Premium
Multi‐tissue proteomic profiling for genetically defined Alzheimer disease cases
Author(s) -
Sung Yun J.,
Yang Chengran,
Suhy Adam J.,
Rhinn Herve,
Norton Joanne,
Wang Fengxian,
Bradley Joseph,
Farias Fabiana H.G.,
Benitez Bruno A.,
Harari Oscar,
Cruchaga Carlos
Publication year - 2021
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.054301
Subject(s) - trem2 , psen1 , disease , cerebrospinal fluid , biology , alzheimer's disease , proteomics , gene , medicine , computational biology , genetics , bioinformatics , immune system , pathology , neuroscience , presenilin , myeloid cells
Background Alzheimer disease (AD) is a complex and heterogeneous disease. Most AD cases present sporadically, while 1‐2% of the cases have autosomal dominant AD (ADAD), carrying mutations in APP , PSEN1 , and PSEN2 . Among sporadic cases, several rare variants in TREM2 , which is involved in immune response, are shown to increase AD risk. Most proteomic studies, while instrumental in identifying novel AD genes and pathways, focus on single tissues and mainly sporadic AD cases. Proteomic signatures for each genetically defined AD case will advance the understanding of the underlying biology of this heterogeneous disease and help us create prediction models for AD risk, onset, and progression. Method We generated deep proteomic profiles of brain (n=370), cerebrospinal fluid (CSF; n=699), and plasma (n=486) for sporadic AD, ADAD, and TREM2 risk‐variant carriers, from the Knight‐ADRC and DIAN cohorts. After stringent QC, 1079 proteins in brain, 713 in CSF, and 931 in plasma remained. Result We identified 150 proteins with differential levels between sporadic cases and controls, 371 for ADAD, and 134 for TREM2 carriers in at least one tissue (with Bonferroni‐corrected statistical significance). Of these, 36, 74 and 31 were replicated across tissue types for sporadic AD, ADAD, and TREM2 carriers, respectively. Using proteins that replicated across tissues, we created a prediction model for CSF and plasma with an AUC (AUC=0.79‐1.0) similar to (or better than) that of the well‐accepted CSF pTau/Aβ42 ratio. A model using 23 TREM2‐specific plasma proteins was able to discriminate between TREM2 carriers and controls (AUC=0.94) or sporadic AD cases (AUC=0.91) with high sensitivity and selectivity. Our pathway enrichment analysis validated pathways previously implicated in AD (such as APP, APOE, GSK3B, NOTCH3, PPP3R1 and PPP3CA) and identified novel pathways such as a Parkinson’s disease pathway (LRRK2 and α‐synuclein) and innate immune response pathways (MEK1, SHC1, and PDGF‐BB), including cytokine‐mediated signaling and a DAP12‐mediated pathway. Conclusion With the first multi‐tissue proteomic study of genetically defined AD cases, we have uncovered multiple novel AD biomarker candidates. These findings not only help to elucidate specific pathways implicated in AD but ultimately support the clinical utility of predictive models based on proteomic signatures.