Gene transfer to the brain is a promising therapeutic strategy for a variety of neurodegenerative disorders including Parkinson‟s disease (PD). PD is the second most common neurodegenerative disease. Although many drugs have been developed and introduced into the market to provide symptomatic treatment, there is still no cure for PD. Glial cell line-derived neurotrophic factor (GDNF) is a potent survival factor for injured nigrostriatal dopamine neurons and is currently being evaluated as a potential treatment for PD. Gene therapy allows localized, long-term and stable transgene expression after a single intervention to obtain a therapeutic effect. Regulatable promoters for transgene expression furthermore allow optimizing GDNF concentration to avoid undesirable biological activity and clinical side effects. In the first part of the study, an autoregulatory tetracycline-inducible recombinant adeno-associated viral vector (rAAV-pTetbidiON) utilizing the rtTAM2 reverse tetracycline transactivator (rAAV-rtTAM2) was used to conditionally express the human GDNF cDNA. Eight weeks after a single intrastriatal injection of the rAAV-rtTAM2-GDNF vector encapsidated into AAV serotype 1 capsids (rAAV2/1), the GDNF protein level was respectively 15 fold higherand undistinguishable from the endogenous level in doxycycline(Dox) treated and untreated animals. However, a residual GDNF expression in the uninduced animals was evidenced by a sensitive immunohistochemical staining. As compared to rAAV2/1-rtTAM2-GDNF, the rAAV2/1-rtTAM2-WPRE-GDNF vector harboring a woodchuck hepatitis post-transcriptional regulatory element, which increases and stabilizes the transgene transcript, expressed a similar concentration of GDNF in the induced state but a basal level ~2.5-fold higher than the endogenous striatal level. However, the distribution of GDNF in the striatum in induced state was more widespread using the rAAV2/1-rtTAM2-WPRE-GDNF vector as compared to rAAV2/1-rtTAM2- GDNF. As a proof for biological activity, for both vectors, downregulation of tyrosine hydroxylase (TH) was evidenced in dopaminergic terminals of Dox-treated but not untreated animals. In the second part of my study, functional (behavioural) and neurochemical changes mediated by delayed intrastriatal GDNF gene delivery in the partial Parkinson‟s disease rat model were investigated. The rAAV2/1-rtTAM2-WPRE-GDNF vector (3.5 108 viral genomes) was administered unilaterally in the rat striatum 5 weeks after intrastriatal injection of 6-hydroxydopamine (6-OHDA) which produces a partial and progressive lesion of the nigro-striatal dopaminergic pathway. Rats were treated with Dox or untreated from the day of vector injection until sacrifice at 4 or 14 weeks (continuous treatment). A sub-group was Dox-treated for 7 weeks (temporary treatment) then untreated until 14 weeks. In the absence of Dox, the GDNF tissue concentration was found to be equivalent to the endogenous level in 6-OHDA-lesioned rats. In the presence of Dox, it was ~10-fold higher. Dox-dependent behavioral improvements were demonstrated 4 weeks post-vector injection. At later time points, spontaneous partial recovery was observed in all rats, but no further improvement was found in Dox-treated animals. Moreover GDNF gene delivery only transiently improved dopaminergic function. Over the long term, TH was more abundant, but not functional, and the increase was lost when GDNF gene expression was switched off. The third part of my study consisted in the evaluation of the respective dose-range of therapeutical and undesirable effects of GDNF. Functional effects appeared after delivery of 3.5 108 viral particles which produced 200-300 pg/mg protein of GDNF in the lesioned rat striatum (see above). In order to evaluate the viral dose producing undesirable effects, we compared two different doses of vector: 3.5x108 and 4.4x109 viral genome. In the low dose group, the GDNF concentration in the striatum was ~300 pg/mg protein in the Dox-treated animals and equivalent to the endogenous level in untreated animals (~20 pg/mg protein). In contrast, in the high dose group, GDNF levels reached ~1200 pg/mg protein in induced animals but up to ~300 pg/mg protein in uniduced animals. In the low dose group, Dox-dependent downregulation of TH but no asymetrical behaviour was evidenced. In the high dose group, TH downregulation was observed in both Dox+ and Dox-rats. In addition, amphetamine-induced rotational behaviour was evidenced in Dox+ but not in Dox-rats. These data suggest that low doses of virus are sufficient to induce therapeutically-relevant but not undesirable functional effects of GDNF. Nevertheless,a neurochemical effect of GDNF (TH down-regulation) did appear at low dose. In order to understand the GDNF-induced motor asymmetry, we investigated the anatomical pattern of TH down regulation in striatum. Strikingly, there was a