"It really is an epidemic," says Alan Saltiel, a diabetes researcher at the University of Michigan in Ann Arbor, "and it's on the rise." Because diabetes is linked with obesity, its incidence increases as people become older and less active. An estimated 143 million people worldwide have the disease—almost five times more than just a decade ago. In the United States, diabetes has been diagnosed in more than 16 million people. When glucose concentrations rise in the blood of healthy individuals—after a meal, for example—the pancreas produces the hormone insulin, which tells muscle and fat cells to extract the sugar from the blood. The insulin also instructs liver cells to stop producing glucose. In people with diabetes, however, glucose levels in the blood remain too high. This is true for both type 1 and type 2 diabetes, though for different reasons.

Four HHMI researchers have chosen different targets, hoping their findings will improve treatments for diabetes or provide clues for a cure. Graeme Bell is searching for culprit genes, Morris White and Morris Birnbaum are studying the cellular pathways of insulin and glucose and Gerald Shulman is investigating the impact of fat on diabetes development. Their work is part of a larger effort being made in scores of laboratories worldwide. Each new discovery of a component of any of these pathways provides a potential target for future drugs. "This is a really exciting time," says Saltiel. "We now have a lot of ideas for how to attack the disease."

Find Faulty Genes
Graeme Bell, an HHMI investigator at the University of Chicago, launched his career as a prospector for diabetes genes some 20 years ago at the University of California, San Francisco. Bell wondered whether mutations related to the human insulin gene might be involved in the development of diabetes—and he soon discovered a genetic variation associated with increased risk for type 1 diabetes. The variation, in a region next to the human insulin gene, turned out to modulate the gene's expression.

"It was a surprising discovery," Bell says, because type 1 diabetes was known to be caused by the outright destruction of the insulin-producing pancreatic beta cells by the immune system. "For 10 years, no one believed it," he recalls.

Although there is no longer any doubt that Bell had identified a gene that increases susceptibility to type 1 diabetes, Bell, Craig Hanis of the University of Texas Health Science Center at Houston and their colleagues now face similar skepticism over their discovery of a gene that appears to be linked to increased risk for type 2 diabetes, the more common type of the disease. Eight years ago, they turned their attention to a long-studied population of Mexican-Americans who are highly prone to the disease. In this population in Starr County, Texas, 2,500 of the 60,000 residents have type 2 diabetes; half of the adults over age 35 are either diabetics themselves or have a relative with the disease. Bell and Hanis obtained DNA samples from 330 pairs of diabetic siblings from 170 families and searched hundreds of locations throughout the genome for a link to disease susceptibility. By 1996, they had narrowed their search to a region of chromosome 2—but they still had to locate the gene and identify the variants that increased the people's risk for diabetes.

In October 2000, they finally announced their success. "People were incredulous," says Bell. Instead of finding a gene known to be involved in insulin action, the team had zeroed in on a gene encoding calpain-10—a protease, which is an enzyme that breaks down proteins. In people at increased risk for disease, the researchers discovered small genetic variations, called single-nucleotide polymorphisms (SNPs), in a region of the gene that controls how much calpain-10 is produced. As a result of the SNPs, these individuals have lower concentrations of calpain-10 in their muscle cells, and their cells are not as sensitive to insulin's commands. The mechanism by which calpain-10 affects how glucose is taken up and used by cells, however, remains unknown.

Bell is anxious to forge ahead. "Once you find a gene, you then have to show how it plays a role in disease." He and his colleagues are trying to learn how calpain-10 might influence insulin signaling or glucose metabolism, particularly in cells involved in diabetes, such as muscle, fat and liver cells.

Unblocking Signals
Morris White focuses his studies on the signals that are switched on and off inside the cell after insulin makes contact with the cell surface. An HHMI investigator at the Joslin Diabetes Center at Harvard Medical School in Boston, White is looking for molecules involved in insulin action and then attempting to learn how they become compromised in people with diabetes. In normal cells, the insulin binds to a specific protein on the cell's surface, called the insulin receptor, triggering a chain reaction in which one protein activates the next to pass insulin's signal deep into the cell. How the cell responds depends on what type of cell it is: Insulin directs muscle cells, for instance, to import and store glucose and tells liver cells to shut down glucose production. It might even alert the brain that enough food has been consumed. In people with diabetes, this signaling cascade fails at some point—but where, and how?

