Premium
IC‐P‐166: CEREBROSPINAL FLUID TAU, BUT NOT AMYLOID, TRACKS ATROPHY IN ALZHEIMER‐VULNERABLE HIPPOCAMPAL SUBFIELDS
Author(s) -
Kerchner Geoffrey A.,
Czirr Eva,
Bernstein Jeffrey D.,
WyssCoray Tony,
Rutt Brian K.
Publication year - 2014
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2014.05.173
Subject(s) - hippocampal formation , atrophy , cerebrospinal fluid , hippocampus , pathology , alzheimer's disease , dementia , psychology , cognitive impairment , autopsy , neuroscience , medicine , cognition , disease
Background: Hippocampal atrophy begins quite focally in Alzheimer’s disease (AD). We have used high-resolution 7.0-Tesla MRI to visualize hippocampal subfields and strata, including the CA1 stratum radiatum / lacunosum-moleculare (CA1-SRLM, a site of key synapses for learning and memory), and showed that CA1-SRLM atrophy is evident in AD, is exacerbated by the ApoE4 gene, and correlates with memory performance (Kerchner et al., 2010, 2012, 2013, 2014). Autopsy studies have suggested that the hippocampus is one of the first brain structures to exhibit neurofibrillary tau pathology but one of the last to develop amyloid plaques. We postulated that CA1-SRLM atrophy would correlate with cerebrospinal fluid (CSF) measures of tau but not Ab42. Methods: We recruited older adults across a cognitive spectrum, including healthy controls (n1⁄412) and patients with Subjective Cognitive Impairment (n1⁄43), amnestic Mild Cognitive Impairment (n1⁄410), and mild AD dementia (n1⁄45). Each subject underwent lumbar puncture for CSF collection, and we determined CSF Ab42, total tau, and phospho-tau levels using standard enzyme-linked immunosorbent assays. Each subject also underwent 7T MRI for hippocampal microstructural analysis at an in-plane resolution of 220 mm, allowing direct visual identification of subfields. Results: Between diagnostic groups, the Ab42:tau ratio varied significantly (one-way ANOVA, p1⁄40.0002). While Ab42 concentrations were not different between groups, tau concentrations were (p1⁄40.004). Across pooled subjects, CA1-SRLM width correlated with CSF concentrations of total tau (r1⁄4-0.48; p1⁄40.006) and phospho-tau (r1⁄4-0.41; p1⁄40.036). CA1 stratum pyramidale and entorhinal cortex, other sites of tau pathology, showed similar correlations. By contrast, the dentate gyrus and CA3 area did not correlate with CSF total tau or phospho-tau. Importantly, CSF Ab42 concentrations did not correlate with any medial temporal lobe structural metric. Finally, among these subjects, episodic memory performance correlated with CA1-SRLM width (r1⁄40.58, p1⁄40.003), CSF total tau (r1⁄4-0.51, p1⁄40.013), and CSF phospho-tau (r1⁄4-0.53, p1⁄40.007), but not CSF Ab42. Conclusions: CSF total tau and phospho-tau levels reflect atrophy in CA1-SRLM and other specific medial temporal lobe structures known to be selectively vulnerable to tau pathology. These data raise the possibility that CSF tau levels may go beyond representing nonspecific effects of neurodegeneration and may reflect the actual focal burden of tau pathology.