NEW YORK (GenomeWeb) – A new transcriptome and expression quantitative trait loci-based analysis has narrowed in on gene variants that cause age-related macular degeneration (AMD).
One of the leading causes of blindness in people over the age of 65, AMD affects about 2 million people in the US, according to the National Eye Institute. The condition damages the macula and leads to the loss of central vision.
Following on a genome-wide association study that implicated 52 SNPs at 34 genetic loci in AMD, NEI researchers conducted a cis-eQTL study, eCAVIAR analysis, and transcriptome-wide association study to further home in on causal genes. As they reported today in Nature Genetics, they uncovered a number of likely target genes, but noted that B3GLCT and BLOC1S1 were the most plausible target genes, as each of three approaches they applied implicated them.
"If we were conducting a criminal investigation, prior research would have localized different crime syndicates to 52 streets within 34 zip codes. These latest findings identify actual suspects — direct targets that we can more closely investigate," senior author Anand Swaroop, the chief of the Neurobiology-Neurodegeneration and Repair Laboratory at NEI, said in a statement.
He and his colleagues first developed what they've dubbed the EyeGEx, a database of retinal gene expression akin to the Genotype-Tissue Expression (GTEx) Project database. To build it, they sequenced RNA from 523 retinas from 517 postmortem donors, with varying stages of AMD or without AMD. They identified 13,662 protein-coding genes and 1,462 non-coding genes, and annotation analysis indicated the transcripts were enriched for genes involved in visual perception, metabolic processes, and energy homeostasis.
They also mapped cis-eQTLs and uncovered nearly 15,000 genetic variants that influenced the expression of more than 10,000 genes. When they folded in results from previous AMD GWAS, the researchers found that nine lead SNPs loci highlighted by those GWAS were significant eQTLs in the retina for 19 SNP-gene associations.
Using eCAVIAR, a statistical tool that estimates that likelihood of a variant to be causal, the researchers pinpointed the most likely causal variants at six AMD loci.
The lead GWAS signal at two loci — B3GALTL and RDH5/CD63 — were considered the most likely causal SNPs, though at four other loci — SLC16A8, ACAD10, TMEM97/VTN, and APOE — the likely causal variant differed from the lead SNP.
The researchers also used their eQTL data and recent GWAS data to conduct a TWAS of their transcriptome data to search for novel risk genes. This analysis uncovered 61 transcriptome-wide significant gene-AMD associations. They uncovered 38 genes located near 13 AMD-GWAS loci, 28 of which passed correction testing. In addition, they found 23 genes located outside those GWAS loci within 16 regions. Three genes — RLBP1, PARP12, and HIC1 — were the only significant genes in the region, leading the researchers to consider them strong new candidate AMD-associated genes.
By combining their eQTL, eCAVIAR, and TWAS data, the researchers identified what they considered to be the most plausible target genes: B3GLCT and BLOC1S1, which were both identified through all three approaches.
B3GLCT, the researchers noted, encodes a glucosyltransferase and, if lost, leads to Peters plus syndrome, a genetic condition marked by eye abnormalities. BLOC1S1, meanwhile, encodes a subunit of a multi-protein complex involved in synaptic function. Altered expression of these genes could contribute to AMD pathogenesis, they added, noting they could also represent treatment targets.