Feature: Dissecting diabetes

Trying to treat diabetes when you do not know the underlying cause is like trying to start a broken-down car without knowing the underlying mechanical problem. For cars, of course, a mechanic will eventually be able to find the dead spark plugs, broken starter motor or a blocked fuel pump at fault; with a new part, the car will up and running again. But for type 2 diabetes, the causes have been elusive and doctors have to rely on treating the outcomes: high blood sugar, blood pressure and cholesterol. Fortunately, with aggressive treatment of these risk factors for cardiovascular disease, the outlook can be improved markedly.

So how can we find the underlying problems in type 2 diabetes? Clues are beginning to emerge from its variability: it may not be the single disease it appears at first. In general, middle-aged, overweight people get diabetes because their tissues become less sensitive to insulin, and the pancreas cannot produce sufficient insulin to prevent the levels of sugar in the blood increasing to potentially damaging levels.

But a closer look reveals that type 2 diabetes is not at all uniform. It may be more common in obese people, but many slim people are affected. Some people respond well to certain drugs, but others have little benefit. And the diabetes found in people from the Indian subcontinent is different in many aspects from that of Caucasians.

"We want to know why some people get diabetes and others don’t, but we still have fundamental problems," says Professor Andrew Hattersley at the University of Exeter. "We know that diet and exercise are key, but the majority of us do too little exercise, eat the wrong food, and we don't get diabetes. One way of finding the primary pathophysiology is to look at the molecular basis of the genetic susceptibility to diabetes."

Now, the genes underlying certain forms of diabetes are being discovered, and are having a direct impact on treatment.

Finding genes

Type 2 diabetes is a classic 'polygenic' disease, caused by the interaction of multiple genes and the environment. But this is not always the case: in 1–2 per cent of cases, a mutation in a single gene can cause diabetes. "When I moved to Oxford at the end of the 1980s, I got some extremely good advice from John Todd, John Bell and David Weatherall," says Professor Hattersley. "They said that if I wanted to look for genes involved in diabetes, it would be best to look at single gene disease rather than polygenic disease. It was the best advice I've had!"

The commonest type of single gene diabetes was originally classified clinically as 'maturity onset diabetes of the young' (MODY). This is a form of young-onset diabetes (typically diagnosed under 25) that is not insulin-dependent and is inherited. Since 1992, mutations in six genes have been found to cause MODY, most cases involving the enzyme glucokinase or one of three gene-regulatory proteins: hepatic nuclear factor-1a (HNF-1a), HNF-4a and HNF-1b.

Mutations in these genes affect insulin production by the beta cells, but the outcomes are markedly different. People with glucokinase mutations are born with mildly raised blood glucose levels, and their condition deteriorates little throughout their life. With a careful diet, there is usually no need for insulin injections or drug treatment. People with mutations in one of the MODY transcription factors are usually born with normal blood glucose but usually develop diabetes between the ages of ten and 25. In fact, they are often misdiagnosed as having type 1 diabetes, as they are in the right age range, are slim and can have high blood glucose levels.

"Recently our work on gene discovery has moved from MODY, where all the common genes have been found, to neonatal diabetes – people who are diagnosed with diabetes within six months of life and who depend on insulin injections," says Professor Hattersley. "As type 1 diabetes is only seen in patients diagnosed after six months, all of these cases are genetic and there is still a lot to be discovered."

In April 2004, for example, his team found that a mutation in the Kir6.2 gene, which produces part of a potassium channel that controls the release of insulin, is responsible for 30–50 per cent of neonatal diabetes. "It's a considerably under-recognised disorder," says Professor Hattersley. "With the invaluable help of paediatricians from all over the world we have now looked at patients from 34 countries and from five continents; these are the people you most want to help – as they would be taking insulin for the rest of their lives."

Different genes, different treatment

Finding the underlying genes can have important implications for treatment in single gene diabetes. For Professor Hattersley this has been the most exciting work he has been involved with: "We had always hoped we would go from the diabetic patient to the gene and back to patient care. Now, this is a reality, not just a dream."

In a randomised trial, his team compared drug responses to the two most commonly used diabetes drugs in people with HNF-1a mutations and matched individuals with type 2 diabetes. The study showed that the fall in blood glucose in response to sulphonylureas (a common diabetes drug) was four-fold greater for people with diabetes caused by defects in HNF-1a.

