You are here

Project Spotlight
Prognosticators in our blood: The search for biomarkers of diabetes

A drop of blood could hold hundreds of clues. 
Image courtesy of ©istockphoto/Mark Hatfield

Imagine taking a snapshot of hundreds of molecules contained in a drop of blood. The levels of vitamins and amino acids, cholesterol and triglycerides, glucose and insulin, and more could help paint a picture of the current health state of the person whose blood was drawn. But they might also show something more. What if a snapshot taken today could be used to predict if a person will be diagnosed with a disease a dozen years from now? Not only could such a discovery allow a potential patient to change lifestyle, habits or medications before the disease progresses, but it could also help scientists understand a disease in ways that have never been possible before.

The idea of finding telltale markers in the blood long before the onset of disease has tantalized researchers for years. Recently, two critical resources have come together to help make this possibility a reality. One is a relatively new platform at the Broad. The other is a project that has been going for more than sixty years: the Framingham Heart Study.

Thomas Wang, a cardiologist at Massachusetts General Hospital (MGH), began working on the Framingham Heart Study when he was a research fellow, but the project had already been going on for decades. The study began in 1948 with over 5,000 men and women from the town of Framingham, Massachusetts. Those enrolled in the study came in for a physical exam, lab tests, and the collection of a detailed medical history. They returned every two years to repeat these examinations. Several years later, their children were included and then, a third generation – the grandchildren of the original cohort – was also enrolled and continues to come back every few years for follow-up exams. This kind of large, longitudinal study in which people are carefully followed over a long stretch of time is rare but incredibly valuable to researchers like Wang. That is because it offers an unparalleled view of how diseases like stroke, heart disease, and diabetes can emerge over time.

Robert Gerszten (left) and Thomas Wang (right) brought the
power of the Framingham Heart Study and metabolite profiling together.
Image by Broad Communications

Robert Gerszten, a senior associate member at the Broad Institute and director of Clinical and Translational Research at the MGH Heart Center, is equally interested in understanding how these diseases progress and finding ways to predict and intervene early. Gerszten had looked at blood from smaller groups of patients for biomarkers of disease – molecular bellwethers that signal a change in health. Gerszten, one of the creators of the Broad’s Metabolite Profiling Initiative, has worked with colleagues to probe the blood for biomarkers that change immediately after exercise and others that can be used to diagnose a heart attack quickly.

As Gerszten and Wang discussed their research interests, the idea to bring together the samples from the Framingham Heart Study and the Broad’s emerging metabolite-profiling technology began to take shape. Interested in the nexus of cardiac and metabolic diseases, the researchers chose a disease to investigate: diabetes.

Diabetes: Metabolism gone awry

Type 2 diabetes is a lifelong disease that affects millions of people. And, like obesity, it is on the rise in the United States. Diabetes is both costly and dangerous – it is the seventh leading cause of death in this country and its complications include heart disease, stroke, high blood pressure, blindness, kidney disease, nervous system disease, and amputation. In 2007, diabetes cost the United States approximately $174 billion dollars.

Normally, the body can break down food into simple components — such as sugars and amino acids — that it can use or store as fuel. In people with diabetes, abnormal chemical reactions in the body disrupt this process. For unknown reasons the body stops responding correctly to insulin, a hormone that controls how sugar is used, so high levels of sugar build up in the bloodstream. Although the root causes of these issues remain unknown, it is clear that diabetes is a disease of metabolism in which molecules like insulin play a key role. “There is no more quintessential metabolic disease than diabetes,” said Gerszten.

Type 2 diabetes is not inevitable. In a study known as the Diabetes Prevention Program (DPP), participants who were overweight and had higher than normal blood glucose levels decreased their risk of developing diabetes by 58 percent by making lifestyle changes such as losing weight and adhering to a diet low in fat and calories. Those at high risk who took the diabetes drug metformin were also able to reduce their risk of disease. David Nathan, Director of the General Clinical Research Center and of the Diabetes Center at Massachusetts General Hospital, and chair of the DPP study, said that these results raise the question: how early can we identify people at highest risk for the disease?

“Having earlier biomarkers for those at risk would be very useful because then we could apply these intervention strategies in a more selective way,” he said. “If we could actually identify those people who are at highest risk of developing diabetes, then we could start developing more focused interactions.”

The tools to get the job done

Clary Clish joined Gerszten and Wang in the search for predictive biomarkers of diabetes when he joined the Broad Institute two years ago. For many years, Clish had been studying the patterns of metabolites – the rise and fall of molecules naturally found in the body such as triglycerides, glucose, vitamins, and amino acids. Drawing upon his previous experience, Clish helped make the metabolite profiling efforts at the Broad robust enough to accommodate the volume of samples available from the Framingham Heart Study.

