Influence of Cerebrospinal Fluid Biomarkers of Alzheimer's Disease on Cognitive and Brain Aging

Abstract

The biological and cognitive heterogeneity observed in aging complicates the distinction of preclinical Alzheimer’s disease (AD) from normal aging. By utilizing well-validated core biomarkers of AD, such as amyloid β (Aβ) and phosphorylated tau (pTau) measured in cerebrospinal fluid (CSF), and analyzing their multivariate structure and relationships with measures of structural brain volumes and cognitive function in cognitively unimpaired (CU) healthy older adults, this approach may help to mitigate the effects of heterogeneity. Additionally, it may facilitate better identification of at-risk populations for recruitment into clinical trials and enhance our understanding of the early pathophysiological cascade of AD. This collection of three studies aims to investigate methods of using CSF biomarkers, specifically Aβ42 and pTau181, either directly by using them to identify healthy older adults at increased risk of cognitive decline, or indirectly by identifying the influence of aging, distinct from AD-related biomarkers and risk factors, by evaluating their impacts on magnetic resonance imaging (MRI) measures of regional brain volume. Initially, I employed a novel hybrid clustering approach to CSF Aβ42 and pTau181 data to identify subgroups within a cohort of CU healthy older adults. The clustering solution identified three subgroups, with a high-risk AD-like cluster containing a higher prevalence of known biomarker risk factors associated with the development of AD pathology, including a greater frequency of the apolipoprotein (APOE) ɛ4 allele, older age, and increased lateral ventricular and white matter hyperintensity (WMH) volumes. The high-risk cluster group had lower baseline composite scores of memory and executive function, poorer performance on two baseline composite measures sensitive to the early cognitive effects of preclinical AD (mPACC), and greater two-year declines on the Clinical Dementia Rating – sum of boxes score (CDR-sb) and an mPACC measure. Next, I investigated how these cluster subgroups related to subcortical brain structures using a multivariate network analysis method. I identified a regionally distributed pattern of subcortical gray matter (SGM) volumes that was related to the three clusters (Cluster-SGM) defined by CSF Aβ42 and pTau181, which was characterized by relative reductions in bilateral hippocampus, amygdala, and thalamus with relative increases bilaterally in the caudate and putamen. Higher expression of the Cluster-SGM pattern was associated with greater two-year declines in the mPACC, but not the CDR-sb. Finally, I investigated the age-related differences in the regional pattern of hippocampal subfield volumes (Age-Hsub) in a separate group of CU older adults without the major genetic risk factor for AD, the APOE ɛ4 allele, and after additionally statistically adjusting for the effects of common risk factors and biomarkers of AD pathology. The Age-Hsub pattern was characterized by reductions of bilateral dentate gyrus (DG) and right Cornu Ammonis (CA) regions with relative preservation in bilateral subiculum. Additionally, multiple regression analysis showed that greater expression of the Age-Hsub pattern was associated with poorer performance on objective and subjective measures of memory. Taken together, these results highlight how CSF biomarkers of AD pathology can be used to enhance our understanding of those at greatest risk for cognitive decline and AD, as well as of the potentially distinct effects of brain aging separate from the risk for AD dementia in CU healthy older adults. The results support the use of clustering techniques for identifying CU healthy older adults at increased risk for cognitive dysfunction, as well as regionally distributed patterns of GM atrophy that are associated with cognitive decline. These findings help to advance our understanding of the early stages of the AD continuum and may also help in identifying those who might benefit most from early disease-modifying treatments and prevention therapies.