What genomics can teach us about polygenic diabetes

Hi, my name’s Tim Frayling. I’m a researcher here at the University of Exeter Medical School, and I have been for more than 15 years now. And I run a small but highly productive research group studying the genetics of diseases, such as type 2 diabetes and related traits, such as obesity. So the genetic basis of type 2 diabetes and obesity and other common diseases, such as heart disease and cancer, is very different to the genetic basis of monogenic single gene diseases you heard about in Week one of the course. With things like type 2 diabetes and obesity, we’re talking about conditions which are very common. More than 5% of people have type 2 diabetes in the British population today.

A very large proportion, unfortunately, are overweight and obese. And so we’re not looking for a single gene which causes that disease. What we’re looking for is common genetic variation which has a subtler predisposing rather than causal effect on the disease. For example, we might be looking for a genetic variant, a version of the gene, what we call an allele, that is present in 20% of the type 2 diabetic population, but only 15% of the general population. Or a genetic variant that is present in 30% of the general population, 35% of people who are obese, sometimes even subtler effects than that.

So the point is that the risk gene, the risk version of the gene can be present in people without the condition. And it might not be always present in the people with the condition. What we’ve found over the last few years, since about 2007, as many as 10, sometimes hundreds of these subtle genetic effects influencing common diseases like type 2 diabetes. The field really took off in 2007 because we had a big change in technology. In 2007 something called a microarray was invented where we could analyse hundreds of thousands of simple units of genetic variation.

What we call single nucleotide polymorphisms, these are regions of the genome which a single letter in the DNA code might be changed, for example, from a G to an A. And some versions of the human genome have an A and some versions have a G. And what we’re looking for, as I say, is those subtle differences in that people with the disease might have more of the G allele than people without the disease. And what microarrays allowed us to do was analyse hundreds of thousands from across the genome and really, for the first time in 2007, we were able to get a snapshot of an individual person’s whole genome from at least a common variation in their whole genome.

And we were able to analyse that common variation against diseases, such as type 2 diabetes. So what did we find? Well, we were able to do those studies in increasingly large numbers of people. There are now studies of tens of thousands of people with diabetes, hundreds of thousands of people with measures of body mass index and obesity. And we’ve been able to pull out the genetic variation in the tens and hundreds of numbers which are influencing those diseases. What does that mean? Well, probably to a doctor and the individual patient, the applications are limited at the moment. It’s more about understanding mechanisms of disease. What have we learned from type 2 diabetes?

I think there’s several important things that have come out of the genetic studies over the last few years. One of the most interesting, perhaps not overly surprising, but one of the most interesting convincing things is that the two main types of diabetes, type 1 and type 2 diabetes, are really very different diseases. There’s virtually no shared genetic aetiology between type 1 diabetes and type 2 diabetes. The genes which are associated with type 1 diabetes are almost always involved in an autoimmune or immunity process, showing that type 1 diabetes is a disease of the immune system going wrong.

The genes hit by the type 2 diabetes, studies are very much about genes which are affecting insulin secretion, and to a lesser extent, insulin resistance. And that’s the second message coming out of the studies, I think, is that at least from an etiological point of view, insulin secretion and something going wrong, genetic predisposition to reduced insulin secretion, really appears to be a stronger risk factor for type 2 diabetes than insulin resistance, the ability of the muscle cells, and the fat cells, and the liver cells to absorb glucose in response to insulin, which is still an important factor in the obesegenic environment.