9th December 2013
by Deepti Gurdasani
This makes the study of lipid regulation important, as the hope is that if we could understand how these levels are regulated genetically, we could perhaps develop drugs that can target the specific biological pathways involved.
The recent work published by the Global Lipids Genetics Consortium (GLGC) in Nature Genetics identified 157 sites across the genome that are associated with lipid levels, of which 62 sites had not been discovered previously. Many of the new loci found to be associated with lipids were close to genes known to be involved in lipid regulation in some way but others were in completely new regions where no known genes associated with lipids are found. Understanding these regions may provide novel methods of regulating lipids in individuals and therefore reduce risk of heart disease.
This is also one of the few studies to make use of the genetic architecture of African populations. As African populations are more genetically diverse than European or Asian populations, the study of genetic factors influencing traits in these populations allows us to more precisely identify loci within genes involved in lipid regulation, as well as identifying new loci that regulate these only in African populations. This raises many exciting possibilities for potential treatment targets in the future.
In addition to identifying the sites of lipid regulation, the GLGC also evaluated the causal role of different lipids in heart disease, separating direct causation from other confounding factors such as lifestyle.
For instance, does having low levels of high-density lipoprotein cause heart disease or does having low levels of high-density lipoprotein make you more likely to be inactive, which could cause heart disease? This distinction is important to make, because if lipid levels were only associated with heart disease because they are associated with another factor that was the actual causal agent, just lowering lipid levels would not reduce the risk of heart disease.
But how do we tease out whether factors are just associated or causally linked to disease? One way is to randomise individuals to groups; reduce the level of the factor experimentally in one group, keep all other conditions the same in both groups and see if this affects their risk of disease. However, clinical trials such as this can take a long time to complete and are very expensive.
Another way can be to identify sites within genes that increase or decrease specific lipid levels. As genes are randomly allocated at birth, if we are able to identify genes that affect specific lipid levels, this provides a natural experiment to study the lifelong effect of altered lipid levels. This can be thought of as similar to a randomised trial, where individuals with a specific genetic variant that increases lipids are our test cases, and individuals who lack the genetic variant are our controls. Therefore, if the genes that increase or decrease lipid levels are also associated with disease through their effect on lipids, it is very likely that lipid levels are causally associated with disease.
This study examined this relationship for two lipids (high-density lipoprotein cholesterol and triglycerides) where the relationship with heart disease is unclear. It was able to show that it’s quite likely that triglycerides are linked to heart disease, while high-density lipoprotein cholesterol is unlikely to be causally associated with heart disease.
Whether these lipids are causally associated with heart disease has important implications for whether high levels of triglycerides would benefit from treatment, which might help prioritise interventional efforts for the reduction of heart disease.
Deepti is a post-doctoral researcher working with Manj Sandhu at the Wellcome Trust Sanger Institute. Her research primarily focuses on the study of genetic diversity in Africa, and how this affects disease susceptibility.
- Do R, et al (2013) Common variants associated with plasma triglycerides and risk for coronary artery disease. Nature Genetics. doi:10.1038/ng.2795