Global neuropathologic severity of Alzheimer’s disease and locus coeruleus vulnerability influences plasma phosphorylated tau levels

Advances in ultrasensitive detection of phosphorylated tau (p-tau) in plasma has enabled the use of blood tests to measure Alzheimer’s disease (AD) biomarker changes. Examination of postmortem brains of participants with antemortem plasma p-tau levels remains critical to understanding comorbid and AD-specific contribution to these biomarker changes. We analyzed 35 population-based Mayo Clinic Study of Aging participants with plasma p-tau at threonine181 and threonine217 (p-tau181, p-tau217) available within 3 years of death. Autopsied participants included cognitively unimpaired, mild cognitive impairment, AD dementia, and non-AD neurodegenerative disorders. Global neuropathologic scales of tau, amyloid-β, TDP-43, and cerebrovascular disease were examined. Regional digital pathology measures of tau (phosphorylated threonine181 and 217 [pT181, pT217]) and amyloid-β (6F/3D) were quantified in hippocampus and parietal cortex. Neurotransmitter hubs reported to influence development of tangles (nucleus basalis of Meynert) and amyloid-β plaques (locus coeruleus) were evaluated. The strongest regional associations were with parietal cortex for tau burden (p-tau181 R=0.55, p=0.003; p-tau217 R=0.66, p<0.001) and amyloid-β burden (p-tau181 R=0.59, p<0.001; p-tau217 R=0.71, p<0.001). Linear regression analysis of global neuropathologic scales explained 31% of variability in plasma p-tau181 (R 2 =0.31) and 59% in plasma p-tau217 (R 2 =0.59). Neither TDP-43 nor cerebrovascular disease global scales independently contributed to variability. Global scales of tau pathology (β-coefficient=0.060, p=0.016) and amyloid-β pathology (β-coefficient=0.080, p<0.001) independently predicted plasma p-tau217 when modeled together with co-pathologies, but only amyloid-β (β-coefficient=0.33, p=0.021) significantly predicted plasma p-tau181. While nucleus basalis of Meynert neuron count/mm 2 was not associated with plasma p-tau levels, a lower locus coeruleus neuron count/mm 2 was associated with higher plasma p-tau181 (R=-0.50, p=0.007) and higher plasma p-tau217 (R=-0.55, p=0.002). Cognitive scores (R 2 =0.31-0.34) were predicted by the global tau scale, but not by the global amyloid-β scale or plasma p-tau when modeled simultaneously. Higher soluble plasma p-tau levels may be the result of an intersection between insoluble deposits of amyloid-β and tau accumulation in brain and may be associated with locus coeruleus degeneration.