The Amyloid-β Protein in Alzheimer's Disease Contributes to Impairments in Glucose Regulation

Abstract

Epidemiological studies of type-2 diabetes mellitus (T2DM) and Alzheimer’s disease (AD)—two common, chronic, costly diseases with rapidly expanding patient populations—provide evidence of a significant link between systemic metabolic disturbances and cognitive decline. The accumulation of (Aβ) plaque pathology is one of the two defining characteristics of AD; the other is formation of neurofibrillary tangles within neurons. Post-mortem cultures from AD brain demonstrate brain insulin resistance, and neurons adjacent to Aβ plaques show a decrease in surface insulin receptors. It is not clear if systemic insulin resistance associated with T2DM promotes AD-specific neuropathogenesis or, on the other hand, increases neuronal vulnerability, which induces the expression of APP, the acute phase protein of the neuron. Likewise, it is unclear if the opposite is true: that AD neuropathologies impair insulin signaling and in this way alter susceptibility to systemic metabolic consequences such as changes in glucose tolerance and insulin tolerance that are characteristic of T2DM. Aβ measurements in human subjects suggest that T2DM does not elevate Aβ deposition. Therefore, I hypothesized that the comorbidity of AD and T2DM is due to the ability of Aβ to impair peripheral glucose regulation.

In order to more directly address my hypothesis, I employed the BRI-Aβ1-42 mouse model for my studies. In this model, human Aβ1-42 is overexpressed in the nervous system without overexpressing APP. This circumvents the problems of using a mutated APP sequence and overexpression of other proteolytic fragments of APP. In order to determine if blood glucose was altered by an alternative proteolytic product, sAPPα, I analyzed a mouse line overexpressing human sAPPα. Comparisons to APP-knockout mice indicated that APP or some fragment thereof may contribute to the glucose dysregulation that occurs in a diet-induced prediabetic syndrome. To determine if endogenous mouse APP contributed to the phenotype of the BRI-Aβ1-42 mice, I crossed the latter with an APP-knockout mouse. I also hypothesized that one of the major metabolic control centers in the brain, the hypothalamus, could be the location where Aβ exerts influence over glucose homeostasis. I analyzed transcript expression levels of peptides in the melanocortin system that control feeding and energy expenditure, parameters which were measured in Comprehensive Lab Animal Monitoring System cages. I finally assessed contribution of glucocorticoids to changes in blood glucose levels by targeting the hypothalamic-pituitary-adrenal (HPA) axis through adrenalectomy.

Our data is consistent with the hypothesis, i.e., that the neural expression of human Aβ1-42 (and not sAPPα) impairs glucose tolerance; however, it appears to do so without impacting insulin production or peripheral effects. Nevertheless, low expression levels of insulin receptor substrate-1 (IRS-1) and glucose uptake in the cerebrum in the Aβ-transgenic mice indicated central insulin resistance and/or deficient glucose transport. Adrenalectomy partially corrected the glucose impairment seen in Aβ-transgenic mice, indicating that the HPA axis contributes to the CNS-to-periphery manifestation of Aβ’s impact on glycemic control.

These findings indicate that the seminal event in AD pathogenesis, central accumulation of Aβ, is capable of perturbing peripheral glucose regulation via the HPA axis without affecting peripheral insulin resistance. This likely involves excessive hepatic glucose regulation and may indicate that other aspects of AD involve the activation of stress-related physiology. The latter is consistent with evidence of elevated glucocorticoids in human AD patients. The role of the HPA axis and glycemic dysregulation in the dementia and other symptoms of AD deserve further scrutiny.