Blood-Based Epigenetic Screen Tests for Diabetes Complications

“I thought it was novel,” remarks Philip Marsden, a physician-scientist at the University of Toronto whose research focuses on epigenetics and cardiovascular disease. “I hadn’t been aware they could do hydroxymethylation screens of circulating free DNA.” However, it’s still early days for the budding technology, he notes.

The team’s tool works by chemically tagging 5-hydroxymethylcytosines (5-hmC), epigenetic modifications that form when cytosine receives a methyl group that is then oxidized into a hydroxymethyl group. While methylation of cytosines is thought to repress gene expression, hydroxymethylation is associated with gene activation, and has been increasingly implicated in complex diseases such as cancer.

Prior to 2011, there was no way to distinguish methylated cytosines from hydroxymethylated cytosines. Then, University of Chicago epigeneticist Chuan He developed a labeling technique —called the 5-hmC seal—that effectively “coats” all 5-hydroxymethylcytosines along a given strand of DNA. Earlier this year using the tool, He’s group identified a set of epigenetic changes in freely circulating DNA in early-stage liver cancer patients, which they propose could serve as a biomarker for the disease.

See “The Role of DNA Base Modifications”

This approach would work in any disease or condition in which injured cells spill free DNA into the bloodstream. The release of free DNA is known to occur in cancer, and He’s team suspected it might also happen in diabetic patients with vascular complications. “We reasoned that . . .  there could be damage to the tissue, damage to the cells, that would release free circulating DNA,” explains Wei Zhang, a geneticist and clinical epidemiologist at Northwestern University’s Feinberg School of Medicine and one of He’s coauthors on the new study. Zhang and He both own shares in Shanghai Epican Genetech Co, a company that has patented the 5-hmC seal.

If vascular complications increase blood levels of freely circulating DNA, that would allow the team to detect any epigenetic changes that may have occurred as a result of vascular complications—perhaps as cells tried to adapt to high blood sugar levels or as they became injured, Zhang and his colleagues proposed.

To test this idea, the team recruited 62 patients at Wuhan University in China. Twelve of them didn’t have any vascular complications, while the remainder experienced a range of symptoms—including atherosclerotic plaques, heart disease, stroke, or damage to nerves, eyes, or kidneys.

The scientists drew a few drops of blood from each patient, extracted their freely circulating DNA, and applied the 5-hmC Seal. They then sequenced the DNA, and for each gene they compared the extent of hydroxymethylation between patients with vascular complications and patients without. This revealed 135 genes with significantly different patterns of hydroxymethylation between the two groups. The genes were often implicated in insulin resistance or inflammation, the researchers noted.

Using a machine learning algorithm, they identified 16 genes that had the most distinct differences between patients with and without vascular complications. They found they could use the hydroxymethylation patterns of these 16 genes to distinguish patients with vascular complications from those without with an accuracy of 85 percent or higher. The tool proved more accurate in distinguishing between the two groups than did concentrations of urinary albumin, a protein considered a marker of kidney disease.

“This study proves that our technology is so sensitive that it can be used not only in cancer . . . but also in chronic disease such as diabetes,” Zhang says. He hopes he and his colleagues can develop the technique further into a blood test, based on just a few drops, that can be used in the clinic to detect vascular complications early on, he adds.

Renu Kowluru, a diabetes researcher at Wayne State University who wasn’t involved in the study, agrees with Marsden that the approach could someday prove useful and is worth exploring. “To find any noninvasive marker for diabetic complications is a big thing,” she says.

However, the study itself is a “messy” one, she finds. When examining the association between epigenetic changes and vascular complications, the team had lumped all vascular complications—from atherosclerosis to retinopathy—together, and didn’t provide information about which complications were the most correlative. Treating all vascular complications as if they were the same is “my major concern with this study,” she says.

Marsden, who also wasn’t involved in the work, would like to see more research investigating the significance of the epigenetic changes seen in patients with complications: whether they have a causative role in the vascular complications, or if they’re just associated with them. “I don’t think they were self-critical here about whether it’s a cause-and-effect relationship or an association, and their data is association at best,” he adds. It would also help to know how much free circulating DNA was present in the patients with vascular complications compared to those without—something the team didn’t report. “It could be that they just saw these [epigenetic] patterns because there’s more total DNA there,” he suggests.

Marsden says he also wonders how expensive the tool will be for scientists and clinicians to use. “And how available is it to the average researcher—do I always have to work with someone in Chicago [to use it]?” he asks. Nevertheless, “this paper definitely stimulated me to want to learn more about measuring hydroxymethylation in cell-free DNA in patients and in normal biology.”

Y. Yang et al., “5-Hydroxymethylcytosines in circulating cell-free DNA reveal vascular complications of type 2 diabetes,” doi:10.1373/clinchem.2019.305508, Clinical Chemistry, 2019.

Katarina Zimmer is a New York–based freelance journalist. Find her on Twitter @katarinazimmer.