Loss of association between plasma irisin levels and cognition in Alzheimer's disease

Background: Irisin, an exercise-induced myokine, has been shown to have beneficial effects on cognitive and metabolic functions. However, previous studies assessing the levels of circulating irisin in patients with Alzheimer's disease (AD) or diabetes mellitus (DM) have provided inconsistent results. This suggests that the normal physiological action of irisin may be altered by disease-associated pathological conditions in target organs. Objective: To investigate the association of plasma levels of irisin with cognition and brain structures according to the presence or absence of AD and DM. Methods: Plasma levels of irisin, multi-domain cognition, and volumes of relevant brain regions were assessed using enzyme-linked immunoassay, neumpsychological test, and magnetic resonance imaging, respectively. We classified 107 participants by cognitive (cognitively normal [CN, n = 23], mild cognitive impairment [MCI, n = 49], and AD [n = 35]) and metabolic (non-DM [n = 75] and DM [n = 32]) states. Results: Disease state-stratified multiple regression analyses showed that plasma levels of irisin were positively associated with cognition only in participants without AD (CN plus MCI). By contrast, in participants with AD, these associations lost significance, and furthermore, higher levels of irisin indicated smaller hippocampal, superior temporal, and inferior frontal volumes. The association between plasma irisin levels and cognition was not affected by the presence of DM. Consistently, moderation analysis revealed that the relationship between plasma irisin levels and cognition or brain structures was significantly modified by the presence of AD, not that of DM. Conclusion: Our findings suggest that the beneficial actions of circulating irisin on cognition may be attenuated by AD-induced pathological conditions in the brain.