Alzheimer’s disease (AD) is on the rise in the U.S. and has no current cure or effective, lasting therapy. AD is associated with circadian and epigenetic disruptions, both of which may contribute to the neurodegeneration, cognitive impairment and dementia characteristic of AD. Epigenetic machinery is known to regulate the circadian clock, however preliminary studies in our lab have suggested that circadian regulators may also regulate epigenetic machinery. We explored the relationship between DNA methylation and circadian molecular mechanisms in the context of AD. Specifically, we investigated the action of circadian regulator Bmal1 in mediating the expression of DNA methyltransferase 1 (Dnmt1) in rat neurons. We hypothesized that a) Dnmt1 is expressed rhythmically in neurons, that b) Bmal1 acts as a transcription factor to regulate oscillation of Dnmt1, and that c) AD pathology alters this rhythmicity and regulation of Dnmt1. To generate an in vitro model of AD, amyloid precursor protein (APP) was overexpressed in adult rat hippocampal neurons. Dnmt1 transcription in the rat neurons was indeed found to be rhythmic and appeared to be altered in APP-overexpressing neurons. Additionally, we showed that Bmal1 does rhythmically bind to canonical E-boxes present at the Dnmt1 gene and that this binding is largely reduced with APP-overexpression. Our results not only provide an example of direct circadian control of DNA methylation in neurons, but also support our hypothesis that AD pathology disrupts this regulation. Defining the mechanism behind this relationship may further elucidate the molecular effects of AD pathology, as well as reveal possible circadian targets as future AD therapeutics.

The circadian clock drives the timing of many biological processes, from molecular dynamics of gene regulation and protein synthesis to animal behavior. Cell- and region-autonomous circadian oscillators within the brain are not well understood; however, it is known that circadian rhythm dynamics are pivotal to health, as circadian rhythm disruptions are a major hallmark symptom and driver of a range of neurological diseases. Additionally, dysregulated circadian rhythmicity is a key feature of advancing age, and may contribute to pathological aging. Alzheimer’s disease (AD) is a slowly progressing, debilitating neurodegenerative disorder affecting over 50 million people globally. AD is a disease of aging; however, the insidious pathological processes driving the disease are thought to begin decades before individuals show severe cognitive impairments. Some of the major and most disruptive symptoms of AD are disruptions in circadian rhythms, including impairment in sleep and daily rhythms in behavior and activity. This dissertation aims to characterize patterns of circadian gene regulation across the mouse brain under healthy conditions as well as in the context of pathological aging and Alzheimer’s disease using novel spatial transcriptomic technology, with the goal of bettering our understanding of region-specific circadian disruption as it relates to vulnerability to pathological aging and AD. In Chapter 1, I detail our work characterizing rhythmicity across the brain using spatial transcriptomics. In Chapter 2, I explore region-specific patterns in rhythmic gene expression in middle-aged mice compared to young adult mice to identify regional vulnerability to aging in the context of circadian rhythm disruption. In Chapter 3, I use a mouse model of AD to investigate region-specific patterns in differential rhythmic gene expression to identify AD pathology early and late into disease progression. Altogether this work shows that rhythmic patterns of gene expression vary across the mouse brain and provides evidence for the first time, that these patterns become disrupted in a region-dependent manner early in the course of aging and AD progression. This work has important implications for early identification and potential disease modifying treatments of AD and other diseases of aging.

Microglia, the brain’s resident macrophage, display a diverse array of phenotypes in the adult brain. Alzheimer’s Disease (AD) is associated with a unique subtype of these cells which transmit excessive inflammatory signals resulting in both synapse and cell loss. To investigate the role of DNA methylation in driving these responses, I profiled genome-wide methylation levels with single-nucleotide resolution in human microglia following in vitro exposure to either lipopolysaccharide or amyloid-beta. I defined specific changes in the methylation landscape and established an important role for both 5-methylcytosine and 5-hydroxymethylcytosine in directing inflammatory and phagocytic processes in response to different stimuli, thus implicating differential methylation as a possible mechanism by which microglia convert to an AD-associated neurotoxic phenotype. Furthermore, I provide preliminary evidence that DNA methylation changes are integral to mounting an inflammatory response in vivo.

Alzheimer’s disease (AD) causes devastating decreases in cognitive ability and quality of life for more than 5 million Americans. Genetic AD risk factors have been well researched, but emerging evidence calls attention to epigenetic changes. In order to screen for changes in DNA methylation during AD, we analyzed genome-wide methylation data from postmortem human frontal cortex samples from AD (n=26), mildly cognitively impaired (n=24), and healthy subjects (n=22). We identified significant changes in methylation in genes related to inflammation, including CXCL17, IKBKAP, IKBKG, IL3, IL13, IL17RB, TNFRSF8, A2M, BDKRB1, and TNF in association with AD pathology.

The relationship of epigenetics and inflammatory pathways was further investigated with an in vitro model. It was hypothesized that the aberrant DNA methylation observed in AD brains causes dysregulation of inflammatory pathways in glial cells, contributing to the neuro-inflammation that is characteristic of AD. Global DNA methylation was modulated in vitro using S-adenosyl methionine to increase or Decitabine to decrease methylation in cultured human microglia, which resulted in activation of inflammation-related genes. Moreover, we identified similar transcriptional changes between cells treated with Decitabine and those activated by amyloid-β (Aβ), including deregulation of TNF, ICAM1, ITGB1, and TNFRSF1B. Importantly, Decitabine treatment of microglia also caused an increase in APP expression, further suggesting a link between changes in DNA methylation and elements of AD pathology.

Alzheimer’s Disease (AD) is a severe neurodegenerative disease characterized by cognitive dysfunction caused by synaptic loss. Its pathogenesis is driven by the development of extracellular plaques due to the accumulation of misfolded amyloid-beta protein and the formation of neurofibrillary tangles (NFTs) caused by the aggregation of misfolded Tau. Due to protein aggregation and cell loss during neurodegeneration, microglia, the innate immune cells of the central nervous system (CNS), acquire a unique activation phenotype that triggers inflammation. My thesis aims to explore the role of DNA methylation in driving these inflammatory responses using human-immortalized microglial cells. We profiled these cells for various inflammatory and disease-associated microglia (DAM) markers following in-vitro exposure to either lipopolysaccharide to model bacterial infections or to amyloid-beta to model neuroinflammation in AD. To investigate global methylation mediating these inflammatory changes, we measured 5-methylcytosine (5-mC) and 5-hydroxymethylcytosine (5-hmC) levels at specific time points. We also explored the potential of TET inhibitors as a therapeutic approach to reduce neuroinflammation by measuring the changes in 5-mC and 5-hmC levels as well as inflammatory markers in cells treated with the TET-inhibitor C35. Taken in all, my results suggest that DNA methylation plays an essential role in eliciting the inflammatory response of microglia to different stimuli, thus implicating this epigenetic mechanism in the pathology of AD.

Background

Methods

Results

Conclusions