Abstract
Abstract
Synapse loss is the strongest pathological correlate of cognitive decline in Alzheimer’s disease, yet the molecular basis of synaptic vulnerability across disease progression remains poorly understood.
We performed proteomic profiling of isolated synaptoneurosomes from
post-mortem
middle temporal gyrus tissue from 67 individuals, comprising 40 individuals spanning all Alzheimer’s disease Braak stages and 27 non-diseased controls. Differential abundance and weighted gene co-expression network analyses were combined with pathway, cell-type and gene set variation analyses to characterise stage-associated changes in the synaptic proteome. Brain-derived candidate proteins were subsequently evaluated using receiver operating characteristic and regression analyses against cognitive examination scores in two independent plasma proteomics datasets.
We identified six distinct stage-associated synaptic proteomic signatures. Mitochondrial function and synaptic signalling declined by mid-stage Alzheimer’s disease, while clathrin-mediated endocytosis increased at mid-stage without further change at later stages. By contrast, proteostasis and perisynaptic astrocytic processes increased progressively throughout the disease course. Global excitatory and inhibitory protein signals also showed a progressive shift in their relative balance (
P
= 7.80 × 10⁻
3
; adjusted
R²
= 0.072). These findings guided targeted investigation of candidate plasma markers linked to distinct synaptic processes in two independent proteomic datasets. We nominated 42 plasma proteins mapping to synaptic signatures of progression, which showed associations with disease stage and cognitive measures including Mini-Mental State Examination, Montreal Cognitive Assessment and Clinical Dementia Rating. PHGDH was prioritised across analyses through associations with cognitive impairment from early disease stages (Clinical Dementia Rating transition 0 to 0.5, false discovery rate = 8.87 × 10⁻
18
), neuropathological progression (
P
= 6.85 × 10⁻
7
) and excitatory-inhibitory protein imbalance (
P
= 5.56 × 10⁻
3
).
Together, our findings define the stage-associated molecular changes at the AD synapse and connect distinct brain pathways to accessible plasma candidates, establishing a mechanistically anchored framework for future patient stratification and evaluation of synapse-directed therapies.