Correlations of synaptic and pathological markers with cognition of the elderly

It has been suggested that the physical basis for dementia is structural or functional loss of synapses. To confirm this finding, we performed an enzyme-linked immunoassay (ELISA) with a monoclonal antibody (EP10) to a synaptophysin-like protein in brain samples from 45 prospectively studied elderly subjects with an average age of 83.3 +/- 10.1 years. We compared the synaptic marker to immunoreactivity with a newly developed PHF antibody (TG3). The cases were selected on the basis of availability of frozen tissue, and included subjects ranging from clinically normal to end-stage dementia. As an initial assessment, we determined Pearson product moment correlations for two clinical measures--the Blessed test of information, concentration, and memory (BICM) and the Fuld object Memory Evaluation (FOME)--with ELISA data and with traditional pathologic markers. We found strong correlations (p < 0.01-0.001) for BICM with brain weight, neuronal loss in the basal nucleus of Meynert (nbM), counts of senile plaques (SP) in the neocortex and hippocampus, and neurofibrillary tangles (NFT) in all areas except the parahippocampal cortex. Except in the occipital lobe, where paired helical filament changes are relatively uncommon, TG3-immunoreactivity also correlated strongly with BICM. Weak correlations (p < 0.05) were found for BICM with EP10-immunoreactivity in only the temporal and parietal lobes. Only the pathologic variables showed any significant correlations with FOME. Because inclusion of normal subjects with few or no pathologic lesions could have been driving the strong correlations with pathologic markers, we limited the analysis to those subjects with dementia (BICM; 8). After making this correction, EP10-immunoreactivity in all cortical areas and the hippocampus correlated better (p < 0.05-0.01) with BICM but not FOME. The present univariate analysis suggests that synaptic markers may not be the best structural correlate of dementia and that markers indicative of cytoskeletal changes, e.g., SP, NFT and PHF protein accumulation, may be better correlates of dementia in the elderly.