In the early 1980s, researchers discovered that when insulin binds to its receptor, the receptor takes on a particular chemical "tag" called phosphotyrosine. White and his colleagues exploited this tag when they set out to identify proteins involved in insulin action. Using the sort of brute-force biochemistry that was the only approach back then, the researchers ground up 200 rat livers—100 from animals that had been treated with insulin, another 100 from animals that hadn't. Then, using an antibody designed to recognize the phosphotyrosine tag, they set out to isolate any proteins that were tagged in the presence of insulin. As it turned out, the protein they were looking for was relatively scant in cells, making the search nearly impossible. After eight long years, however, they had purified enough of the insulin-activated, tagged protein to learn something about it. No one had seen anything like it before, and unfortunately, when White examined the protein more carefully, he could detect no obvious function. It appeared to have no activity at all.

At this point, White's efforts, like Bell's, were raising a few eyebrows. The protein, which he dubbed IRS-1 (for insulin receptor substrate 1), picked up a phosphotyrosine tag within seconds after cells were exposed to insulin. But how was it involved in insulin signaling? While White continued to ask this question, other researchers produced mice that lacked IRS-1, because sometimes the fastest way to figure out what a protein does is to see what happens when it's eliminated.

Mice without IRS-1 were small, but they never developed diabetes—not what one would expect if IRS-1 were critical to insulin's action. "People started asking me what I was going to do with my career," White recalls with a chuckle.

In 1992, Jacalyn Pierce at the National Cancer Institute told White about a line of blood cells that was resistant to insulin and was missing a protein that looked a lot like IRS-1. "It was just different enough to convince me that there was a second IRS protein," White explains. The researchers then succeeded in isolating and cloning that protein, which they called IRS-2.

Mice lacking IRS-2, they found, develop diabetes as they enter puberty, when insulin requirements increase. What's more, these mice show defects both in the way they respond to insulin and in the amount of insulin they produce—as do people with type 2 diabetes. To White, this meant that IRS-2 not only mediates insulin action in liver and muscle, it also appears to help the beta cells meet the body's need for insulin. "This was the most exciting moment for me in years, because it meant that a single problem might cause both insulin resistance and reduced insulin secretion—which leads to type 2 diabetes."

Interestingly, most people with diabetes do not show mutations in the IRS-2 gene. So a simple genetic explanation doesn't fit this puzzle. "Some other component in the insulin signaling pathway, perhaps calpain-10, or something not yet discovered, might break the IRS-2 branch of the pathway just enough to cause diabetes," White says. "Finding this defect may point to a relatively simple way to increase production of IRS-2, reduce insulin resistance and promote beta cell function and insulin secretion."

Enable Glucose Transport
White's research dovetails with that of Morris J. Birnbaum, an HHMI investigator at the University of Pennsylvania, whose interest in figuring out how insulin stimulates glucose uptake led him to the IRS-2 pathway. While a faculty member at Harvard Medical School in the late 1980s, Birnbaum cloned the gene for Glut4, the "transporter" protein that brings glucose into cells in response to insulin stimulation. He then set out to understand how insulin lures this transporter from inside the cell to the cell membrane, where it works to remove glucose from the bloodstream.

At Penn, Birnbaum focused on PI 3-kinase (phosphoinositide 3-kinase), a protein needed for insulin to stimulate glucose uptake. As it happens, PI 3-kinase is activated when it binds to the IRS proteins discovered by White. PI 3-kinase, in turn, activates a protein called Akt. When Birnbaum and colleagues tried adding an overactive form of Akt to fat cells, they found that it stimulated glucose uptake by causing Glut4 to journey to the cell membrane. "In fact, most of the things that insulin does were mimicked by putting activated Akt into cells," he notes.

To learn more about this protein's role in insulin signaling, the researchers also produced mice lacking the Akt2 gene, because the Akt2 protein is the form of Akt prevalent in insulin-sensitive tissue. As published in Science, June 1, 2001, the mutant mice showed marked insulin resistance and had elevated blood-sugar levels—the same problems experienced in type 2 diabetes. These data establish Akt2 as an essential gene in the maintenance of normal glucose balance, the authors wrote. Further studies of these mice, in collaboration with HHMI investigator Gerald Shulman, showed that their insulin could no longer shut down glucose production in the liver, another hallmark of diabetes.