greater loss of TH labeling in striosomes than in the surrounding matrix. Receptors which are known to be differentially expressed in the striosomes i.e. µ-opioid receptor(MOR-1) and N-methyl-D-aspartic acid (NMDA) receptor 1 (NR1) as compared to the matrix were analyzed in the high-dose group of animals. MOR-1 was not affected by GDNF gene delivery. In contrast, NR1 was down regulated. The potential relationship between TH and NR1 down-regulation as well as other previously described neurochemical effects of GDNF (as enhancement of DA release and metabolism, of DA neurons excitability or of TH phosphorylation) and behavioural asymmetry remains to be clarified. As summary, our data suggest that behavioural and neurochemical effects of striatal delivery of GDNF can be controlled by Dox by using the autoregulatory rAAV2/1-TetON- GDNF vector, provided the dose range of gene delivery is carefully adjusted. / Doctorat en Sciences biomédicales et pharmaceutiques / info:eu-repo/semantics/nonPublished

Dehydroepiandrosterone, a C-19 steroid, is found endogenously with the highest circulating serum levels. It is converted to important steroids such as the sex hormones oestrogen and testosterone. DHEA has come under the spotlight as a purported “fountain of youth” due to its well-characterised age-related decline. The supplementation of DHEA in both the elderly and those with a pathophysiological deficiency has been shown to be of benefit, particularly with regard to wellbeing and depression. The role of DHEA in the periphery has not been elucidated beyond its role as a precursor hormone in sex steroid biosynthesis, though it has been established as a neuroactive neurosteroid, capable of exerting neuroprotective effects in the brain. Since the importance of free radicals in aging and neurodegeneration is well established, investigations were conducted on the ability of DHEA to inhibit free radical generation or scavenge existing free radicals. DHEA was able to significantly inhibit quinolinic acid-induced lipid peroxidation, a measure of membrane damage, over a range of concentrations, although the reduction did not appear to be dose-dependent. This was observed in both in vitro and in vivo studies. Thus, the ability of a compound to reduce the degree of lipid peroxidation may indicate its value as a neuroprotectant. However, DHEA did not significantly reduce cyanide induced generation of the superoxide free radical, suggesting that DHEA is not an effective free radical scavenger of the superoxide anion and that the reduction in lipid peroxidation does not occur through a scavenging mechanism. Apoptosis is a physiological process which is necessary for development and homeostasis. However, this form of programmed cell death can be initiated through various mechanisms and too much apoptotic cell death results in deleterious effects in the body. DHEA was shown not to induce apoptosis. Even the lowest concentration of DHEA investigated in this thesis shows a remarkable decrease in the degree of apoptosis caused by intrahippocampal chemical insult by the neurotoxin quinolinic acid. Cresyl violet was used to visualise tissue for histological examination which revealed that DHEA is able to preserve the normal healthy morphology of hippocampal cells which have been exposed to quinolinic acid. Cells maintained their integrity and showed little evidence of swelling associated with necrosis. Organ culture studies were performed by assessing the impact of DHEA on several pineal metabolites. The study revealed that DHEA exerted an effect on the metabolism of indoleamines in the pineal gland. Melatonin, the chief pineal hormone, did not appear to be affected while the concentrations of N-acetylserotonin, serotonin and methoxytryptamine showed significant alterations. Thus, the neuroprotective mechanism of DHEA does not appear to be mediated by an increase in the presence of melatonin. The biological importance of metal ions in neurodegeneration is also well established and thus the potential interaction between DHEA and metal ions was considered as a mechanism of action. Spectroscopic and electrochemical analyses were performed to determine whether DHEA is able to interact with metal ions as a ligand. These reveal that DHEA does not form a strong bond with the metals investigated, namely copper (II) and iron (III), but that a weak interaction is evident. These investigations were conducted in a rodent model, which has neither large amounts of endogenous DHEA, nor the enzymatic infrastructure present in humans. Thus, the theory that DHEA exerts its effects through downstream metabolic products is unlikely. However, these investigations reveal that there is merit in the statement that DHEA itself is a neuroprotective molecule, and confirm that the further investigation of DHEA is an advisable strategy in the war against neurodegeneration and aging.

