New Gene Variants Linked to Lipid Levels
Published: August 04, 2010
In a genetic analysis of more than 100,000 people of European ancestry, researchers have found 95 common genetic variants -- 59 of them previously unknown -- that influence lipid levels by regulating nearby genes.
Some of the variants are clustered near genes already known to be involved in regulating lipid levels but "scores" are near genes not previously linked to lipoprotein metabolism, the researchers wrote in two papers in the Aug. 5 issue of Nature.
* Explain to interested patients that two basic science articles in Nature examined new genetic variants associated with lipid disorders and risk of coronary artery disease.
* The investigators advocate that this approach can potentially define new pathways of lipid metabolism and targets for diagnostic tests and new therapeutic agents.
And at least one actually causes increases in low-density lipoprotein cholesterol (LDL-C) by decreasing the expression in the liver of a gene called Sort1. The gene lies in a region previously linked to about a 40% increase in the risk of heart attack, but the mechanism has not been understood.
The multinational team led by Sekar Kathiresan, MD, of Massachusetts General Hospital in Boston, and Daniel Rader, MD, of the University of Pennsylvania in Philadelphia, reported that the 95 single nucleotide polymorphisms (SNPs) play a role both in normal variation in lipid levels and in people whose lipids are at extreme levels.
They also found that the SNPs have an impact on lipid levels in three-non-European populations and link several novel regions with plasma lipid levels and coronary artery disease.
In mice, the researchers validated the findings in three novel genes -- dubbed GALNT2, PPP1R3B and TTC39B -- by showing that changing their expression altered lipid levels in the animals.
"It's safe to say that this is the largest genetic research study done to date for any trait or any disease," Kathiresan told MedPage Today.
As the size of such studies has increased, the number of discoveries has gone up in a linear fashion, but Kathiresan said there is likely to be a plateau level of 200 to 400 genes that would explain all of the genetic variability in lipid levels.
In contrast, the current study explains about a quarter of the genetic variation, he said.
"Larger sample sizes for most diseases are going to help find some more gene regions," he said, "but then we probably need to turn to complementary approaches to find the rest of it" -- approaches such as direct sequencing of regions of interest.
The findings offer a chance for scientists interested into translating basic science into clinical science to "roll up their sleeves," according to Alan Shuldiner, MD, and Toni Pollin, PhD, of the University of Maryland School of Medicine in Baltimore.
Writing in an accompanying editorial, they said the research strategy has led to several insights that have raised "many clinically relevant questions," including:
* Would a panel of 95 genetic tests have diagnostic value, beyond conventional measurements of blood lipids?
* What mechanisms underlie the effects of the novel genes?
* And would any of the loci make targets that might respond to drugs?
To make their discoveries, Kathiresan, Rader, and colleagues analyzed data from 46 genome-wide association studies that looked at plasma concentrations of total cholesterol, LDL-C, high density lipoprotein cholesterol (HDL-C), and triglycerides.
They found 95 regions that had genome-wide significant association with one of the four traits, where the significance level was set at P<5x10-5. Of those, 36 had previously been reported and of the remainder, 39 were associated with total cholesterol, 22 with LDL-C, 31 with HDL-C, and 16 with triglycerides.
In three non-European populations -- East Asians, South Asians and African Americans -- most of the lead SNPs showed the same direction of association as they did in the larger European population, suggesting that many of the 95 lipid loci "contribute to the genetic architecture of lipid traits widely across global populations," the researchers wrote.
At least some of the SNPs have clinical relevance, the researchers found by comparing people with and without coronary artery disease or hyperlipidemia. For both conditions, they found associations that were highly significant, at P<0.001. For instance, they reported, 14 of the SNPs were associated with coronary artery disease.
To test the biological relevance of the findings, the researchers manipulated three of the novel genes in experimental animals.
For instance, they used a viral vector to induce overexpression of the mouse version of GALNT2 in the livers of experimental animals. Within four weeks, the treated animals had a 24% decrease in HDL-C, compared with control animals, which was significant at P=0.002.
In contrast, when they knocked down the gene, using a short hairpin RNA segment, the mice had a 71% increase in HDL-C levels, compared with control animals (P<0.0001).
The researchers found similar results when they manipulated PPP1R3B (associated with HDL-C, LDL-C and total cholesterol) and TTC39B (associated with HDL-C).
In the companion study, the research team used similar methods to try to tease out the mechanism behind the observation that people with two copies of the region in which Sort1 is found have an increased risk of heart attack.
They showed that an SNP (dubbed s12740374) on chromosome 1 -- which itself does not code for any protein -- nevertheless can create or disrupt a transcription factor binding site and thereby alter the expression of Sort1 in the liver.
In mice, they showed, over and underexpression of the gene in the liver leads to decreases and increases, respectively, in total plasma cholesterol and LDL-C.
Taken together, they wrote, the experiments demonstrated that Sort1 is the causal gene that leads to lipid changes and increased heart attack risk, and that its expression is under the control of a noncoding section of DNA, the SNP s12740374.
The findings suggest a "promising new target for therapeutic intervention" to reduce LDL-C and heart attack risk, they concluded.
That target may not be the gene product of Sort1 -- a protein dubbed sortilin -- according to Kathiresan. But the study "exposes basically a new pathway, a new way of regulating lipids, LDL cholesterol, that's actually independent of or different than previous approaches," he said.
"So sortilin itself or other genes that interact with sortilin may in time turn out to be good drug targets," he added, "but time will tell."
The most immediate clinical impact, he said, could be a panel of SNPs that would help diagnostic and treatment decisions.
The paper by Kathiresan, Teslovich et al was supported by the National Heart, Lung, and Blood Institute, the Donovan Family Foundation, and the Broad Institute Center for Genotyping and Analysis.
Kathiresan did not declare any competing interests. However, several authors are full-time employees of GlaxoSmithKline and others report financial links with the company. Also several authors are full-time employees of deCODE genetics.
The paper by Rader, Musunuru, et al had support from the National Heart, Lung, and Blood Institute, as well as the Harvard Stem Cell Institute, the National Cancer Institute, the Swedish Medical Research Council, Quest Diagnostics, and Bristol-Myers Squibb.
Rader declared no competing interests. However, several authors are employees of Alnylam Pharmaceuticals and one author reported research support from Quest Diagnostics related to work presented in this manuscript.
Shuldiner and Pollin did not report any external support for their analysis. Shuldiner is a consultant for ISIS Pharmaceuticals.
Primary source: Nature
Teslovich TM, et al "Biological, clinical and population relevance of 95 loci for blood lipids" Nature 2010; 466: 707-713.
Additional source: Nature
Musunuru K, et al "From noncoding variant to phenotype via SORT1 at the 1p13 cholesterol locus" Nature 2010; 466: 714-721.
Additional source: Nature
Shuldiner AR, Pollin TI "Variations in blood lipids" Nature 2010; 466: 703-704.