"I think this is the first time anyone has proven that a downstream signaling pathway governs insulin action in a mammal," says Birnbaum. "People have been working on this for over 40 years."

Determine Fat's Role
Obesity and type 2 diabetes often go hand in hand—and both are on the rise. Gerald Shulman, an HHMI investigator and clinical endocrinologist at Yale University School of Medicine, is getting to the heart of what links weight gain with altered glucose metabolism and an inability to respond properly to insulin.

Most insulin-resistant individuals have elevated levels of fatty acids in their bloodstreams. To investigate how fat might cause insulin resis-tance in humans, Shulman and his colleagues used nuclear magnetic resonance spectroscopy to measure fatty acids, sugars and their metabolites in the tissues of two groups of human volunteers: one healthy, the other insulin resistant. People who are resistant to insulin had greater amounts of fat inside their muscle cells than did their healthy counterparts.

To go a step further and find out how fat might block muscle's response to insulin, they developed a method to measure the concentration of glucose and an intermediary metabolite, glucose-6-phosphate, in muscle. When healthy volunteers were given intravenous fat, they temporarily became severely insulin resistant and had lower levels of glucose and glucose-6-phosphate in their muscle cells. "These data challenge conventional beliefs by strongly suggesting that fatty acids cause insulin resistance in muscle by directly interfering with insulin activation of glucose transport activity," says Shulman. "This is the same step found to be defective in patients with type 2 diabetes."

How does fat interfere with insulin-stimulated glucose transport in muscle? Does it directly interfere with movement of the glucose transporter Glut 4 to the cell's membrane, or does it interfere with the insulin signaling cascade? To find out, the scientists examined insulin activation of PI 3-kinase, a key enzyme that stimulates glucose transport activity. Using muscle biopsies from the earlier subjects, they found that an intravenous infusion of fat totally abolished insulin activation of PI 3-kinase. Shulman's group teamed with Morris White to figure out how, and they discovered which steps in the insulin signaling pathway are blocked by excess fats (see figure).

What do these findings mean for diabetics? Because accumulation of fatty acids in liver and muscle has such a profound effect on insulin sensitivity, researchers could try to develop drugs that block their action. In an August 2001 article in the Journal of Clinical Investigation, Shulman's group and Steve Shoelson's group at the Joslin Diabetes Center showed that high-dose aspirin can prevent insulin resistance caused by high levels of fat in mice, and they determined where in the pathway the impact occurred.

More directly, people with type 2 diabetes can clear fatty acids from their own muscles: "Exercise and/or weight reduction is a great way to do this," notes Shulman. Meanwhile, he plans to continue moving back and forth between his studies of mice and of humans. "The transgenic mice are an incredibly powerful system for testing hypotheses," Shulman says, "but in the end, what you really care about is the patient with the disease."

The University of Michigan's Saltiel says he is encouraged by the many scientists who are tackling diabetes and the varied routes they are traveling. "No single approach will work," he says, and "no one researcher can move forward alone." Birnbaum, White, Shulman and others hope to identify all the molecules involved in the disease and to discover how these components act together to disrupt metabolism. With access to DNA samples from almost all of the 2,500 diabetic residents of Starr County, Bell is hoping to conduct an exhaustive search for all the factors—genetic and environmental—that contribute to type 2 diabetes in that population. With a closer look at calpain-10, he may discover that the enzyme somehow controls the activity of a protein in the IRS signaling pathway.

"What has emerged," says Bell, "is an appreciation that individuals can have diabetes for very different reasons." As the genetic work comes together with biochemical and physiological studies, the reasons for those differences should become clearer.

Photos: (from top) Dan Cohen, Todd Buchanan, Asia Kepka, David Graham, Gale Zucker

Download this story in Acrobat PDF format.
(requires Acrobat Reader)

Reprinted from the HHMI Bulletin,
December 2001, pages 16-21.
©2001 Howard Hughes Medical Institute

Gerald Shulman is getting to the heart of what links weight gain with altered glucose metabolism and an inability to respond properly to insulin. Gerald Shulman's research abstract Dr. Shulman's Faculty Page   How does excess fat cause insulin resistance?

Loss of Enzyme Produces
Diabetes-Like Symptoms
—June 1, 2001

Scientists Link Energy
Metabolism and Fertility
—September 21, 2000

Researchers Discover
New Genetic Culprit in
Type 2 Diabetes
—September 27, 2000