Clary Clish leads the Broad's Metabolite Profiling Initiative,
which uses machines like the HPLC to scour metabolites.

Image by Broad Communications

With the help of Clish and the rest of the Metabolite Profiling Initiative, the researchers have been able to look at approximately 250 metabolites in blood plasma (the yellow-colored, liquid component of blood) for hundreds of samples. These molecules span an intricate web of interconnected pathways. Clish has a wall chart of known metabolic pathways – a complicated image with arrows connecting thousands of molecules, the names of which are squeezed onto the large chart in tiny type.

“The goal is to cover as many of these pathways as we can,” said Clish. This is especially challenging since some of these molecules are intermediates, existing for only a short period of time before being converted into something else. However, the connectivity of these molecules makes the search feasible. Instead of measuring every metabolite, one molecule can serve as a read-out for a whole metabolic pathway.

By looking across the chart of pathways, the researchers are shining light into all of the corners of metabolism in their search for signals that may predict diabetes. “When people think about diabetes, they always think of glucose metabolism,” said Clish. “But this gives us a chance to bring light to the other pathways that may not have been previously appreciated.”

Clish and his colleagues have been adjusting and perfecting tools to precisely measure molecules that are abundant as well as those that are present only in tiny amounts. “We’re still pushing the limit of the technology,” Clish explained. The researchers must carefully tune their machinery to measure molecules, much like one must carefully tune a radio to hear a distant broadcast. Many of the variables are left to the control of the platform’s scientists who try to attain breadth of coverage, stability, and precision.

Journeying back in time

The Framingham Heart Study spans six decades and three
generations of participants.

Image courtesy of the Framingham Heart Study Archives

In previous studies, scientists had compared biomarkers from the blood of patients with diabetes to the blood of those without the disease. But by the time diabetes is diagnosed, it is difficult to distinguish metabolic changes that are there due to complications from those that are due to the root cause of diabetes. “You want to look before any of the complications are there,” said Gerszten. “The earlier you look, the more likely the metabolites you find are closer to the cause.” By the time the disease is full blown, those early signals may be washed out.

Diabetes is diagnosed when a person’s blood sugar levels are tested (either after fasting, after consuming a glucose-containing beverage, or at random). But glucose intolerance may be the most obvious consequence of an initial derangement in metabolism. Gerszten compares this to looking for biomarkers in the bloodstream of a patient with heart disease. “The real problem was a cholesterol build up that occurred fifty years earlier,” he said. “What you want to do is to look long before they have a heart attack.”

One of the advantages of the Framingham Heart Study is that participants’ blood has been collected and stored for years. Additionally each patient has had his or her genome scanned. Since the 1990s, the subjects have been tested for diabetes using glucose tolerance tests. If a person in the study is diagnosed with diabetes today, researchers can look at a sample of the person’s blood drawn over a decade ago. They can look at snapshots of metabolites in the intervening years to see what kinds of changes take place long before diagnosis. “The real power of the Framingham Heart Study is that you can look back in time before a person had the disease,” said Eugene Rhee, a renal fellow at MGH and visiting postdoctoral scholar at the Broad Institute who has been working on the project for the last two years. “That’s very different than taking blood from people who already have diabetes and those that don’t.”

Another advantage to this strategy is that a person can serve as his or her own control. Metabolites may vary greatly from person to person so instead of comparing snapshots from two different people, the researchers can compare snapshots taken from the same person at different times.

The path forward

Researchers can use the metabolite data they have gleaned from the Framingham Heart Study to investigate many other diseases, including stroke, peripheral vascular disease, hypertension, and kidney disease. Rhee, who specializes in kidney function, is especially interested in how kidney disease affects human metabolism. “Diabetes is our flagship and first project,” said Gerszten. “But the beauty of the Framingham Heart Study is that because both phenotypic and genetic information has been collected, we can now use the data to study many, many diseases, including liver disease, heart disease, and kidney disease.”

The researchers have also found collaborative partners in Malmo, Sweden where a similar study of patients over many years has been conducted. By sharing their data, the scientists have been able to compare and confirm their preliminary results.

The ultimate goal of the project is to find biomarkers of disease that not only accurately predict diagnosis, but also illuminate new pathways that could help scientists develop better preventative drugs. “That’s our long-range hope,” said Wang. “We hope that insights gained from metabolite profiling will contribute to both prediction and prevention.”

To view other Projects visit the Project Spotlight Archives.