Neurobiological correlates of grey matter morphometry changes in preclinical Alzheimer’s disease

Abstract Background The driving mechanisms of structural brain alterations in the earliest stages of Alzheimer's disease (AD) are not well understood. Previous heterogeneous findings in preclinical AD, including subtle atrophy and also increased grey matter (GM) volume, underscore the need for further exploration. This study uses an extensive fluid biomarkers panel to identify pathological drivers behind longitudinal GM changes in cognitively unimpaired (CU) adults. Method We investigated 632 CU individuals (age 61.8±6.3, mean±SD) from the ALFA+, Wisconsin ADRC, and WRAP cohorts (Table 1), with available longitudinal MRI and cerebrospinal fluid biomarkers of Aβ 42/40 , p‐tau 181 , neurogranin, NfL, total‐tau, α‐synuclein, GFAP, YKL‐40, sTREM2, s100B, and IL‐6 measured using the NeuroToolKit panel of robust prototype assays (Roche Diagnostics International Ltd). We used non‐negative matrix factorization to decompose biomarkers into latent components. The per‐subject component weights were associated with voxel‐wise GM volume changes (years 3.5±0.9, mean±SD) and longitudinal performance on the preclinical Alzheimer's cognitive composite (PACC) using mixed model analysis. Result Six biomarker components (C1‐C6) were identified (Figure 1). High expression of astrocyte biomarkers (C1) was related to GM volume increases in frontotemporal and subcortical regions, and faster cognitive decline in A+ individuals, while axonal damage with glial reactivity (C2) was associated with temporal atrophy. Microglial reactivity (C3) showed subtle associations with both increases and decreases in GM volume in different regions. A component primarily based on Aβ pathology (C4) was strongly associated with widespread GM volume loss and faster cognitive decline. IL‐6 (C5) reflected non‐AD related inflammation. Finally, the co‐expression of tau pathology, neuronal injury, and synaptic damage (C6) showed robust associations with frontotemporal atrophy and cognitive decline in A+ individuals. Conclusion In a large longitudinal sample, we provide in vivo evidence that in the earliest stages of AD, Aβ pathology is also a strong contributor to future neuronal loss and cognitive decline besides neurofibrillary tangles and neuroaxonal damage. Astroglial reactivity (i.e. GFAP, s100B) in preclinical AD was associated with an increase in GM volume, which may suggest a neuroinflammatory response, and was associated with worse cognitive performance. These findings underscore the dynamic nature of structural brain changes within preclinical AD, shaped by the contributions from diverse pathophysiological processes.