The functional compartmentalization underlying neuronal polarity makes tightly regulated intracellular transport between the cell body, axons, and dendrites essential for proper development and homeostatic maintenance. Disruptions to neuronal trafficking are a major cause of neurodegenerative disease. Pathogenic variants in the microtubule motor protein KIF1A cause KIF1A Associated Neurological Disorder (KAND), a spectrum of rare neurodegenerative conditions. KAND is clinically and genetically heterogeneous, with a broad phenotypic spectrum and over a hundred pathogenic variants identified. KAND is poorly understood at both the clinical and molecular level, and there is currently no treatment. This work characterizes the natural history of KAND and describes a novel heuristic severity score. This severity score is then used to show how the location of pathogenic missense variants within the KIF1A motor domain correlates with disease severity, providing evidence the clinical phenotypic heterogeneity in KAND reflects and parallels the molecular phenotypes. Insights from the neuropathology of deceased KAND patients is used to focus a histopathologic assessment of the C3-Kif1aLgdg mouse model. C3-Kif1aLgdg/Lgdg mice have a cerebellar axonal torpedo phenotype, paralleling some of the pathological changes seen in the patients. Phenotypically, the C3-Kif1aLgdg mice were found to recapitulate some of the symptoms seen in patients including progressive spasticity and gait abnormalities associated with hind limb paralysis. To model the disease at a cellular level, iPSCs were derived from affected individuals and successfully used to generate neural stem cells and neurons. These patient-derived neurons were found to have increased markers of protein aggregates, a cellular phenotype that can be used to test potential treatments. Taken together, these studies provide foundational knowledge for future therapeutic development.

Indiana University-Purdue University Indianapolis (IUPUI) / Friedreich’s Ataxia (FRDA) is a rare and poorly understood autosomal recessive disease caused by a pathological deficiency of the mitochondrial protein frataxin. Patients suffer neurodegeneration, ataxia, diabetes, and heart failure. In an effort to understand the mechanisms of heart failure in FRDA, we investigated the role of the protein modification acetylation, which is highly abundant on mitochondrial proteins and has been implicated in regulating intermediary metabolism. Using mouse models of FRDA, we found that cardiac frataxin deficiency causes progressive hyperacetylation of mitochondrial proteins which is correlated with loss of respiratory chain subunits and an altered mitochondrial redox state. Mitochondrial protein hyperacetylation could be reversed by the mitochondria-localized deacetylase SIRT3 in vitro, suggesting a defect in endogenous SIRT3 activity. Consistently, frataxin-deficient cardiac mitochondria showed significantly decreased rates of fatty acid oxidation and complete oxidation to carbon dioxide. However, the degree of protein hyperacetylation in FRDA could not be fully explained by SIRT3 loss. Our data suggested that intermediary metabolites and perhaps acetyl-CoA, which is required for protein acetylation, are accumulating in frataxin-deficient mitochondria. Upon testing the hypothesis that mitochondrial protein acetylation is non-enzymatic, we found that the minimal chemical conditions of the mitochondrial matrix are sufficient to cause widespread non-enzymatic protein acetylation in vitro. These data suggest that mitochondrial protein hyperacetylation in FRDA cardiomyopathy mediates progressive post-translational suppression of mitochondrial oxidative pathways which is caused by a combination of SIRT3 deficiency and, likely, an accumulation of unoxidized acetyl-CoA capable of initiating non-enzymatic protein acetylation. These findings provide novel insight into the mechanisms underlying a poorly understood and fatal cardiomyopathy and highlight a fundamental biochemical mechanism that had been previously overlooked in biological systems.

