NEW YORK (GenomeWeb) – In a new study published online today in Nature Genetics, an international team led by investigators at Stanford University used genetic and phenotypic clues from a large cohort of US veterans to more fully spell out genetic contributors to blood lipid levels — insights that are expected to serve as a resource in efforts to find, treat, or prevent heart disease or diabetes.
"The idea is to use genetic data linked to electronic health records from a very large number of individuals to find genetic variants that simultaneously improve lipid profiles and protect against cardiovascular disease," corresponding and co-senior author Themistocles Assimes, a cardiovascular medicine researcher affiliated with Stanford University and the Veterans Affairs Palo Alto Health Care System, said in a statement. "From there, you can figure out what the best potential drug targets are."
The researchers started with a genome-wide association study involving nearly 298,000 participants in the Million Veteran Program who had one or more blood lipid measurements available for that analysis. As they reported, their results led to 141 suspicious loci with genome-wide significant ties to blood lipid levels.
The team subsequently shored up links between blood lipid levels and variants at 118 loci through a meta-analysis encompassing more than 600,000 additional individuals enrolled through the Global Lipids Genetics Consortium, using transcriptome-wide and phenome-wide association approaches to dig into new associations.
In particular, the researchers noted that the blood lipid-associated sites included variants expected to interfere with the functions of PDE3B — a gene that codes for a phosphodiesterase enzyme inhibited by the blood circulation drug cilostazol (Otsuka Pharmaceutical's Pletaal) — and two other genes.
The researchers used a custom MVP 1.0 Affymetrix Axiom biobank array to genotype 312,571 individuals from the Million Veteran Program at roughly 32 million variants. For their GWAS, they focused on 297,626 Million Veteran Program participants — including 57,332 African American individuals and 24,742 individuals with Hispanic ancestry — who had at least one laboratory-based blood lipid measurement in their electronic health records.
Through the multi-stage discovery and replication analyses, they narrowed in on variants with significant ties to blood lipid traits, particularly blood levels of low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, total cholesterol, and triglyceride levels.
The team brought in available array- and sequencing-based gene expression data to explore the associations further, focusing in on the predicted loss-of-function alterations in PDE3B and other genes. In the case of PDE3B, for example, data from the Million Veteran Program and for some 45,000 participants in the DiscovEHR study suggested that loss-of-function mutations in the gene lead to lower-than-usual HDL cholesterol and triglyceride levels.
For the PheWAS analysis, meanwhile, the researchers found nearly three dozen phenotypes that appeared to coincide with blood lipid traits in a subset of 176,913 participants of European ancestry. Those findings are expected to provide a clearer understanding of how treatments that alter blood lipid levels might impact other conditions such as type 2 diabetes or heart disease risk.
But Assimes also urged some caution. "The genetics help suggest that this drug can decrease the risk of heart disease by lowering triglycerides, but it's not proof," he said. "I would not prescribe it until a large randomized trial is completed with cilostazol or a related drug looking specifically at heart disease outcomes."
He concluded, "Better cholesterol profiles can look great, but if the drug doesn't affect the outcome you're aiming for, which is heart attack in this case, then it's useless."