Pathway-specific polygenic risk score of AD-associated genetic variants associated with AD risk, resilience against AD, and progression to AD

Journal Article (2021)
Author(s)

N. Tesi (Amsterdam UMC)

S.J. van der Lee (TU Delft - Pattern Recognition and Bioinformatics)

Marc Hulsman (TU Delft - Pattern Recognition and Bioinformatics)

Iris E. Jansen (Amsterdam UMC)

N. Stringa (TU Delft - Pattern Recognition and Bioinformatics)

Natasja M. van Schoor (TU Delft - Pattern Recognition and Bioinformatics)

Martijn Huisman (Amsterdam UMC)

Philip Scheltens (Amsterdam UMC)

Marcel Reinders (TU Delft - Pattern Recognition and Bioinformatics)

Wiesje M. van der Flier (Amsterdam UMC)

Henne Holstege (Amsterdam UMC, TU Delft - Intelligent Systems)

Research Group
Pattern Recognition and Bioinformatics
DOI related publication
https://doi.org/10.1002/alz.053500
More Info
expand_more
Publication Year
2021
Language
English
Research Group
Pattern Recognition and Bioinformatics
Volume number
17
Pages (from-to)
e053500

Abstract

BACKGROUND: Several collaborative genome-wide-association studies (GWAS) have characterized the genetic landscape of Alzheimer's disease (AD), which now counts >70 single-nucleotide polymorphisms (SNPs) associated with AD-risk. METHOD: We linked these SNPs to their affected biological pathways, and combined the effect of multiple SNPs into pathway-specific polygenic-risk-scores (PRS). Using a genetically homogeneous dataset of 2,458 AD-patients (age=70.2±10.4), 3,848 healthy controls (age=61.1±14.8), 343 cognitively healthy centenarians (age=101.1±1.8) and 705 individuals with mild-cognitive-impairment (MCI, age=69.1±8.9), we studied the association between pathway-PRS and AD-risk, MCI-to-AD conversion, and resilience against AD. RESULT: With an integrative strategy, we linked 72 AD-associated SNPs to 5 clusters of biological pathways: beta-amyloid cluster, immune cluster, vascular cluster, endocytosis cluster and a trafficking cluster (Figure 1). A PRS comprising all 72 SNPs (excluding APOE-SNPs) was significantly associated with AD-risk (OR=1.53, p<2e-16), progression of MCI to AD dementia (OR=1.45, p=5.5e-4), and, albeit in the opposite direction, resilience against AD in healthy centenarians (OR=0.82, p=6.4e-4) (Figure 2, Figure 3). At the pathway level, all pathway-PRS significantly associated with increased AD-risk while specifically the vascular-, endocytosis- and trafficking-PRS associated with resilience against AD (p<0.05) (Figure 4). Interestingly, the beta-amyloid-, endocytosis and trafficking-PRSs were significantly associated with MCI-to-AD progression (p<0.05), with a higher-PRS leading up to 1.5-year shorter conversion time. CONCLUSION: Our results add on the efficacy of pathway-specific PRSs to identify individuals at highest or the lowest risk for the development of AD. Further, our results suggest that specific biological pathways are more important at different stages of the disease or in fact, associate with the escape of disease. AD diagnostic procedures may benefit from the identification of individual, pathway-specific vulnerabilities. In the future, pathway-specific PRSs may contribute to the selection of patients who may benefit most from pathway-specific treatment strategies.

No files available

Metadata only record. There are no files for this record.