Distinct regional patterns of synaptic vulnerability across hippocampal and parahippocampal subregions in Alzheimer's disease

Abstract Synaptic loss in the (para)hippocampus is a major contributor to cognitive decline in Alzheimer's disease (AD), yet its regional specificity and pathological correlates remain poorly understood. Here, we quantified synaptophysin‐positive puncta across hippocampal subregions (CA4, CA2, CA1, and subiculum) and parahippocampal cortex (entorhinal cortex, parahippocampal and fusiform gyrus) in postmortem tissue from 28 AD and 17 controls, and assessed relationships to neuropathological severity and cognitive decline. Amyloid‐β, phosphorylated tau (p‐tau) load and neurofilament light (NfL) immunoreactivity were quantified, and Clinical Dementia Rating scores were collected as a cognitive measure. Group differences were analyzed with linear mixed models; correlations between synaptic puncta, pathology and cognitive scores with linear regressions, corrected for age, sex and multiple comparisons. We found selective synaptic loss in the entorhinal cortex (−14%, p = 0.017) and parahippocampal gyrus (−14%, p = 0.021) in AD versus controls. Hippocampal synaptic density negatively correlated with amyloid‐β in controls ( r = −0.52, p < 0.001) and positively in AD ( r = 0.25, p = 0.007), suggesting a disease‐dependent shift. In AD, hippocampal, but not parahippocampal, subregions with higher p‐tau burden showed greater synaptic density ( r = 0.21, p = 0.003), raising questions about the role of p‐tau in synaptic loss at late stages. Axonal damage (i.e., higher NfL load) in the parahippocampal cortex associated with synaptic loss locally and in interconnected subregions. Worse cognitive performance was strongly associated with synaptic loss in the CA1 ( r = −0.64, p = 0.003), subiculum ( r = −0.62, p = 0.012), entorhinal cortex ( r = −0.60, p = 0.008), parahippocampal ( r = −0.48, p = 0.041) and fusiform gyrus ( r = −0.67, p = 0.002). These findings highlight the vulnerability of the entorhinal cortex and parahippocampal gyrus to synaptic loss in AD. Our results suggest that amyloid‐β and p‐tau pathology may play a limited role in synaptic loss at end‐stage disease, and the strong link between synaptic loss in (para)hippocampal subregions and cognitive decline underscores the need for monitoring synaptic change in light of disease progression and evaluating therapeutic interventions.