Discussion
We observed opposite level changes between insoluble and soluble FLNA in prodromal AD. As this stage coincides with the appearance of cognitive symptoms, this may be a key event in the transition from preclinical to prodromal AD. Insoluble FLNA could be useful to identify prodromal AD among subjects with an MCI, indicating that it might be a hallmark of prodromal AD.
Methods
From parietal cortex samples of 57 individuals (19 with no cognitive impairment (NCI), 19 mild cognitively impaired (MCI) and 19 with dementia) from the Religious Orders Study (ROS), we quantified total tau, phosphorylated tau (pTau), FLNA, synaptophysin, vesicular acetylcholine transporters (VAChT) and choline acetyltransferase (ChAT) by Western blot. Aβ42 and neuritic plaques (NP) were quantified by ELISA and Bielschowsky silver impregnation, respectively. AD staging was determined using ABC method combining Thal, Braak and the CERAD staging. From this, clinicopathologic stages of AD were established by subdividing subjects with neuropathological AD between preclinical AD, prodromal AD and AD dementia (ADD). Receiver operating characteristics analyses were performed to predict AD neuropathology from FLNA quantifications.
Results
Insoluble FLNA was significantly and positively correlated with Aβ42, NP, Thal stages, ABC scores and AD clinicopathologic stages (p < 0.05 False discovery rate-corrected). No correlation of FLNA with tau measures was found. Insoluble FLNA levels were significantly higher in the prodromal AD, ADD and intermediate ABC groups. This was consistent with significantly lower levels of soluble FLNA specifically in prodromal AD. Insoluble (AUC: 0.830) and soluble FLNA levels (AUC: 0.830) as well as the ratio of soluble over insoluble FLNA (AUC: 0.852), were excellent predictors of prodromal AD among subjects with MCI from the ROS cohort.