Specific associations between plasma biomarkers and post-mortem amyloid plaque and neurofibrillary tau tangle loads

Several promising plasma biomarkers have recently been developed that could serve as diagnostic and/or prognostic tools for Alzheimers disease (AD). However, their neuropathological correlates have not yet been fully determined. Therefore, we aimed to investigate the independent associations between multiple plasma biomarkers (i.e., phosphorylated tau217 [p-tau217], p-tau181, p-tau231, the amyloid-{beta}42/40 [A{beta}42/40] ratio, glial fibrillary acidic protein [GFAP] and neurofilament light [NfL]) and core semi-quantitative measures of AD pathology (i.e., amyloid plaques and tau neurofibrillary tangles) as well as common co-pathologies (i.e., cerebral amyloid angiopathy, Lewy body disease, TAR DNA-binding protein 43, cerebral white matter rarefaction and argyrophilic grain disease). We included 105 participants from the Arizona Study of Aging and Neurodegenerative Disorders and Brain and Body Donation Program with antemortem collected plasma samples and a post-mortem neuropathological exam (mean(SD) time: 482(355) days), 48 of whom had longitudinal p-tau217 and p-tau181 (mean(SD) follow-up time: 1,378(1,357) days). Participants ranged from cognitively unimpaired to Alzheimers and non-Alzheimers dementia. All markers except NfL were associated with plaques (| {beta}|[&ge;]0.37, p<0.001) and tangles (| {beta}|[&ge;]0.27, p<0.008), in univariable analyses adjusted for age, sex and time between blood sampling and death. In multivariable models, when including both plaques and tangles as independent variables, the A{beta}42/40 ratio and p-tau231 were only associated with plaques ({beta}A{beta}42/40 [95%CI]=-0.59[-0.80,-0.38], R2plaques/R2=77.6%; {beta}p-tau231[95%CI]=0.32[0.09,0.56], R2plaques/R2=45.9%, all p[&le;]0.007), while GFAP was only associated with tangles ({beta}GFAP[95%CI]=0.39[0.19,0.59], p<0.001, R2tangles/R2=30.4%). In contrast, p-tau217 and p-tau181 were associated with both plaques ({beta}p-tau217[95%CI]=0.46[0.30,0.62], R2plaques/R2=40.4%; {beta}p-tau181[95%CI]=0.41[0.22,0.60], R2plaques/R2=35.7%, both p<0.001) and tangles ({beta}p-tau217[95%CI]=0.40[0.24,0.57], p<0.001, R2tangles/R2=30.7%; {beta}p-tau181[95%CI]=0.30[0.10,0.49], p=0.004, R2tangles/R2=17.1%). A parsimonious model predicting plaque load included p-tau217 and A{beta}42/40, while a parsimonious model for tangle burden included only p-tau217. Further, combining p-tau217 and A{beta}42/40 ratio yielded the highest accuracy for predicting intermediate/high AD neuropathological change ([ADNC], AUC[95%CI]=0.90[0.84,0.96],R2=0.66). High plasma NfL levels were predictive of presence of cerebral white matter rarefaction (AUC[95%CI]=0.76[0.66,0.85],R2=0.25). Finally, p-tau217 ({beta}[95%CI]=0.13[0.02,0.24], p=0.018), but not p-tau181 ({beta}[95%CI]=0.12[-0.05,0.29], p=0.152), levels increased more over time in participants with intermediate/high ADNC compared with those with none/low ADNC. In this relatively large neuropathological study with multiple plasma biomarkers available, we showed that the A{beta}42/40 ratio and p-tau231 were specific markers of plaque pathology, and GFAP of tangle pathology, while p-tau181 and, especially, p-tau217 were markers of both plaque and tangle pathologies. Our results suggest that high-performing assays of plasma p-tau217 and A{beta}42/40 might be an optimal biomarker combination to detect ADNC in vivo.