"At present, we tend to treat people depending on the level of their blood sugar," says Professor Hattersley. "But patients with HNF-1a are different. We've found that they are a lot more sensitive to sulphonylureas, so sensitive that some patients can even stop injecting insulin, take sulphonylurea tablets, and get excellent control of their blood sugar. This is a clear example of how knowing the genetic aetiology can alter treatment choice."

The most dramatic instance where defining the gene alters the treatment is in neonatal diabetes. The potassium channel is a key part of the link between glucose metabolism by the beta cell and insulin release. Normally, the metabolism of glucose produces ATP, which binds to and closes the channel. The mutation in the Kir6.2 gene stops the potassium channel responding to ATP; the channel stays open and insulin cannot be released.

But the potassium channel could be closed by sulphonylureas, which bind to a different part of the channel, and hence insulin would be secreted. This works not only in cells in the laboratory but also in trials with people. "We've found that these patients who we have shown are secreting no insulin at all can get extremely good control of their blood sugar with sulphonylurea drugs, better than they ever had with insulin," says Professor Hattersley. "So far, 29 patients are insulin-free – it has transformed their lives."

Spreading the word

"The genetics are interesting, and help us to understand how changes to the structure of the channel lead to diabetes. But, more importantly, this is now altering clinical practice: as soon as we diagnose someone with a Kir6.2 mutation, improved treatment is there ready for them."

At the Diabetes Genetics Centre in Exeter, Professor Hattersley and colleagues have set up diagnostic testing in diabetes in their NHS laboratory, so it is available to any NHS Trust in the UK. Most doctors and nurses have not heard about the genetic subtypes of diabetes, so there is a considerable educational challenge. With support from the Department of Health, the team has been training doctors and diabetes specialist nurses to recognise and manage genetic diabetes.

"We spend a lot of time teaching diabetologists and diabetic nurses how to identify the important subtle differences between patients. It takes a long time for scientific discoveries to reach clinical practice, so we have to try to spread the news," he says. "For permanent neonatal diabetes, this has happened very quickly, especially now that it is altering treatment. We only described it in April 2004, and we've already had nearly 300 referrals."

Common variations

Although MODY is relatively rare, might common variations in the genes involved influence the risk of developing the more common form of type 2 diabetes? "This is a really exciting question," says Professor Hattersley. "If a severe mutation results in diabetes, as is the case in MODY, it may be that a common, mild variation might be a susceptibility gene increasing the chances of getting diabetes. And it follows that there may be subgroups in the type 2 diabetes population with certain shared variations. So, type 2 may not be a single disease at all, it may well be a collection of disorders that are characterised by high blood sugar levels and by insulin resistance.” Already, work has shown that common variations in the glucokinase and Kir6.2 genes are important in the general population and in type 2 diabetes.

Equally, the same principle that has been applied to the treatment of the single gene forms of diabetes – finding the treatments that work best for each group – could be applied to subgroups within the wider population of people with type 2 diabetes. "We already know that diabetes patients respond differently to treatment," says Professor Hattersley. "Thin patients are different to fat patients; Indian patients are different to Caucasian patients; some people respond well to the glitazone drugs, others don't; and so on. The differences could be environmental, but I think that genetics will play a large role. So it would be great if we could define categories of people who respond well to particular drugs and to make treatment more specific. If you could identify these categories straightaway using physiological or genetic tests, it would be an enormous step forward."

Further reading

• Stride A, Hattersley AT. Different genes, different diabetes: lessons from maturity-onset diabetes of the young. Ann Med 2002;34(3):207–16.
• Pearson ER et al. Genetic aetiology of hyperglycaemia determines response to treatment in diabetes. Lancet 2003;362(9392):1275–81.
• Shepherd M, Hattersley AT. ‘I don’t feel like a diabetic any more’: the impact of stopping insulin in patients with maturity onset diabetes of the young following genetic testing. Clin Med 2004;4(2):144–7.
• Gloyn AL et al. Activating mutations in the gene encoding the ATP-sensitive potassium-channel subunit Kir6.2 and permanent neonatal diabetes. N Engl J Med 2004;350(18):1838–49.
• Sagen JV et al. Permanent neonatal diabetes due to mutations in KCNJ11 encoding Kir6.2: patient characteristics and initial response to sulfonylurea therapy. Diabetes 2004;53:2713–18.