Indiana University-Purdue University Indianapolis (IUPUI) / Huntington’s Disease (HD) is an inherited, autosomal dominant, neurodegenerative disorder. There is no cure for HD and the existing therapies only alleviate HD symptoms without eliminating the cause of this neuropathology. HD is linked to a mutation in the huntingtin gene, which results in an elongation of the poly-glutamine stretch in the huntingtin protein (Htt). A major hypothesis is that mutant Htt (mHtt) leads to aberrant Ca2+ homeostasis in affected neurons. This may be caused by increased Ca2+ influx into the cell via the N-methyl-Daspartate (NMDA)-subtype of glutamate receptors. The contribution of two major Ca2+ removal mechanisms, mitochondria and plasmalemmal Na+/Ca2+ exchangers (NCX), in neuronal injury in HD remains unclear. We investigated Ca2+ uptake capacity in isolated synaptic (neuronal) and nonsynaptic mitochondria from the YAC128 mouse model of HD. We found that both Htt and mHtt bind to brain mitochondria and the amount of mitochondriabound mHtt correlates with increased mitochondrial Ca2+ uptake capacity. Mitochondrial Ca2+ accumulation was not impaired in striatal neurons from YAC128 mice. We also found that expression of the NCX1 isoform is increased with age in striatum from YAC128 mice compared to striatum from wild-type mice. Interestingly, mHtt and Htt bind to the NCX3 isoform but not to NCX1. NCX3 expression remains unchanged. To further investigate Ca2+ homeostasis modulation, we examined the role of collapsin response mediator protein 2 (CRMP2) in wild-type neurons. CRMP2 is viewed as an axon guidance protein, but has been found to be involved in Ca2+ signaling. We found that CRMP2 interacts with NMDA receptors (NMDAR) and disrupting this interaction decreases NMDAR activity. CRMP2 also interacts with and regulates NCX3, resulting in NCX3 internalization and decreased activity. Augmented mitochondrial Ca2+ uptake capacity and an increased expression of NCX1 in the presence of mHtt suggest a compensatory reaction in response to increased Ca2+ influx into the cell. The role of NCX warrants further investigation in HD. The novel interactions of CRMP2 with NMDAR and NCX3 provide additional insight into the complexity of Ca2+ homeostasis regulation in neurons and may also be important in HD neuropathology.

Indiana University-Purdue University Indianapolis (IUPUI) / Introduction: During glutamate excitotoxicity, overstimulation of glutamate receptors leads to sustained elevation in cytosolic Ca2+ ([Ca2+]c), or delayed Ca2+ dysregulation (DCD), which is causally linked to cell death. There are two major hypothetical mechanisms for DCD: the continuous activation of N-methyl-D-aspartate-subtype of the ionotropic glutamate receptors (NMDAR) and the reversal of the plasmalemmal Na+/Ca2+ exchanger. However, the contribution of each of these mechanisms in DCD is not completely established. Major results: Neurons exposed to excitotoxic glutamate produced DCD, an increase in cytosolic Na+ ([Na+]c), and plasma membrane depolarization. MK801 and memantine, noncompetitive NMDAR inhibitors, added after glutamate, completely prevented DCD; however AP-5, a competitive NMDAR inhibitor, failed to do so. The NMDAR inhibitors had no effect on lowering elevated [Na+]c or on restoring plasma membrane potential, which are conditions suggesting NCXrev could be involved. In experiments inducing NCXrev, MK801 and memantine completely inhibited Ca2+ dysregulation after glutamate while AP-5 did not. Inhibition of NCXrev, either with KB-R7943 or by preventing the increase in [Na+]c, failed to avert DCD. However, NCXrev inhibition combined with NMDAR blocked by AP-5 completely prevented DCD. Overall, these data suggested that both NMDAR and NCXrev are essential for glutamate-induced DCD, and inhibition of only one mechanism is insufficient to prevent collapse of calcium homeostasis. Based on the data above, we investigated a NMDA receptor antagonist currently in clinical trials for reducing the effects of glutamate excitotoxicity, ifenprodil. Ifenprodil is an activity-dependent, NMDAR inhibitor selective for the NR2B subunit. We found that ifenprodil not only inhibited the NR2B-specific NMDAR, but also inhibited NCXrev. If ifenprodil is combined with PEAQX, a NMDAR inhibitor selective for the NR2A subunit, low concentrations of both inhibitors completely prevent DCD. Conclusion: The inhibition of a single Ca2+ influx mechanism is insufficient in preventing DCD, which requires simultaneous inhibition of both the NMDAR and NCXrev. These findings are critical for the correct interpretation of the experimental results obtained with these inhibitors and for better understanding of their neuroprotective actions.

