroskopie tegnieke (TEM). Vel biopsies is ook geneem en fibroblaste is gekweek vir die gebruik in: i) mitokondriese morfologiese assesering; ii) mitokondriese netwerk analiese; iii) funksionele studies waar vlakke van reaktiewe suurstof spesies (ROS) gemeet is; iv) proteoom analiese met behup van ‘n LTQ Orbitrap Velos massa spektrometer. RNA is ook geisoleer vanaf perifere bloedmonsters vir die gebruik in geen-uitdrukkings studies met behulp van ‘n RT² Profiler PCR Array en ‘n RT² Primer Assay. Selle vanaf famielie lede wat heterosigotiese draers is van die mutasie, asook normale (geen parkin mutasie) selle is gebruik as kontroles in die studie. TEM resultate vanaf die spier monsters het subtiele mitokondriese veranderinge getoon. Hierdie sluit in die teenwoordigheid van tipe II vesels, atrofiese vesels, teenwoordigheid van lipiedes, assook waarnemings van rimpeling van die sarcolemmal membraan. Vergrote mitokondrias is ook in een van die pasiënte opgelet. TEM resultate vanaf die fibroblaste het toename in die aantal elektron-digte vakuole vertoon, moontlik geidentifiseer as autolisosome. Gefragmenteerde en onderbreekte mitokondria netwerke is gelet tydens netwerk analiese van die fibroblaste, ‘n indikasie van beskadigde mitokondria. ‘n Toename in mitokondriese ROS vlakke is gevind in drie van die vier pasiënte. Af-regulering van 14 gene, geassosieerd met vier uit die vyf mitokondria komplekse, is verneem tydens die geen-uitdrukkings studie. Saam met dit is ‘n totaal van 688 proteïene geidentifiseer wat slegs teenwoordig is in die kontrole monsters en nie in die pasiënt monsters nie. Hierdie proteïene is almal uitgedruk en betrokke in die mitokondriese wanfunsionerings-weë. Hierdie resultate dui dat die afwesigheid van parkin mitokondriese afwykings tot gevolg het wat kan lei tot die afsterwing van selle. Dit dra ook by tot die vorming van ‘n beter-verstaande siekte-model vir PS: Mutasies in parkin (wat lei tot die afwesigheid van parkin) kan dus moontlik lei tot die af-regulasie van gene geassosieerd met die elektron-vervoer ketting komplekse in die mitokondria. Dit lei tot gebrekkige oksidatiewe fosforilering en veroorsaak ‘n toename in die vorming van ROS, wat dan ‘n toename in oksidatiewe stres binne in die sel tot gevolg het. Uiteindelik lei dit dus tot die beskadiging van die mitokondria wat gepaard gaan met fragmentering van die mitokondriese netwerk. Beskadigde mitokondrias word geetiketeer vir afbraking. Hierdie etiketering aktiveer omringende autophagosome wat die beskadigde mitokondrias dan verwyder deur middel van ‘n verswelgende proses genaamd mitophagy. Dit veroorsaak die aktivering van ‘n aantal gekorreleerde sellulêre prosesse wat lei tot apoptose (afsterwing van die sel). Hierdie studie dra by tot die verklaring van die spesifieke effek wat parkin mutasies het op die funksionering van die mitokondria. Resultate hier lê ook die grondslag vir toekomstige studies met die doel tot die ontwikkeling van terapeutiese strategeë en biochemiese merkers wat kan bydrae tot die genesing van beide pasiënte met PS, asook pasiënte met ander neurodegeneratiewe afwykings.

A Neurodegenerative Disease (ND) is progressive damage to brain neurons, which the human body cannot repair or replace. The well-known examples of such conditions are Dementia and Alzheimer’s Disease (AD), which affect millions of lives each year. Although conducting numerous researches, there are no effective treatments for the mentioned diseases today. However, early diagnosis is crucial in disease management. Diagnosing NDs is challenging for neurologists and requires years of training and experience. So, there has been a trend to harness the power of deep learning, including state-of-the-art Convolutional Neural Network (CNN), to assist doctors in diagnosing such conditions using brain scans. The CNN models lead to promising results comparable to experienced neurologists in their diagnosis. But, the advent of transformers in the Natural Language Processing (NLP) domain and their outstanding performance persuaded Computer Vision (CV) researchers to adapt them to solve various CV tasks in multiple areas, including the medical field. This research aims to develop Vision Transformer (ViT) models using Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset to classify NDs. More specifically, the models can classify three categories (Cognitively Normal (CN), Mild Cognitive Impairment (MCI), Alzheimer’s Disease (AD)) using brain Fluorodeoxyglucose (18F-FDG) Positron Emission Tomography (PET) scans. Also, we take advantage of Automated Anatomical Labeling (AAL) brain atlas and attention maps to develop explainable models. We propose three ViTs, the best of which obtains an accuracy of 82% on the test dataset with the help of transfer learning. Also, we encode the AAL brain atlas information into the best performing ViT, so the model outputs the predicted label, the most critical region in its prediction, and overlaid attention map on the input scan with the crucial areas highlighted. Furthermore, we develop two CNN models with 2D and 3D convolutional kernels as baselines to classify NDs, which achieve accuracy of 77% and 73%, respectively, on the test dataset. We also conduct a study to find out the importance of brain regions and their combinations in classifying NDs using ViTs and the AAL brain atlas. / <p>This thesis was awarded a prize of 50,000 SEK by Getinge Sterilization for projects within Health Innovation.</p>