Large study helps unearth cholesterol’s genetic roots

Cholesterol chartImage courtesy of iStockphoto

In 1948, thousands of residents of Framingham, Massachusetts enrolled in a new kind of health study. A team of scientists wanted to follow these people for years, even decades, in an effort to discover traits that might cause heart disease, which was poorly understood at the time. Just thirteen years later, the work had already begun to shed light on the mysteries of heart disease, pointing to cigarette smoking, blood pressure, and cholesterol level as risk factors.

In the decades since those first revelations, a total of more than 15,000 people have taken part in the Framingham Heart Study, including descendants of those in the original cohort, and the work continues to yield insights into heart disease, still a leading cause of death in the United States. Now, the children and grandchildren of those original study participants are helping to reveal how genes can, in turn, influence the factors initially found to raise heart disease risk, such as cholesterol level.

In a new large-scale study known as a meta-analysis, data from a recent genome scan of these second- and third-generation Framingham Heart Study participants has been studied along with data from 11 other genome studies. With collaborators across the globe, researchers at the Broad Institute of MIT and Harvard and elsewhere combined data on more than 40,000 people to find 11 new spots in the genome that contribute to fat levels in the blood. Among the largest meta-analyses performed to date, the new work brings the total number of genomic regions associated with cholesterol and triglycerides to 30 and helps explain roughly 20% of the genetic basis for blood lipid levels. The findings, which appear in the December 7 online edition of Nature Genetics, may also help scientists identify new drug targets, understand how blood cholesterol is regulated, and even predict whether a person is genetically predisposed to high levels of blood lipids.

In earlier work, the researchers had found 19 sites in the genome that contributed to differing cholesterol levels. To expand that list even more, they scrutinized 2.5 million sites in the genomes of 20,000 people. DNA spots that seemed important were then studied in another 20,000 people to see if they affected cholesterol levels. Most of the eleven new gene regions they found had no prior link to blood cholesterol in humans, said Sekar Kathiresan, who is first author of the new study, a researcher in the Broad Institute’s Program in Medical and Population Genetics, and director of preventive cardiology at Massachusetts General Hospital. The new loci not only explain more of the inter-individual variation in blood lipid levels, but the implicated genes may open new doors to understanding how cholesterol is regulated in the body.

Studies like this look for single-letter changes, or SNPs, in DNA linked to a trait or disease. They help scientists hone in on an important region of the genome, but they often fail to reveal the exact cause. The SNP itself could substitute one amino acid along a protein’s chain for another, giving a different protein shape and function. Or the SNP could merely be a marker for a separate DNA change, such as extra or missing DNA, to which the SNP is linked. To find the important bits of DNA that the SNP signifies, geneticists can sequence the region of DNA near it, reading the exact sequence of As, Ts, Cs, and Gs to find some variation from the norm that might be important.

Another way to determine the biological counterpart of the SNP is to measure how different DNA spellings affect the amounts of protein being produced. By testing the levels of RNA — an indirect measure of gene activity, and therefore, protein levels — in hundreds of human liver samples, Kathiresan and his colleagues discovered that 7 of the 30 cholesterol-influencing SNPs also altered the activity of nearby genes.

Some rare cases of extremely high cholesterol are due to mutations that dramatically change or shut down the function of a protein. Changing gene activity, however, might simply alter how much protein is made. “We were surprised to find that so many of the SNPs altered gene expression in human liver,” said Kathiresan. “So it appears that for the common genetic variants that nudge cholesterol one way or another, what’s important could be the quantity of proteins, not the quality.”

Another insight into rare and common causes of altered cholesterol comes from the discovery that genes can vary in multiple ways. For example, a rare mutation in the PCSK9 gene found in only a few families in the world changes LDL cholesterol — the so-called “bad cholesterol” — by roughly 100 mg/dL (the optimal level is below 100). A more common mutation present in 1 in 50 people changes LDL cholesterol by 16 mg/dL. And one of the 30 SNPs, near PCSK9 and found in 1 in 5 people, changes that level by a modest 3 mg/dL. So it’s possible that for each of the 30 common SNPs, other mutations in the same genes could give rise to rare forms of extremely high cholesterol.

The exact biological pathways through which the SNPs alter cholesterol have yet to be found, and more research is needed. But unlike other genetic traits or diseases that lack effective treatments, prediction is an important goal for cholesterol researchers because of the availability of cholesterol-lowering drugs, like statins. A comprehensive list of genes controlling blood lipid levels could give researchers the power to identify people at risk of having extremely high cholesterol, and who should be given statins early in life to prevent heart disease and reduce their risk of heart attack. To find all the genes that influence cholesterol levels, researchers must scan the genomes of even greater numbers of people from diverse ethnic backgrounds, and sequence the genes in hundreds to thousands of people to discover all the ways DNA is altered in those genes.

The findings also raise the possibility of unearthing new drug targets. “Getting from genes to treatment is a long road, but our list of genes is a promising start, because we know they’re important in humans,” said Kathiresan. Among the 30 gene regions in this study is HMGCR, a gene blocked by statin medications. Kathiresan and his fellow researchers wonder if any of the other 29 regions, or others yet to come, could lead to new treatments. “The development of statins was one of the most dramatic stories of cardiovascular medicine over the last 30 years,” he said. “If we found even one more drug target like that, it would make this whole enterprise worthwhile.”

Other Broad Institute researchers contributing to this work include Benjamin Voight, Gabriel Crawford, Aarti Surti, Candace Guiducci, Noel Burtt, Paul I W de Bakker, Leena Peltonen, and David Altshuler. The meta-analysis combined data from seven genome-wide association studies in the first phase (the Framingham Heart Study, the London Life Sciences Prospective Population Cohort study, the Supplementation en Vitamines et Mineraux Antixoxydants study, the Invecchiare in Chianti study, the Diabetes Genetics Initiative, the Finland-United States Investigation of NIDDM Genetics (FUSION), and the SardiNIA Study of Aging), and five in the second phase (the Malmo Diet and Cancer Study, the FINRISK97 study, FUSION Stage 2, the Metabolic Syndrome in Men study, and the International Study of Infarct Survival).

Paper(s) cited

Kathiresan et al., Common variants at 30 loci contribute to polygenic dyslipidemia. Nature Genetics advance online publication. December 7, 2008. DOI: 10.1038/ng.291.