Identifying the role of transcriptome in amyloid‐β and tau pathologies in Alzheimer’s disease
Meichen Yu, Shannon L Risacher, Kwangsik Nho, Andrew J. Saykin,- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Amyloid‐β and tau proteins accumulate at distinct neuronal systems in Alzheimer’s disease (AD). However, why specific brain regions are more vulnerable to amyloid‐β and tau pathologies than others remain unclear.
Method
Using two independent cohorts from the ADNI and the IADRC as discovery and replication samples, we studied the relationship between gene expression profiles and amyloid‐β and tau pathologies. In ADNI, we included 583 and 564 participants with amyloid‐β PET and tau PET data, respectively; in IADRC, we included 110 and 78 participants with amyloid‐β PET and tau PET data, respectively. In both cohorts, amyloid‐β pathology was measured by [18F]florbetapir (AV45) and [18F]florbetaben (FBB) PET; tau pathology was measured by [18F]flortaucipir (AV‐1451) PET. The regional vulnerability to amyloid‐β and tau pathologies were estimated by contrasting amyloid‐β and tau PET data between cognitively unimpaired (CU and/or SCD) and patients (MCI and AD). Regional expression profiles for 20,736 protein‐coding genes were derived from brain‐wide microarray‐based transcriptome data from the Allen Human Brain Atlas. We computed spatial associations between gene expression levels and amyloid‐β (gene‐to‐amyloid‐β associations) and tau (gene‐to‐tau associations) pathologies for 60 AD susceptibility genes selected from recent large‐scale GWAS studies.
Result
In both cohorts, AD and MCI showed the expected higher amyloid‐β loads than CU, particularly in the medial parietal cortex, temporal lobe, medial and inferior prefrontal cortices, and superior and middle frontal cortices. In both cohorts, AD and MCI showed higher tau deposits than CU, mainly in the medial temporal lobe, medial and inferior parietal cortices, and inferior and middle temporal cortices. In both cohorts, we identified consistent significant gene‐to‐amyloid‐β and gene‐to‐tau associations after multiple testing adjustment (PFDR < 0.05). Specifically, CNTNAP2 and CLU showed the strongest positive gene‐to‐amyloid‐β associations, whereas TNIP1 and INPP5D showed the strongest negative associations. APOE and BIN1 showed the strongest positive gene‐to‐tau associations, whereas SORL1 showed the strongest negative associations.
Conclusion
We identified replicable regional vulnerability patterns of amyloid‐β and tau pathologies and their relationships to regional gene expression profiles. Better understanding of the regional association of gene expression and amyloid‐β and tau pathologies may foster identification of novel diagnostic and therapeutic targets.