NEW YORK – The genetic variants influencing blood lipid levels and related coronary artery disease (CAD) risk may be influenced by other genetic variants as well as environmental factors, according to new research by investigators in Iceland, Denmark, and the US.
"One of the things that has proven difficult in human genetics is to show interactions of variants in the sequence of the genome with other sequence variants (epistasis) and with components of the environment," co-senior and co-corresponding author Daniel Gudbjartsson, a researcher affiliated with Amgen subsidiary Decode Genetics and the University of Iceland, said in an email. "However, it is likely that such interactions contribute significantly to necessary homeostasis in human biology."
To dig into those interactions in the context of lipid traits and heart disease, the researchers searched for sequence variants with apparent ties to variability in blood lipid levels, rather than focusing exclusively on associations linked to mean measurements of these quantitative traits.
As they reported in Cell on Thursday, Gudbjartsson and his colleagues analyzed genotyping profiles for more than 743,700 individuals from the UK Biobank, Copenhagen Hospital Biobank in Denmark, and the Icelandic population. With these data, they searched for variants that coincided with mean measurements for eight lipid traits as well as variants corresponding to the variance in the same set of lipid traits.
By combining these data, the team was able to distinguish between variant associations marked by dominance effects and those featuring between-variant interactions, gene-environment interactions, or correlated effects. They then considered how these lipid effects impacted CAD risk using data for 149,007 affected individuals and more than 942,400 unaffected controls.
"We show that sequence variants can have strong effects on the correlation between lipid traits and their variability, which gives us a better understanding of their overall impact on biology," Gudbjartsson said, noting that "effects of sequence variants on phenotypic variance are mostly caused by their interactions with other sequence variants and the environment, which gives us a foothold for systematic studies of interactions."
For example, the study found a potentially clinically relevant interaction involving an APOE2 variant and levels of "bad" cholesterol, namely low-density lipoprotein (LDL) and other non-high-density lipoprotein (non-HDL) cholesterol.
APOE2, an allele known for decreasing Alzheimer's disease risk when present in both copies of the gene, corresponded with lower-than-usual levels of a cholesterol-carrying apolipoprotein B (apoB) particle, Gudbjartsson explained. Even so, APOE2 variant homozygotes had CAD risk that fluctuated in relation to LDL and non-HDL cholesterol levels in a manner that is comparable to individuals without the Alzheimer's protective variant.
"This suggests that monitoring bad cholesterol is clinically more important than apoB levels, which runs contrary to most clinical guidelines," he said.
The investigators also flagged a genetic variant in the ADH1B alcohol dehydrogenase gene that stretches out the process of alcohol metabolism, dialing down the LDL cholesterol consequences of alcohol drinking. In addition, they highlighted interactions between a variant in FUT2, a blood group-related gene, cholesterol levels, and CAD that was dependent on individuals' ABO blood group.
Among the other interactions identified, meanwhile, the team saw a relationship between TM6SF2 variants and lipid levels that differed between individuals who did or did not report having "oily fish" such as sardines or herring in their diets.
Together, the authors explained, the study "demonstrates how variants in the sequence of the genome can have complex effects, such as dominance effects, effects on variance and correlation, and interactions among variants and between variants and environmental components."