Investigation of Monoaminergic Neurotransmitter Systems as a Method to Study Alzheimer's Disease-Related Neurodegeneration

Abstract

Alzheimer’s disease (AD) is the most common neurodegenerative disease and leading cause of dementia currently afflicting 5.4 million people in the United States. Memory impairment and cognitive dysfunction are the most recognized clinical manifestations of AD, however, disturbances in vision, sleep, olfaction, psychosis, and depression also impair the normal daily functions of AD patients. These symptoms can be attributed to the defining histopathological changes in AD and loss of neuronal populations in the hippocampus, cortex, olfactory bulb, brainstem, and retina – regions that are responsible for learning, memory, cognition, olfaction, wakefulness, affect, and vision. Current treatments for AD include reversible acetylcholinesterase inhibitors and N-methyl-D¬-aspartate (NMDA) receptor antagonists, however, these medications only provide mild symptomatic benefits in a subset of AD patients. A great deal of work aimed at understanding the ongoing pathological processes in AD and its progression have resulted in a wealth of transgenic mouse models currently saturating the field. However, despite the abundance of AD mouse models available, most of these models have failed to recapitulate substantial neuronal loss in regions relevant to AD [1-4]. Specifically, mouse models of Aβ pathology fail to show subsequent NFT/NT formation and profound neurodegeneration. In contrast, models of tauopathy vary in their neurodegenerative phenotype, but typically rely on mutant tau transgenes that are not seen in AD [1, 5]. Taken together, these findings have led to the notion that Aβ/APP models of AD only recapitulate the earliest changes of the disease and are alone insufficient to cause progressive neurodegeneration [6]. To address whether Aβ pathology is indeed sufficient for progressive neurodegeneration, recent work looked at neuronal populations in the brainstem where the dorsal raphe (DR) and locus coeruleus (LC) degenerate early in AD [7-11]. These DR and LC regions contain serotonergic and noradrenergic neurons, respectively, which send out abundant projections throughout the entire brain [10, 12-16]. Early work on human AD samples showed that the DR and LC neurons are among the earliest affected in the disease process [9-11, 17-20]. Interestingly, this degeneration is limited to the anteromedial portions of the DR and LC, while caudal regions are left unaffected. This topography of tau pathology and neurodegeneration is consistent with the fact that rostro-medial portions send their axonal projections to cortical and hippocampal regions, which display abundant amyloid pathology. The caudal DR and LC, however, send their projections to the spinal cord and cerebellum – areas without amyloid deposition – and are, thus, spared from neurodegenerative changes [10, 21, 22]. Findings in the APPSwe/PS1ΔE9 line showed that progressive axonal and somatic degeneration of serotonin and noradrenergic neurons followed persistent Aβ deposition/pathology [7]. Furthermore, this degeneration could be attenuated with the administration of anti-Aβ immunotherapy [8]. The analyses of brainstem monoaminergic (MAergic) systems in the APPSwe/PS1ΔE9 was the first to show that Aβ pathology could induce progressive neurodegeneration in vivo [23]. Subsequent studies expanded on these observations showing degeneration initiating at axon terminals and progressing to cell body degeneration [7]. Furthermore, defining a potential cellular or molecular mechanism for MAergic neurodegeneration remains unanswered. To better identify the factors leading to MAergic neurodegeneration in mouse models of Aβ pathology, we set out to address certain questions that were left open. First, is degeneration of MAerigc systems a common feature of Aβ mouse models? Second, is MAergic degeneration dependent on an interaction between mutant APP and PS1 transgenes? Third, is degeneration of MAergic systems dependent on Aβ signaling through cellular prion protein (PrPC)? By utilizing a mouse model expressing mutant APP transgenes in the absence of mutant PS1, known as the J20 model, we show that MAergic degeneration is also recapitulated in this mouse model (chapter 2). The presence of MAergic degeneration in this model, in addition to other previously observed models, supports the argument that Aβ pathology is sufficient for MAergic neurodegeneration. Furthermore, the absence of a mutant PS1 transgene in the J20 model indicates that MAergic neurodegeneration is indeed due to Aβ toxicity, not an artificial interaction between mutant APP and PS1 transgenes. Interestingly, by comparing the results from the J20 model to the previously reported