Predicting development of AD clinical symptoms and their progression through a collection of novel plasma Aβ immunoassays

Abstract

AbstractBackgroundFor many years, fluid biomarkers for amyloid β‐protein were described exclusively for Aβ42 and Aβ40, with Aβ40 being the major secreted form and Aβ42 being the predominant component of amyloid‐plaques in AD brain. However, the C‐termini of these two Aβ peptides and the N‐terminus of the AICD are not aligned. It is now clear that secreted Aβ peptides are produced along two proteolytic pathways by γ‐secretase: Aβ49→46→43→40→37 and Aβ48→45→ 42→39/38. We now asked whether a more comprehensive analysis of Aβ peptides in plasma and CSF will help understand and predict brain amyloid pathology and its consequences.MethodWe designed novel immunoassays to detect essentially all the secreted forms of Aβ (Aβ x‐37, 38, 39, 40, 42 and 43), initially on the MesoScale‐Discovery platform (sensitive enough for human CSF samples). To measure same analytes in plasma, we built ultrasensitive immunoassays on the Quanterix SP‐X planar platform with 10‐20 folds enhanced sensitivity. We term Aβ37‐39 short and 40‐43 long Aβs.ResultWe quantified 6 Aβ‐variants in 79 CSF samples from Mayo‐Clinic divided into three diagnostic groups: cognitively normal (CN), mild cognitive impairment (MCI) and Alzheimer's disease (AD). In contrast to Aβ42/40, the short/long Aβ ratios in CSF distinguished CN from AD & MCI more effectively, with AUC (ROC‐analysis) of 0.9609 for Aβ37/42 in distinguishing CN and AD, while AUC is 0.8623 for Aβ42/40. Moreover, both 37/42 and 38/43 correlated significantly with the MMSE scores of these subjects. Next, we tested 88 plasma samples from same cohort, the Aβx‐37/1‐42 ratio in plasma clearly distinguished AD and CN (AUC of 0.8572), while the traditional Aβ1‐42 analyte could not. Currently, we are testing the Aβ37/42 and other ratios in a longitudinal Harvard‐Aging‐Brain‐Study (HABS) cohort.ConclusionCompared to the classic Aβ42/40 ratio in CSF or the Aβ1‐42 absolute level in plasma, a comprehensive analysis of Aβ peptides in plasma and CSF yields better Aβ biomarkers to distinguish CN and AD. Moreover, both 37/42 and 38/43 ratios in CSF correlated significantly with MMSE scores. Importantly, we are assessing whether plasma Aβ37/42 correlates most tightly with the development of AD clinical symptoms and their progression, using longitudinal aging cohorts.