Neurological disorders present a significant challenge in global health. With the increasing availability of imaging datasets and the development of precise machine learning models, early and accurate diagnosis of neurological conditions is a promising and active area of research. However, several characteristic factors in neurology domains, such as heterogeneous imaging, inaccurate labels, or limited data, act as bottlenecks in using deep learning on clinical neuroimaging. Given these circumstances, this dissertation attempts to provide a guideline, proposing several methods and showcasing successful implementations in broad neurological conditions, including epilepsy and neurodegeneration. Methodologically, a particular focus is on comparing a two-dimensional approach as opposed to three-dimensional neural networks. In most clinical domains of neurological disorders, data are scarce and signals are weak, discouraging the use of 3D representation of raw scan data. This dissertation first demonstrates competitive performances with 2D models in tuber segmentation and AD comorbidity detection. Second, the potentials of ensemble learning are explored, further justifying the use of 2D models in the identification of neurodegeneration. Lastly, CleanNeuro is introduced in the context of 2D classification, a novel algorithm for denoising the datasets prior to training. CleanNeuro, on top of 2D classification and ensemble learning, demonstrates the feasibility of accurately classifying patients with comorbid AD and cerebral amyloid angiopathy from AD controls. Methods presented in this dissertation may serve as exemplars in the study of neurological disorders using deep learning and clinical neuroimaging. Clinically, this dissertation contributes to improving automated diagnosis and identification of regional vulnerabilities of several neurological disorders on clinical neuroimaging using deep learning. First, the classification of patients with Alzheimer’s disease from cognitively normal group demonstrates the potentials of using positron emission tomography with tau tracers as a competitive biomarker for precision medicine. Second, the segmentation of tubers in patients with tuberous sclerosis complex proves a successful 2D modeling approach in quantifying neurological burden of a rare yet deadly disease. Third, the detection of comorbid pathologies from patients with Alzheimer’s disease is analyzed and discussed in depth. Based on prior findings that comorbidities of Alzheimer’s disease affect the brain structure in a distinctive pattern, this dissertation proves for the first time the effectiveness of using deep learning on the accurate identification of comorbid pathology in vivo. Leveraging postmortem neuropathology as ground truth labels on top of the proposed methods records competitive performances in comorbidity prediction. Notably, this dissertation discovers that structural magnetic resonance imaging is a reliable biomarker in differentiating the comorbid cereberal amyloid angiopathy from Alzheimer’s disease patients. The dissertation discusses experimental findings on a wide range of neurological disorders, including tuberous sclerosis complex, dementia, and epilepsy. These results contribute to better decision-making on building neural network models for understanding and managing neurological diseases. With the thorough exploration, the dissertation may provide valuable insights that can push forward research in clinical neurology.

T cells infiltrate the degenerating brain and influence central nervous system (CNS) inflammation and neuronal health. In mice, the choroid plexus and the meninges have been implicated in regulating T cell entry and egress from the CNS, respectively. Further, antigen presenting cells in the mouse meninges present CNS-derived antigens to T cells and may represent a method for the peripheral immune system to sense and respond to CNS immune triggers. However, whether these processes occur in the human choroid plexus and meninges has not been comprehensively studied. Further, the antigens towards which T cells in the degenerating human brain and its borders respond remain unknown. Therefore, I implemented a multi-omics approach using fresh postmortem tissue from patients diagnosed with amyotrophic lateral sclerosis (ALS), Alzheimer’s disease (AD), Parkinson’s disease (PD), and non-neurodegenerative controls to identify not only the T cell-associated changes that occur in the degenerating human CNS and surrounding tissues but also identified a library of putative antigen targets for disease-associated T cell populations. Specifically, using single cell RNA and TCR sequencing information from paired postmortem choroid plexus, leptomeninges, and brain I lineage traced T cells using their TCR information and found that T cell access to leptomeninges and brain is likely limited and controlled by anti-inflammatory macrophage activity at the blood/CSF barrier (BCSFB). Once past the BCSFB, I present evidence that T cells access the CNS where they interact with MHC expressed by microglia. T cells also accumulate in the leptomeninges where they become tissue resident memory T cells. These tissue resident memory T cells likely serve as a reservoir for a rapid antigen-driven immune response to future CNS inflammatory insults. Finally, by performing immunopeptidomics to identify peptides presented by MHC in the same patients’ CNS and border tissues, I identified a library of putative antigenic triggers that may drive high levels of T cell clonal expansion in the brain and surrounding tissues. Altogether, this thesis serves as a resource for understanding the trajectory of T cells as they travel into the degenerating human brain and as a foundation for the development of antigen-specific precision medicines to treat neurodegeneration.

