Interaction of CSF α-synuclein and amyloid beta in cognition and cortical atrophy

Abstract

Introduction: Lewy body-related pathology is commonly observed at autopsy in individuals with dementia, but in vivo biomarkers for α-synucleinopathy are lacking.

Methods: Baseline cerebrospinal fluid (CSF) biomarkers, polygenic risk score (PRS) for Parkinson's disease (PRS-PD) and Alzheimer's disease (PRS-AD), longitudinal cognitive scores, and magnetic resonance imaging were measured in 217 participants from the Alzheimer's Disease Neuroimaging Initiative. Linear mixed models were used to find the relationship of CSF biomarkers and the PRS with cognition and cortical atrophy.

Results: Higher PRS-PD and PRS-AD were associated with lower CSF α-synuclein and amyloid beta (Aβ), respectively. Lower CSF α-synuclein and the interaction of CSF α-synuclein and Aβ were associated with lower cognitive scores and global cortical atrophy most prominently in the occipital cortex.

Discussion: Lower CSF α-synuclein could be a biomarker for α-synucleinopathy, and the simultaneous evaluation of CSF biomarkers for AD and CSF α-synuclein could reveal the independent and interactive effects on cognition and cortical atrophy.

Keywords: Alzheimer's disease; Lewy body disease; cerebrospinal fluid biomarkers; polygenic risk score; α‐synuclein.

FIGURE 1

Flowchart of the study participants. Aβ, amyloid beta; AD, Alzheimer's disease; ADNI, Alzheimer's Disease Neuroimaging Initiative; CSF, cerebrospinal fluid; PD, Parkinson's disease; PRS, polygenic risk score

FIGURE 2

Association between cerebrospinal fluid (CSF) biomarkers and polygenic risk scores. Data are results of general linear models for CSF α‐synuclein (A and B), CSF β‐amyloid (C and D), and CSF tau (E and F) using polygenic risk score‐Alzheimer's disease (PRS‐AD), polygenic risk score‐Parkinson's disease (PRS‐PD), and other CSF biomarkers as predictors. Covariates included age, sex, and education. Each dot represents individuals and predicted regression lines (black line) with confidence intervals (gray shade) are displayed

FIGURE 3

Effect of CSF biomarkers and the genetic risk score on regional cortical thickness. A, Results are based on a linear mixed model for voxel‐wise cortical thickness using baseline CSF α‐synuclein, CSF Aβ1‐42, CSF tau, PRS‐AD, PRS‐PD, α‐synuclein × Aβ1‐42, and tau × year as predictors after controlling for age, sex, education, and intracranial volume. B,C, Results for subgroups differing by Aβ positivity are based on a linear mixed model without CSF Aβ1‐42 and α‐synuclein × Aβ1‐42. Positive correlations were tested for Aβ1‐42 and PRS‐PD, while negative correlations were tested for α‐synuclein, tau, and PRS‐AD. The RFT was used to correct for multiple comparisons over the whole cortical mantle. In the p‐maps, blue color indicates the brain regions where the RFT‐corrected cluster P‐values were significant; yellow to red indicates the brain regions where the RFT‐corrected vertex P‐values were significant. Brain images are displayed in the neurologic convention. Aβ, amyloid beta; AD, Alzheimer's disease; CSF, cerebrospinal fluid; PD, Parkinson's disease; PRS, polygenic risk score; RFT, random field theory

FIGURE 4

Correlation of PRS‐PD with mean cingulate cortical thickness. Scatter plots of PRS‐PD against mean cingulate cortical thickness in separate subgroups according to (A) cognitive status and (B) Aβ positivity. Results are based on a mixed model showing correlation between mean cingulate cortical thickness and PRS‐PD in each subgroup. Age, sex, years of education, and intracranial volume were corrected. Gray dashed line indicates the relationship without the group separation as a reference. The cingulate region was defined by a surface parcellation map provided in the anatomical pipeline (green area). Aβ, β amyloid beta; MCI, mild cognitive impairment; NC, normal cognition; PD, Parkinson's disease; PRS, polygenic risk score