The relationship between APOE genotype, CSF Tau and cognition across the Alzheimer's disease spectrum, moderation and mediation role of insula network connectivity

Abstract

AbstractAimsTo investigate whether insula network connectivity modulates the relationship between apolipoprotein E (APOE) ε4 genotype, cerebrospinal fluid (CSF) biomarkers (Aβ, Tau, and pTau) and cognition across Alzheimer's disease (AD) spectrum.MethodsForty‐six cognitive normal (CN), 35 subjective memory complaint (SMC), 41 mild cognitive impairment (MCI), and 32 AD subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) were obtained. Multivariable linear regression analyses were conducted to investigate the main effects and interaction of the APOE genotype and disease status on the insula functional connectivity (IFC) network. Mediation and moderation analysis were performed to investigate whether IFC strengths regulate the association between APOE genotype, CSF biomarkers and cognition. Additionally, the support vector machine (SVM) model integrating APOE genotype, CSF biomarkers, and neuroimaging biomarkers (insula volumes and altered regional IFCs) was used to classify the AD spectrum.ResultsThe interactive effect of the APOE genotype and disease on the insula network was found in the left medial superior frontal gyrus (SFGmed.L), right anterior medial prefrontal cortex (aMPFC.R), and bilateral thalamus (THA.B). The functional connectivities (FCs) in the left insula (LIns) connecting with the left posterior middle temporal gyrus (pMTG.L), SFGmed.L, and right lingual gyrus (LING.R) were correlated with cognition. LIns‐SFGmed.L and LIns‐pMTG.L FCs could moderate the effects of Tau on cognition. Furthermore, LIns‐SFGmed.L FC may suppress the association between APOE genotype and cognition. More importantly, the integrated biomarkers from the SVM model yielded strong powers for classifying the AD spectrum.ConclusionsInsula functional connectivity regulated the association between APOE genotype, CSF Tau and cognition and provided stage‐dependent biomarkers for early differentiation of the AD spectrum. The present study used a cross‐sectional design. Follow‐up studies are needed to validate the relationship.