Depressive symptoms in relation to markers of neurodegeneration, cerebrovascular injury and Alzheimer’s Disease (AD) in community‐dwelling older adults: a plasma and multimodal neuroimaging study

Abstract

AbstractBackgroundDepressive symptoms are common with aging and Alzheimer’s Disease (AD). Both neurodegenerative and cerebrovascular diseases have been postulated as underlying mechanisms, but it’s unclear if they contribute to depressive symptom pathogenesis in preclinical stages of AD. Probing these mechanisms using both central (neuroimaging) and peripheral (plasma) biomarkers is critical to gain insight into the neurobiology of depression and provide opportunities for clinical translation.MethodLongitudinal data were analyzed from 350 Harvard Aging Brain Study participants, an observational cohort study of older adults who are cognitively unimpaired and have no/minimal depressive symptoms at baseline. Depressive symptoms were assessed annually with the 30‐item Geriatric Depression Scale (GDS) (average follow up = 7.2+/‐3.4yrs). At baseline, all participants underwent blood sampling for a plasma marker of general neurodegeneration, neurofilament light (NfL) and 11 cerebrovascular markers in plasma. Participants also underwent baseline amyloid (PiB) PET‐ wherein a cortical amyloid composite was derived‐ and baseline and subsequent MRI scans every three years‐ to derive slopes of regional cortical thickness and global white matter hyperintensities (WMH). Linear mixed effects models examined relationships between longitudinal GDS scores and predictors: (NfL or cerebrovascular markers)*time, and covariates: age, sex, education. Secondary model examined plasma markers and amyloid as interactive predictors of GDS, and repeated analyses using predictors: regional cortical thickness (neurodegeneration) or global WMH (cerebrovascular injury) slopes.ResultHigher baseline levels of NfL and of ICAM1, a marker of cerebrovascular injury, predicted increasing GDS scores over time (NfL: β = 0.03+/‐0.01, t = 2.7, p = 0.007; ICAM1: β = 0.02+/‐0.01, t = 2.2, p = 0.03). Each also interacted with amyloid in predicting GDS scores: (NfL X PiB: β = ‐0.03+/‐0.01, t = ‐2.42, p = 0.02; ICAM1 X PiB: β = 0.023+/‐0.01, t = 2.44, p = 0.02). Regarding neuroimaging correlates, medialorbitofrontal cortical thickness slope (β = ‐0.03+/‐0.01, t = ‐2.94, p = 0.004), but not global WMH slope, predicted increasing GDS over time. None of the other cerebrovascular markers interacted with amyloid in predicting GDS.ConclusionFindings suggest that cerebrovascular injury, and peripheral and central biomarkers of regional neurodegeneration, alone and in synergy with amyloid burden, are associated with depressive symptoms in cognitively unimpaired older adults. Future studies incorporating additional measures of peripheral and central neurodegeneration, tauopathy, and cerebrovascular injury are needed to guide AD prevention efforts targeting depressive symptoms.