Purpose
To enhance our understanding of age related macular degeneration (AMD) genome wide association study (GWAS) results by using signals from studies of expression quantitative trait loci (eQTL). In particular, eQTL results can help narrow down the causal variants and point to specific genes which may be dysregulated by the variants.
Methods
We analyzed publicly available RNA-seq and genotype data (Lappalainen et al., Nature, 2013; Abecasis et al., Nature, 2012) from 335 lymphoblastoid cell lines (LCLs) to look for eQTLs which affect the genes located in the vicinity of 19 loci from a prior GWAS. To detect eQTLs, we used weighted linear regression, adjusting for the first 5 principal components of the genetic covariance and the first 10 principal components from the expression covariance. We focus only on variants which overlapped between the eQTL data and the AMD GWAS.
Results
On chromosome 14, eQTL variants for the gene ZFP36L1 are shared with AMD GWAS hits, suggesting that dysregulation of ZFP36L1 may be involved in AMD. ZFP36L1 is an RNA binding protein from the tristetrapolin family, a set of zinc-finger proteins that bind to the 3’ ends of RNA and mediate programmed decay. Specifically, ZFP36L1 is responsible for VEGF RNA stability. On chromosome 19, variants for an eQTL of the gene GPR108 are shared with a GWAS hit for AMD, suggesting that GPR108 may be important. Chromosome 6 (Figure 1) contains a very interesting regulatory region with eQTLs that control expression of at least 3 different genes in the same area as the GWAS, supporting recent papers that suggest more than one gene/AMD signal in this region. In addition, the genes located near the two strongest GWAS signals (ARMS2/HTRA and CFH) were not expressed in LCLs at quantifiable levels. Because only 13 of the 22 genes near GWAS signals were expressed at quantifiable levels, we also examined recent RNA-seq from publicly available retinal tissue. We determined that 20 of the 22 genes are expressed in the retina. RNA-seq of a native fetal RPE cell line from our lab suggests that 16 of these genes are also expressed in fetal RPE primary cells (e.g., CFH).
Conclusions
This work highlights the usefulness of eQTL data to enhance our interpretation of GWAS results in terms of biological mechanisms, but it also strongly suggests that eQTLs from studies of ocular tissues would be far more enlightening in understanding disease pathogenesis.
Keywords: 412 age-related macular degeneration •
533 gene/expression •
534 gene mapping