HAI Book 2025 - Flipbook - Page 335
Klinger, Hannah
70
Whole-blood gene expression associations with Aβ-PET
accumulation and cognitive decline
Hannah M Klinger1, Vaibhav A Janve2, Mabel Seto1,3, Jane A Brown1, Colin Birkenbihl1, Gillian
Coughlan1, Diana L Townsend1, Jane Zyski2, Ting-Chen Wang2, Rebecca E Amariglio3, Kathryn
V Papp3, Dorene M Rentz3, Hyun-Sik Yang3, Jasmeer Chhatwal1, Michelle Clifton2, Michael C
Donohue4, Rema Raman4, Robert A Rissman4, Paul Aisen4, Keith A Johnson1, Reisa Sperling1,3,
Logan Dumitrescu2, Timothy Hohman2, Rachel Buckley1
1
Massachusetts General Hospital/Harvard Medical School, Boston, MA, US
Vanderbilt Memory & Alzheimer’s Center, Vanderbilt University Medical Center, Nashville, TN, US
3
Brigham and Women’s Hospital, Boston, MA, US
4
Alzheimer’s Therapeutic Research Institute, University of Southern California, San Diego, CA, US
2
Objective: Bulk brain-tissue studies of gene expression have discovered novel biological pathways associated
with AD neuropathology and pre-morbid cognitive decline. However, these findings are primarily tuned to much
older adults and are focused on end-of-life changes. The objective of this study was to ascertain transcriptomic
signals from whole-blood associated with A´-PET accumulation and cognitive decline in clinically unimpaired
older adults from the A4/LEARN studies. Identifying genes in whole-blood that are associated with these in vivo
AD phenotypes can elucidate biological pathways implicated in these earliest disease processes.
Methods: We selected 1,737 participants (Nlongitudinal=770; 71years(±4.7); 63% Female; 35% APOEε4; 26% A´+) from
the A4 clinical trial placebo and treatment arms (and screen fail at baseline) and the adjoining LEARN
observational study who had A´-PET (18F-Florbetapir), cognitive assessments (using PACC), and whole-blood gene
expression data from the autosome and X chromosome (20,621 genes) available. Linear mixed-effects models
determined the association between gene expression alone and the interactions with sex, APOEε4, and A´continuous
on longitudinal A´-PET and PACC. All analyses included random intercepts/slopes, and is adjusted for age,
education level, cohort, and story version. Analyses were FDR corrected.
Results: A SAT1-DT*APOEε4 interaction, from a long non-coding (lnc) RNA, was found to weakly associate with
A´-PET accumulation (Fig.1A & Fig.2A; b=-0.008(0.002),pFDR=0.024). For PACC decline, one gene (TRBV4-2) was
moderated by APOEε4*sex, 21 genes by A´, and 128 genes by A´*sex (Fig.1B-D). Select analyses revealed that
lower ETF1P2 expression, another lncRNA, and higher baseline A´ was associated with faster cognitive decline
(Fig.2B; bINT=2.38(0.5),pFDR=0.003). Lower ZSCAN2, implicated in regulating inflammation, resulted in similar
patterns, but only in males (bINT=-7.08(1.3),pFDR=0.001; Fig.2C).
Conclusion: Whole-blood transcriptomic signals were not robustly associated with A´-PET accumulation but
were associated with cognitive decline via interactions with A´-PET and sex. Further work and validation in
external cohorts will elucidate relevant biological pathways.
HAI2025 - 335
