MRI Signature of α-Synuclein Pathology in Asymptomatic Stages and a Memory Clinic Population

Abstract

Importance: The lack of an in vivo measure for α-synuclein (α-syn) pathology until recently has limited thorough characterization of its brain atrophy pattern, especially during early disease stages.

Objective: To assess the association of state-of-the-art cerebrospinal fluid (CSF) seed amplification assays (SAA) α-syn positivity (SAA α-syn+) with magnetic resonance imaging (MRI) structural measures, across the continuum from clinically unimpaired (CU) to cognitively impaired (CI) individuals, in 3 independent cohorts, and separately in CU and CI individuals, the latter reflecting a memory clinic population.

Design, setting, and participants: Cross-sectional data were used from the Swedish BioFINDER-2 study (inclusion, 2017-2023) as the discovery cohort and the Swedish BioFINDER-1 study (inclusion, 2007-2015) and Alzheimer's Disease Neuroimaging Initiative (ADNI; inclusion 2005-2022) as replication cohorts. All cohorts are from multicenter studies, but the BioFINDER cohorts used 1 MRI scanner. CU and CI individuals fulfilling inclusion criteria and without missing data points in relevant metrics were included in the study. All analyses were performed from 2023 to 2024.

Exposures: Presence of α-syn pathology, estimated by baseline CSF SAA α-syn.

Main outcomes and measures: The primary outcomes were cross-sectional structural MRI measures either through voxel-based morphometry (VBM) or regions of interest (ROI) including an automated pipeline for cholinergic basal forebrain nuclei CH4/4p (nucleus basalis of Meynert [NBM]) and CH1/2/3. Secondary outcomes were domain-specific cross-sectional cognitive measures. Analyses were adjusted for CSF biomarkers of Alzheimer pathology.

Results: A total of 2961 participants were included in this study: 1388 (mean [SD] age, 71 [10] years; 702 female [51%]) from the BioFINDER-2 study, 752 (mean [SD] age, 72 [6] years; 406 female [54%]) from the BioFINDER-1 study, and 821 (mean [SD] age, 75 [8] years; 449 male [55%]) from ADNI. In the BioFINDER-2 study, VBM analyses in the whole cohort revealed a specific association between SAA α-syn+ and the cholinergic NBM, even when adjusting for Alzheimer copathology. ROI-based analyses in the BioFINDER-2 study focused on regions involved in the cholinergic system and confirmed that SAA α-syn+ was indeed independently associated with smaller NBM (β = -0.271; 95% CI, -0.399 to -0.142; P <.001) and CH1/2/3 volumes (β = -0.227; 95% CI, -0.377 to -0.076; P =.02). SAA α-syn+ was also independently associated with smaller NBM volumes in the separate CU (β = -0.360; 95% CI, -0.603 to -0.117; P =.03) and CI (β = -0.251; 95% CI, -0.408 to -0.095; P =.02) groups. Overall, the association between SAA α-syn+ and NBM volume was replicated in the BioFINDER-1 study and ADNI cohort. In CI individuals, NBM volumes partially mediated the association of SAA α-syn+ with attention/executive impairments in all cohorts (BioFINDER-2, β = -0.017; proportion-mediated effect, 7%; P =.04; BioFINDER-1, β = -0.096; proportion-mediated effect, 19%; P =.04; ADNI, β = -0.061; proportion-mediated effect, 20%; P =.007).

Conclusions and relevance: In this cohort study, SAA α-syn+ was consistently associated with NBM atrophy already during asymptomatic stages. Further, in memory clinic CI populations, SAA α-syn+ was associated with NBM atrophy, which partially mediated α-syn-induced attention/executive impairment.

Figure 1.. Voxel-Based Morphometry (VBM) and Region of Interest (ROI)–Based Analyses in the Whole Cohort Associating Seed Amplification Assay (SAA) α-Synuclein Positivity (α-syn+) With Magnetic Resonance Imaging (MRI) Structural Measures

A, VBM analyses in the whole cohort (cognitively unimpaired and cognitively impaired) of the BioFINDER-2 study. VBM results are shown in orange (P < .05, after false discovery rate correction). Regions in purple and blue depict the atlas-based basal forebrain nuclei. B-D, ROI-based analyses of the association of cerebrospinal fluid SAA α-syn+ with MRI structural measures in the BioFINDER-2 study (B) and replicated in the BioFINDER-1 study (C) and the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (D). The violin plot representing the data distribution is overlaid on the box plot showing the IQR and the median of the distribution. The red line indicates a significant difference between SAA α-syn negative (neg) and positive (pos) individuals (P < .05) after Bonferroni correction over all 8 regions. CH1/2/3 reflects the nuclei in the basal forebrain, not including the NBM; NBM, nucleus basalis of Meynert.

Figure 2.. Region of Interest (ROI)–Based Analyses in Cognitively Unimpaired (CU) and Cognitively Impaired (CI) Associating Seed Amplification Assay (SAA) α-Synuclein Positivity (α-syn+) With Magnetic Resonance Imaging Structural Measures

A-C, ROI-based analyses in CU individuals. D-F, ROI-based analyses in CI individuals. Main analyses are performed in the BioFINDER-2 study and replicated in the BioFINDER-1 and Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohorts. The violin plot representing the data distribution is overlaid on the box plot showing the IQR and the median of the distribution. The red line indicates a significant difference between SAA α-syn negative (neg) and positive (pos) individuals (P < .05) after Bonferroni correction over all 8 regions. CH1/2/3 reflects the nuclei in the basal forebrain, not including the NBM; NBM, nucleus basalis of Meynert.

Figure 3.. Association of Seed Amplification Assay (SAA) α-Synuclein Positivity (α-syn+) With Attention/Executive Functioning Partially Mediated by Nucleus Basalis of Meynert (NBM) Volumes in Cognitively Impaired Individuals

The symbol digit modalities test is used in BioFINDER-1 and BioFINDER-2 cohorts, and the harmonized composite executive function score is used in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort. ^aMediation models are shown for the BioFINDER-1, BioFINDER-2, and ADNI cohorts.