NEW YORK (GenomeWeb News) – Through an epigenome-wide association study, researchers have linked DNA methylation at three spots in one gene to body-mass index, as they reported in the Lancet yesterday evening.
Nilesh Samani, a cardiology professor at the University of Leicester in the UK, and his colleagues examined the methylation status of nearly 480 people, finding a handful of probes associated with BMI. The researchers confirmed three of those probes — all in intron one of the HIF3A gene — in two additional cohorts.
HIF3A is a subunit of the hypoxia inducible transcription factor (HIF), which regulates cellular and physiological response to low-oxygen conditions. HIF also, the researchers said, has been linked to metabolism, energy expenditure, and obesity.
The researchers also found that the methylation changes occurred in adipose tissue, but not in skin samples. They noted, though, that methylation at these HIF3A sites was likely a consequence rather than a cause of increased BMI.
"To find that the methylation of HIF3A is increasingly altered as someone becomes more obese is remarkable and raises the possibility that HIF may also be involved in mediating some of the deleterious effects of becoming overweight," Samani said in a statement.
To home in on DNA methylation changes that might be associated with BMI, Samani and his colleagues typed DNA isolated from whole-blood samples from 479 people of European ancestry using the Illumina Infinium HumanMethylation450 array. From this, they uncovered five sites across three genes linked to BMI, which they then tested in two replication cohorts, also comprised of people of European ancestry.
Three probes — cg22891070, cg27146050, and cg16672562 — were confirmed in both cohorts. The probes, the researchers reported, are located near to each other in intron 1 of the HIF3A gene, and methylation levels at those probes are highly correlated with each other.
DNA methylation at HIF3A was not, the researchers reported, associated with height, which is a component of the BMI calculation, nor was it linked to other traits related to BMI, such as physical activity and type 2 diabetes.
The researchers also calculated that for every 10 percent increase in methylation at the most significant of the probe sites, a person in the discovery cohort with an average BMI of 27 kg/m2 would have a 3.6 percent rise in BMI, or gain about 0.98 kg/m2.
As methylation states can vary between tissue types, the researchers examined its levels in both adipose and skin biopsy samples, finding a strong association between methylation at the HIF3A gene sites and BMI in adipose, but not in skin, tissue. Additionally, they noted that the direction of the association between HIF3A methylation in adipose tissue and BMI was the same as in the blood, though the percent change was greater.
They also reported a weak, though significant, inverse correlation between methylation at one of the probes and HIF3A gene expression. Previous studies, they noted, have found that the three BMI-related CpG sites in the HIF3A gene are within an open chromatin region that may have a regulatory function.
Additionally, as DNA sequence can also influence methylation status, the researchers searched for possible associations between SNPs lurking within 1 megabase of one of the probes and its methylation. Two SNPs, located upstream of HIF3A, were associated with methylation, the researcher reported, but they were not linked to BMI in either the discovery cohort or in a dataset from the GIANT consortium, an international collaboration investigating genetic loci that modulate human body size and shape.
"These findings suggest that the association between increased methylation and higher BMI is not causal," the researchers wrote in the paper, adding, "our findings suggest that increased methylation at the HIF3A locus is a result of increased BMI."
Even though there might not be a direct causal relationship between HIF3A and BMI, there could be a role for HIF signaling in mediating the response to increased BMI.
"A non-causal association between methylation and a phenotype could still be informative as a diagnostic or prognostic biomarker," Therese Murphy and Jonathan Mill from the University of Exeter wrote in a related Lancet comment piece. "[For instance,] HIF3A methylation might predict disease phenotypes associated with BMI, such as cancer and cardiovascular disease."