NEW YORK – An international team led by researchers from Finland and the UK has characterized the genetic underpinnings of human metabolic processes, laying the foundation for studies of how metabolism affects a variety of diseases.
For their genome-wide association study, published in Nature on Wednesday, the investigators examined 233 circulating metabolic traits, quantified by nuclear magnetic resonance (NMR) spectroscopy, in up to 136,000 individuals from 33 cohorts. They discovered more than 400 genetic loci associated with those traits and identified likely causal genes for two-thirds of them. The results helped them characterize lipoprotein metabolism as well as a pregnancy-related liver disorder.
The new study builds on a 2016 GWAS by many of the same authors that used a similar approach to examine genetic influences on 123 circulating metabolic traits in up to 25,000 study participants. That study identified eight novel loci for amino acids, pyruvate, and fatty acids and suggested a new intervention to reduce cardiovascular risk.
"The fivefold increase in sample size and doubling of the number of metabolic traits compared to our previous GWAS meta-analysis of NMR metabolic traits led to a marked increase in the number of significant associations," corresponding author Minna Karjalainen, a researcher at the University of Oulu in Finland, and her colleagues wrote in their latest paper.
For their new study, the researchers quantified metabolic traits in plasma or serum samples from 33 cohorts, most of them of European ethnicity but also including participants of Chinese or South Indian origin.
The 233 traits measured included 213 lipid and lipoprotein parameters or fatty acids as well as 20 non-lipid traits such as amino acids, ketone bodies, glycolysis and gluconeogenesis, fluid balance, and inflammation metabolites.
Their GWAS included almost 13.4 million imputed SNPs for up to 136,016 participants. It revealed 8,000 genetic associations with the metabolic biomarkers examined and more than 400 genetic loci, for almost 300 of which manual curation identified a total of 231 likely causal genes.
The researchers investigated how these associations held up across ethnic groups — South Asian, East Asian, Finnish, and non-Finnish Europeans — and found that they were broadly transferable, though some effects were stronger in one group than another.
In addition, they checked whether they could replicate their associations in UK Biobank participants, for whom NMR data are also available, and found that the majority of associations were present in both cohorts. Sample type and fasting status were the main reasons for not being able to replicate a finding.
When they focused on associations across lipoprotein measures, the researchers found clusters of genes with similar metabolic profiles. In particular, the metabolic profile of the gene TRIM5 aligned well with genes known to affect LDL cholesterol uptake by hepatocytes. TRIM5 had previously been implicated in antiviral host defense but also in liver biology, and the new results suggested it could become a therapeutic target for lowering blood lipid levels to prevent cardiovascular disease. "Although we specifically chose the TRIM5 association for further investigation, our clustering analysis suggests there are several other novel loci worthy of further in-depth investigation," the authors noted.
They also took a closer look at seven metabolic trait-associated loci that showed a link to intrahepatic cholestasis of pregnancy (ICP), a liver disorder that can appear in the second and third trimester of pregnancy. A pathway analysis of those loci revealed an enrichment for processes related to bile acid, glucose, and lipid metabolism in ICP. Through this analysis, "we exemplify the value of combining the metabolic association information with disease associations to clarify the metabolic underpinnings of poorly understood conditions," the investigators wrote.
However, the researchers cautioned that the predominantly European ancestry of the cohorts in the study could limit how broadly generalizable the findings are. Also, mass spectrometry could have measured a larger number of metabolites, but the lower cost, higher throughput, and greater robustness of NMR meant they could include a larger number of individuals.
The identification of new genes associated with metabolic traits through manual curation "provides a useful resource to further biological understanding of the associations and allows high-confidence identification of causal genes for disease associations that colocalize," the authors wrote. "For the remaining loci, our results provide a starting point for identification of genes that have so far not been known to be involved in metabolic regulation."