The relationship between amyloid-β peptide spectrum and the spastic paraparesis phenotype in autosomal dominant Alzheimer's disease.

backgroundMore than 300 mutations in presenilin 1 (PSEN1) lead to autosomal dominant Alzheimer's disease (ADAD). PSEN1, as the catalytic subunit of γ-secretase, generates amyloid-β (Aβ) peptides through a sequential proteolysis of the amyloid precursor protein (APP). While ADAD typically presents with progressive cognitive decline, ~ 25% of PSEN1 mutation carriers develop spastic paraparesis (SP), a debilitating motor condition. The molecular basis of this phenotypic heterogeneity remains unknown. This study examines Aβ profiles generated by PSEN1 variants associated with different clinical presentations with the aim of exploring potential associations between different Aβ profiles and clinical heterogeneity.

methodsWe analysed reported Aβ peptide profiles generated in vitro by 160 PSEN1 variants, categorized by their associated AD or AD + SP phenotype. We employed an integrated analytical approach combining univariate comparisons of Aβ profiles with machine learning classification.

resultsAD + SP-linked mutations showed significantly higher Aβ43 levels and more severe impairments in γ-secretase processivity compared to pure dementia associated variants. Machine learning consistently identified Aβ43 as the most important feature allowing for the phenotypic classification. Unlike processivity impairments, total Aβ production was comparable between groups, suggesting specific rather than global alterations in γ-secretase function.

conclusionsOur analysis reveals a robust association between elevated Aβ43 levels and SP development in PSEN1 mutation carriers. While this correlation does not establish causation, the distinct impact of SP-associated mutations on γ-secretase function, resulting in elevated Aβ43 production, suggests that mutation-specific mechanisms may underlie clinical heterogeneity in ADAD, with potential implications for biomarker and translational research.