Vascular endothelial growth factor associated dissimilar cerebrovascular phenotypes in two different mouse models of Alzheimer's Disease

Highlights

• Two mouse models of Alzheimer's Disease showed distinct cerebrovascular phenotypes.

• 5xFAD exhibit early hypoperfusion followed by a rebound to normal blood flow.

• Biphasic blood flow pattern in 5xFAD is driven by increased cerebral blood volume.

• TgCRND8 showed decreased cerebral blood flow and hypometabolism at later stage.

• Vascular endothelial growth factor may be responsible for increased vascular tone.

Abstract

Vascular perturbations and cerebral hypometabolism are emerging as important components of Alzheimer's disease (AD). While various in vivo imaging modalities have been designed to detect changes of cerebral perfusion and metabolism in AD patients and animal models, study results were often heterogenous with respect to imaging techniques and animal models. We therefore evaluated cerebral perfusion and glucose metabolism of two popular transgenic AD mouse strains, TgCRND8 and 5xFAD, at 7 and 12 months-of-age under identical conditions and analyzed possible molecular mechanisms underlying heterogeneous cerebrovascular phenotypes. Results revealed disparate findings in these two strains, displaying important aspects of AD progression. TgCRND8 mice showed significantly decreased cerebral blood flow and glucose metabolism with unchanged cerebral blood volume (CBV) at 12 months-of-age whereas 5xFAD mice showed unaltered glucose metabolism with significant increase in CBV at 12 months-of-age and a biphasic pattern of early hypoperfusion followed by a rebound to normal cerebral blood flow in late disease. Finally, immunoblotting assays suggested that VEGF dependent vascular tone change may restore normoperfusion and increase CBV in 5xFAD.

Introduction

Increasing evidence suggests that vascular perturbations are a potential driver of neuronal degeneration and cognitive decline in Alzheimer's disease (AD) (de la Torre, 2004; Ostergaard et al., 2013; Zlokovic, 2011). For instance, there is a well-established association between cerebral hypoperfusion and memory impairment as the disease progresses (Austin et al., 2011; Sierra-Marcos, 2017). The most profound hypoperfusion occurs in areas that are known to be sites of early plaque accumulation and those that control memory storage and retrieval, drawing a link between vascular perturbations and the primary symptoms of the disease. Furthermore, the brains of AD patients are less able to increase blood flow in response to neural activity (neurovascular coupling) (Girouard and Iadecola, 2006; Iadecola, 2004), causing already affected areas to be deprived of blood flow when attempting critical tasks. In addition to having vascular abnormalities as the result of the disease, there is also evidence that vascular disturbances such as those encountered with hypertension and diabetes could play a role in the cause or progression of AD (Viswanathan et al., 2009). Whether cause, effect, or some combination of both, there is a clear association between altered cerebrovascular function and disease progression in AD. In addition to hypoperfusion, AD patients also frequently show decreased cerebral metabolism as the disease progresses (Herholz et al., 2011). The decrease in cerebral metabolism may be a consequence of decreased perfusion as reduced blood delivery would entail lower glucose availability. While the exact mechanisms of hypometabolism are still under investigation, there is a clearly consistent pattern of decreased glucose uptake in AD patients.

In order to aid in diagnosis of AD, noninvasive imaging technologies have been employed to detect hemodynamic abnormalities and hypometabolism. Recent advances in such neuroimaging techniques have contributed greatly to earlier and more reliable detection in AD patients (Montagne et al., 2016). Of the two classical pathologies, in vivo amyloid imaging has become well established, with tau imaging also showing promising recent developments (Brosch et al., 2017; Chandra et al., 2019). Methods originally developed to assess blood flow in ischemic or traumatic brain injury have also been adapted for the diagnosis of AD. Imaging techniques commonly employ magnetic resonance imaging (MRI) or positron emission tomography (PET) to investigate aspects of blood flow, metabolism, or amyloid and tau accumulation that have been observed to change throughout the course of the disease. For example, arterial spin labeling (ASL) MRI and ¹⁸F-Fluorodeoxyglucose positron emission tomography (FDG-PET) have revealed that patients with AD show decreased perfusion and cerebral glucose metabolism compared to non-demented control subjects (Chen et al., 2011; Du et al., 2006). Furthermore, hypometabolism of specific regions can be seen in clinical as well as presymptomatic AD patients (Mosconi et al., 2010), making FDG-PET a valuable tool for early diagnosis. As a result of continued positive findings, noninvasive imaging has become more widely adopted as an early diagnostic tool in AD. PET imaging is commonly employed to complement behavioral testing in diagnosis of early AD and mild cognitive impairment, with FDG and amyloid targeting tracers used most commonly (Caminiti et al., 2018; Kato et al., 2016; Rice and Bisdas, 2017).

Considering the clinical and diagnostic significance of neurovascular and metabolic abnormalities in human AD patients, it is crucial to understand if animal models of AD display similar characteristics and if separate models vary in regard to these modalities. Transgenic murine models of AD have long been the mainstay for early and preclinical AD research. These mice usually contain mutated human copies of genes for amyloid-β precursor protein (APP), presenilin (PS1 or PS2), or some combination thereof. These mutations lead to cerebral amyloid deposition and result in effects that resemble what is observed in human AD, such as cognitive decline, inflammation, and neurodegeneration (Hall and Roberson, 2012).

Various imaging modalities have been previously employed in AD mice and have yielded assorted and sometimes contradictory results (Deleye et al., 2016; Dubois et al., 2010; Hebert et al., 2013; Klohs et al., 2014; Macdonald et al., 2014; Rojas et al., 2013; Waldron et al., 2015; Xiao et al., 2015). Findings have included hypoperfusion, atrophy of the hippocampus and cortex, hypometabolism and hypermetabolism. It can be difficult to interpret or compare experiments that are conducted by separate researchers because these studies are often heterogeneous with respect to imaging procedures, anesthetic protocols, handling conditions, data acquisition, or image analysis. It is also often assumed that different AD mouse strains will show similar pathology, which entails that finding in one strain may guide future studies in another. We set out to test these assumptions by performing imaging studies in two AD mouse strains under identical conditions, focusing primarily on aspects of cerebral blood flow and metabolism. The two strains used were TgCRND8 and 5xFAD, which were selected for specific differences in inserted genes and predominant type of Aβ, which may result in different cerebrovascular phenotypes. Specifically, 5xFAD miceharbor presenilin mutations (in addition to APP mutations) and accumulate predominantly Aβ₄₂, whereas TgCRND8 mice carry only APP mutations, resulting in accumulation of a more heterogeneous amyloid pool with the majority being Aβ₄₀ (Van Vickle et al., 2007). These mice are also reported to exhibit different rates of disease progression, with 5xFAD mice showing earlier amyloid deposition and more rapid disease course along with neuronal cell loss, which is not prevalent in TgCRND8 mice (Oakley et al., 2006).

We used three in vivo imaging modalities including ASL-MRI, dynamic susceptibility contrast-enhanced magnetic resonance imaging (DSC-MRI), and FDG-PET. These methods are meant to measure cerebral blood flow (CBF), cerebral blood volume (CBV), and cerebral metabolism, respectively. In addition to characterizing the cerebrovascular phenotype of these mice, these imaging modalities were also evaluated for their ability to monitor disease progression in AD mice. We have also investigated the molecular mechanism behind alteration in perfusion in these two strains focusing on neuroinflammation and vascular tone.