Amyotrophic lateral sclerosis (ALS) is a progressive fatal neurodegenerative disease affecting mostly spinal motor neurons. Its hallmarks include motor neuron loss, muscle atrophy, paralysis and death. Studies in ALS using genetic mutants in mouse models have identified numerous changes in motor neurons at early stages, including excitotoxicity, oxidative stress, denervation at neuromuscular junction (NMJ), mitochondrial abnormalities and motor neuron excitability. Such changes are accompanied by abnormal motor unit function and motor output. Despite attempts to identify early changes in ALS, the causal relationship between abnormal motor output – broadly defined as motor unit dysfunction – and the reported pathological changes are poorly understood. Here, we used C57BL6/J-SOD1 G93A male mice as the animal model of ALS, together with behavioral, physiological and morphological assays to identify transcriptional changes in vulnerable motor neurons involved in motor dysfunction during ALS onset. We found that ALS mice travelled shorter distances as early as one month old, which progressively reduced to peak around ~P80 and dramatically decreased afterwards until death at ~P150. To identify changes responsible for the observed reduction in distance travelled in SOD1 mutants, we investigated mice at ages in which behavioral differences became apparent. We opted to evaluate the extent of motor unit pathology by focusing on a motor pool involved in running that is known to be vulnerable in ALS, the tibialis anterior (TA) muscle. Although TA motor units do not exhibit any detectable overt changes between SOD1 and control mice at P30, SOD1-G93A mice at P50 exhibited a significant ~40% loss of functional TA motor units, determined by in vivo experiments in which individual motor units were counted following incremental stimulation intensities and quantifying the elicited muscle force. This reduction was not accompanied by motor neuron loss. However, the loss in functional motor units was accompanied by significant decrease in the twitch force alongside ~40% NMJ denervation in the TA muscle. Taken together, the functional changes in motor units and NMJ denervation are deficits in motor unit function that can account for the reduction in distance travelled daily by SOD1 male mice. We next wanted to identify transcriptional changes in the TA motor neurons of SOD1-G93A mice that could be involved in the observed motor unit pathology. To this end, large TA motor neurons (presumed to be motor neurons from the affected vulnerable motor units) from control and SOD1 males at P30 and P50 were subjected to RNAseq which were selectively collected via laser capture microdissection following labelling with a fluorescent tracer through intramuscular injection. The resultant differentially expressed genes included 6 upregulated and 10 downregulated transcripts. Among these, a mitochondrial gene, Chchd10, which mutations have recently been associated with ALS, revealed elevated levels in motor neurons. Reduction of Chchd10 protein levels within CNS, including motor neurons, via intracerebroventricular injection with an antisense oligonucleotide, we observed a significant protection against motor unit loss that was translated with an increase in distance travelled daily by ALS-treated mice. These results suggest that dysregulation of Chchd10 in SOD1 G93A is involved in the pathogenesis of disease. Furthermore, whilst the downregulation of Chchd10 did not rescue all characteristics of motor unit pathology it implies that multiple genes dysregulation may be involved and be responsible for the different characteristics of motor unit pathology early in the disease onset. It further highlights the downregulation of Chchd10 as a potential novel therapeutic target for ALS.

Indiana University-Purdue University Indianapolis (IUPUI) / Alzheimer disease (AD) results, in part, from the excess accumulation of the amyloid-β peptide (Aβ) as neuritic plaques in the brain. The short Aβ peptide is derived from a large transmembrane precursor protein, APP. Two different proteolytic enzymes, BACE1 and the gamma-secretase complex, are responsible for cleaving Aβ peptide from APP through an intricate processing pathway. Dysregulation of APP and BACE1 levels leading to excess Aβ deposition has been implicated in various forms of AD. Thus, a major goal in this dissertation was to discover novel regulatory pathways that control APP and BACE1 expression as a means to identify novel drug targets central to the Aβ-generating process. MicroRNAs (miRNA) are short, non-coding RNAs that act as post-transcriptional regulators of gene expression through specific interactions with target mRNAs. Global analyses predict that over sixty percent of human transcripts contain evolutionarily conserved miRNA target sites. Therefore, the specific hypothesis tested was that miRNA are relevant regulators of APP and BACE1 expression. In this work, several specific miRNA were identified that regulate APP protein expression (miR-101, miR-153 and miR-346) or BACE1 expression (miR-339-5p). These miRNAs mediated their post-transcriptional effects via interactions with specific target sites in the APP and BACE1 transcripts. Importantly, these miRNA also altered secretion of Aβ peptides in primary human fetal brain cultures. Surprisingly, miR-346 stimulated APP expression via target sites in the APP 5’-UTR. The mechanism of this effect appears to involve other RNA-binding proteins that bind to the APP 5’-UTR. Expression analyses demonstrated that these miRNAs are expressed to varying degrees in the human brain. Notably, miR-101, miR-153 and miR-339-5p are dysregulated in the AD brain at various stages of the disease. The work in this dissertation supports the hypothesis that miRNAs are important regulators of APP and BACE1 expression and are capable of altering Aβ homeostasis. Therefore, these miRNA may possibly serve as novel therapeutic targets for AD.