3D chromatin structure and massively parallel reporter assays in hiPSC‐derived microglia and human macrophages identifies novel putative Alzheimer’s Disease risk genes
Ivana Yoseli Quiroga, Marielle L Bond, Susan D'Costa, Jessica L Bell, Jessica C McAfee, Kathleen S Metz Reed, Isha R Sahasrabudhe, Mary Patrucco, Hyejung Won, Douglas H Phanstiel- Psychiatry and Mental health
- Cellular and Molecular Neuroscience
- Geriatrics and Gerontology
- Neurology (clinical)
- Developmental Neuroscience
- Health Policy
- Epidemiology
Abstract
Background
Alzheimer’s disease (AD) genome‐wide association studies (GWAS) identified 75 disease‐associated loci; however, interpretation of these results has proven difficult for two primary reasons. First, linkage disequilibrium between nearby variants makes it difficult to identify the causal variant(s) at each locus. Second, because the vast majority of OA risk variants are non‐coding and can regulate genes from distances of greater than one million base pairs, the genes impacted by AD risk variants are largely unknown. Recently developed genomic approaches can address these hurdles but must be applied to the correct cell types and biological contexts.
Method
We quantified the impact of 3,576 AD‐associated genetic variants on enhancer activity by performing massively parallel reporter assays (MPRAs) in resting and activated human macrophages. To map these variants to the genes they regulate we built regulatory networks by mapping 3D chromatin structure (Hi‐C), chromatin accessibility (ATAC), histone K27 acetylation (CUT&RUN), and gene expression (RNA‐seq) in both resting and activated iPSC‐derived microglia.
Result
In total we identified 20,992 chromatin loops across resting and activated iPSC‐derived microglia. Among these, 9,544 loops linked 14,576 enhancers to 11,487 genes. 58 loops connected 414 AD‐risk variants queried by our MPRA assays to 85 genes. 212, 19,256, and 2,790 of these loops, enhancers, and genes (respectively) changed in response to activation, providing further insight into their regulatory mechanisms and highlighting the need to study these events in the correct biological context.
Conclusion
Ongoing analyses and data integration should reveal further mechanistic details and provide novel AD risk genes for further research and therapeutic development. By intersecting our MPRA results and dynamic regulatory networks with AD GWAS, we were able to identify novel putative genes regulated by non‐coding AD risk variants in a cell‐type and condition‐specific manner.