Diabetes Breakthrough: Genes Linked to Type 2

Type 2 diabetes is a severe and growing problem in the United States. In the last seven years, there has been a 54 percent increase in the disease rate, to roughly 20 million cases nationwide. According to the Centers for Disease Control and Prevention, 1 in 3 children born in 2000 will develop the disease in his or her lifetime. Much of the increase is due to obesity and lack of exercise—two factors that have long been associated with the disease. But the genetic underpinnings that put people at risk have been largely a mystery. That may not be the case for much longer.

In a study appearing Sunday in the online version of the journal Nature, an international team of scientists reports on a group of four new gene variants that appear to play a role in influencing susceptibility to type 2 diabetes. These are clearly not the only genes involved in diabetes risk. But their discovery marks an important step toward unraveling the genetic "fingerprint" of the disease. NEWSWEEK's Anne Underwood spoke with Dr. Constantin Polychronakos of McGill University in Montreal, one of the lead authors on the paper. Excerpts:

NEWSWEEK: How did you conduct this study?
Constantin Polychronakos:
We screened 7,000 people—3,500 with diabetes and 3,500 without—using DNA samples from the Pasteur Institute in France. We found four genetic variants that confer type 2 diabetes risk, in addition to confirming a known association with another gene. This is just the beginning. When we've finishing screening, we expect to have another 10 or so gene variants.

Why publish now, then?
The first four are interesting enough. The most interesting is SLC30A8, which codes for the zinc transporter in beta cells [the insulin-producing cells in the pancreas]. You might wonder, what does zinc have to do with insulin? The form of insulin that circulates in the blood is basically four molecules of insulin stuck together. The glue that holds them together is zinc. People who have a very active form of the gene have better insulin secretion in response to glucose in the bloodstream.

How might this information be applied?
It tells us that maybe a zinc deficiency in the diet could be involved in diabetes. Or maybe we could come up with a drug that influences the zinc transporter. It suggests a therapeutic intervention, based on whether this person needs zinc and that one doesn't.

What were the other genetic variations you found?
Two involve blocks of more than one gene that are found together in the same stretch of DNA. The advantage of a block is that it's easier to detect the disease association in the first place. But it also makes it harder to determine which of the [variations within that block] is responsible for the effect. It's a double-edged sword.

Do you see some kind of diagnostic test coming out of this?
To diagnose diabetes, you don't need a genetic test. A blood-sugar reading will tell you whether or not you have the disease. The purpose of a gene test would be to predict diabetes beforehand. If we knew all the genes, we could put them on a chip and screen every newborn with one drop of blood. We could then tell parents, this child has a 50 to 70 percent chance of developing diabetes if she doesn't watch her diet and exercise. Instead of trying to convince the entire population to live healthy, which is hard to do, we could concentrate on the portion of the population that is most susceptible.

Would it help doctors to prescribe the best treatment, as well?
That's the more important application. It would help us understand why people get diabetes. We know that people get diabetes for different reasons. The most common cause of insulin resistance is obesity. But there must be other causes as well, because some people who develop insulin resistance are not obese, and the same degree of obesity can result in different levels of insulin resistance in different people. What causes [the pancreas to stop secreting insulin] is not known at all. There may well be more than one cause. When we have all the genes, we will be able to say, this person will respond to this treatment and that one will respond to another.

How common are the gene variants you identified?
They're very common. You don't need rare alleles [forms of the gene] to predispose you to diabetes.

You say in the paper that the effect of these gene variants together is relatively small.
That's right. I expect that by the time we're finished, we will find roughly 20 gene variants that affect diabetes risk. We should have them all in the next six months to a year.

You call these gene variants "ancestral." What does that mean?
We compare the genome of humans with that of the chimpanzee. If the disease-causing variants in humans are found in the chimpanzee, too, it means that form of the gene dates back to the common ancestor of humans and chimps. When a novel variant arises, it can remain rare over many generations or increase in frequency. If it increases, that's because it confers some advantage. In this case, the [gene variants that confer risk] are the ancestral alleles.

What are the implications of that?
It may be consistent with the hypothesis that the susceptibility genes for diabetes were adapted to the environment of the ancient human population [when food was not so abundant]. They must have been advantageous when people were starving. Today, in a different environment, they increase risk.

Diabetes Breakthrough: Genes Linked to Type 2 | Culture