Neural Substrates of Insulin-Mediated Memory Facilitation in Early Alzheimer's Disease; The Impact of the Apolipoprotein E-Epsilon-4 Allele on Hippocampal Insulin Responses

Abstract

Background: Several studies have demonstrated that insulin delivered by nasal spray acutely improves cognitive performance in early Alzheimer's disease. Furthermore, the apolipoprotein E-episilon-4 allele, a known AD risk factor, appears to influence those cognitive responses in a dose dependent manner, with carriers exhibiting impaired cognitive performance at higher insulin doses. The neural correlates of this phenomenon are however presently undefined. Hippocampal neurons are known to express insulin receptors in high densities, and neuronal insulin signaling is thought to modulate long-term potentiation mechanisms, suggesting that insulin might act at the hippocampus to generate cognitive enhancement. How the apolipoprotein E-epsilon-4 allele might interact with hippocampal insulin responses is not clear however. Methods: We studied the effects of a single 40IU dose of insulin aspart on hippocampal activation in 10 cognitively intact and 18 early AD subjects in a double-blind, counterbalanced, crossover memory encoding BOLD-based functional MRI study. We also assessed the effects of insulin on non-cognitive brain function with an fMRI motor task, and mean cerebral blood flow with arterial spin-labeled MRI. Last, insulin's effect on cognitive performance was assessed. We hypothesized AD subjects would demonstrate insulin-related increases in hippocampal activation, and that insulin-related changes would be greatest in apolipoprotein E-epsilon-4 negative AD. We further hypothesized insulin-related increases in mean cerebral blood flow in all subjects. Results: We observed an apolipoprotein E-epsilon-4 based interaction in insulin-related changes in hippocampal activation in AD, with carriers exhibiting decreased activation with insulin. Mean cerebral blood flow did not change significantly with insulin administration. Cognitive performance increased in the apolipoprotein E-epsilon-4 negative AD subjects after insulin administration. Conclusions: There is growing body of research indicating that faulty insulin signaling might be a component of AD neuropathology. The results herein are the first description of the neural correlates of insulin-related cognitive enhancement in AD, and provide direct evidence that insulin might serve an important neuromodulatory role in AD. Insulin also appears to modulate non-cognitive functions in a manner similar to hippocampal activity, suggesting that insulin might have wide spread effects on brain physiology, while the arterial spin-labeled MRI results suggest that insulin's effects are independent of cerebral blood flow. Our results furthermore underscore the importance of the apolipoprotein E-epsilon-4 allele in brain insulin responses.