This study mapped plasma proteomic correlates of cerebral amyloid burden using quantitative amyloid PET integrated with large-scale plasma proteomics (~7,000 proteins; SomaScan v4.1) in 1,429 participants across two independent cohorts (Knight-ADRC discovery, n=558; Bio-Hermes validation, n=871), focusing on molecular heterogeneity within amyloid-beta-positive (Aβ+) individuals.
454 plasma proteins were nominally associated with amyloid load in Aβ+ individuals; 54 replicated cross-cohort with concordant effect directions. A weighted 54-protein proteomic score correlated with higher amyloid burden, worse CDR-SB (r=0.19, p=0.01), A+T+ biomarker status (p=0.03), and AD diagnosis (p=0.02) within Aβ+ individuals. Five protein clusters mapped onto distinct clinical trajectories including early-onset aggressive disease, slower-progressing early-onset, putative resilience, rapid late-onset decline, and systemic neurodegeneration.
- Cross-sectional design precludes causal inference and limits ability to track dynamic proteomic changes over time. - Aβ+ subsets were relatively small (n=178 Knight-ADRC; n=241 Bio-Hermes), reducing power for stringent multiple-testing correction. - Cohorts were predominantly of European ancestry recruited at specialized centers, limiting generalizability to diverse or community-based populations.
A 54-protein plasma proteomic score can stratify disease severity and molecular subtype within Aβ+ individuals, complementing p-tau217 and amyloid PET without requiring invasive or costly imaging. Clinicians and trialists may eventually use such panels to identify high-risk subgroups or monitor treatment response, though longitudinal validation in diverse cohorts is still needed.
Explore related topics
