Locus coeruleus integrity as a proxy of initial tau burden: in vivo versus ex vivo observations

Abstract

AbstractBackgroundAutopsy studies reported the locus coeruleus (LC) as one of the initial regions in the brain harboring hyperphosphorylated tau pathology. MR signal intensity in the LC has been proposed as a measure of LC integrity, but the contribution of tau pathology, commonly found in LC, to this signal intensity is not currently measurable in vivo. We related in‐vivo LC MR signal intensity to Alzheimer’s disease (AD) phenotype and PET measures and used ROSMAP autopsy data to guide interpretation of the MR LC measure as tau‐related or cellular integrity‐related.MethodOne‐hundred seventy‐two participants from the Harvard Aging Brain Study (HABS) underwent 3T LC‐MRI imaging, FTP and PiB‐PET imaging (median age: 75 years, 58% female, n=22 with CDR=0.5/1). In 160 individuals from ROSMAP, LC tangle density and neuronal density were quantified at autopsy (median age at death: 89 years, 67% female, 66 control, 53 mild cognitively impaired, 41 AD patients; Figure 1). Groups were compared using ANCOVA. LC measures were related to Mini‐Mental State Examination (MMSE) scores using Spearman correlations (age‐adjusted). In HABS, LC integrity was vertex‐wise regressed on tau‐PET. In ROSMAP, Spearman correlations related LC tangle or neuronal density to cortical tangle density or Braak staging.ResultLC measures correlated with measures of disease progression: in HABS, LC integrity was lower in impaired individuals compared to controls, and correlated positively with MMSE score. In ROSMAP, LC tangle density was higher in MCI/AD patients compared to controls; and correlated negatively with MMSE scores. No associations were found for LC neuronal density (Figure 2). Furthermore, lower LC integrity correlated with higher tau in medial‐lateral temporal regions in HABS, mirroring the consecutive Braak‐staging. Similarly, In ROSMAP, higher LC tangle density, but not LC neuronal density, correlated positively with cortical tangle burden and Braak staging (Figure 3).ConclusionThe comparable associations of MRI‐LC integrity or LC tangle density with disease progression and tau suggest that MRI‐measures of LC integrity may be a proxy for tau pathology, rather than neuronal density. Future verification using head‐to‐head comparison is warranted, but our results indicate that LC integrity holds promise as marker for detection of initial AD‐related processes.