APPSwe/PS1ΔE9 model, we observe differential onsets on Aβ deposition and MAergic neurodegeneration. Specifically, the J20 model has an approximate 3-4 month delay in Aβ pathology compared to the APPSwe/PS1ΔE9 mouse model. This results in a delay in the onset of MAergic degeneration. Whereas axonal degeneration begins at 8 months in the APPSwe/PS1ΔE9 model [7], when Aβ deposits are abundant, J20 mice just begin to show Aβ plaques at this time point and MAergic axons are still intact. Looking at aged mice (~16 months), J20 mice have significantly less noradrenergic axonal afferents in cortical and hippocampal regions, but no loss of MAergic cell bodies. In contrast, APPSwe/PS1ΔE9 animals at this age show a significant loss of MAergic neurons consistent with the earlier onset of MAergic axonal degeneration (Table 1.1.) [7]. To address whether PrPC is essential for MAergic neurodegeneration, we utilized a genetic approach to conditionally remove the Prnp gene, which encodes PrPC. Recent work by several labs have shown that PrPC is essential in Aβ toxicity. Specifically, constitutive removal of the Prnp gene, which encodes PrPC, fully rescues spatial learning and memory impairments in APPSwe/PS1ΔE9 mice [24]. However, given that the current diagnosis of AD occurs after Aβ deposition has already occurred, understanding whether conditional deletion of Prnp may provide a better understanding to the potential efficacy of targeting PrPC in patients. By using a Cre-Lox system to conditionally remove PrPC in APPSwe/PS1ΔE9, we are able to address if inhibiting Aβ oligomer (Aβo) signaling through PrPC is able to attenuate ongoing MAergic neurodegeneration (chapter 3). Despite efforts to identify mediators of Aβ toxicity in AD, the mechanistic insight triggering neurodegeneration remains poorly understood. Some of this ambiguity may be due to the inability of preclinical mouse models to capture the complexity of human AD cases. However, by looking at other systems affected in AD, it may be possible to better understand the cellular events leading to neurodegeneration. The classical view of AD is that it is memory disorder, but multiple faculties become compromised in people affected by the disease. Vision impairment has been noted as an early symptom in AD cases and has remained an interesting topic of investigation [25]. The association between vision and cognition or specifically, age-related macular degeneration (AMD) and AD have been explored several times in the past 25 years [25-37]. While AMD and AD, or vision and cognition, may seem as an unlikely link, multiple studies have confirmed the ability of poor vision to negatively impact cognition [25, 34-36]. Population based studies following individuals with early and late AMD have shown increased incidences of cognitive impairment compared to controls without AMD [26-28, 35]. Cognitive impairment, on the other hand, does not seem to affect vision [32]. Thus, the current interpretation is that vision impairment occurs early in AD and may represent a possible mode of detecting early AD and predicting disease severity. Still, conclusions about AMD and AD are controversial. Possible causes of this controversy may include the different methods of testing for visual and cognitive functions as well as how to properly control for risk factors. Indeed, Baker and colleagues (2009) found that early AMD was associated with lower cognitive tests. When they evaluated their results again using a modified mini mental state exam (mMMSE), this association was no longer present. Another study found associations between AMD and risk of AD, but not after controlling for smoking and atherosclerosis [26]. By evaluating common pathologies in AMD and AD, it may be possible to understand the potential mechanisms leading to cell loss. Using the links between AMD and AD, (such as risk factors, inflammation, and oxidative stress) we set out to establish a method to investigate cell loss in AMD and AD. Specifically, inflammation, oxidative stress, and mitochondrial DNA (mtDNA) damage are seen in degenerating regions of the eye in AMD and degenerating regions of the brain in AD [38-44]. By looking at human control and AMD tissues, we show that mtDNA damage is present in AMD retinal pigment epithelium (RPE) cells (chapter 4). However, mtDNA damage is not increased in the neural retina [41, 44]. Following up on these results, we show that at a stage characterized as “dry” AMD, there is no relationship RPE and retinal mtDNA damage in individual donors [41]. This demonstrated that mtDNA damage as an important change leading to cell death. By using the long-extension PCR (LX-PCR) technique to evaluate mtDNA damage for AMD, this technique may also be used to evaluate degenerating DR and LC neurons in AD.