Researchers discover major diabetes susceptibility gene
First success for positional cloning of a gene affecting susceptibility to a common disorder
September 26, 2000
In a finding that provides an enormous boost for scientists interested in either diabetes or genetics, researchers from the University of Chicago, the University of Texas-Houston Health Science Center, and several supporting centers have identified the major susceptibility gene for type 2 or non-insulin-dependent diabetes mellitus (NIDDM) in Mexican Americans. The gene also plays a role in at least two separate European populations.
The discovery, published in the October 2000 issue of Nature Genetics, pinpoints a new and unexpected biochemical pathway leading to diabetes and suggests novel approaches to prevention, diagnosis and treatment.
Identifying this new diabetes gene is also a major coup for genetics. This is the first time that a genome-wide approach has successfully led to the identification of a susceptibility gene responsible for a common, genetically complex disorder.
"Bell and colleagues' accomplishment is a 'tour-de-force'," said Allen Spiegel, MD, director of the National Institute of Diabetes and Digestive and Kidney Diseases. "Actually identifying susceptibility genes for diseases such as diabetes with 'complex' as opposed to simple Mendelian inheritance has proved exceedingly difficult."
"Moreover," added Dr. Spiegel, "identification of this gene--one that would not have leaped to mind as an obvious candidate--is exciting because it should lead to greater understanding of the pathogenesis of type 2 diabetes and possibly to new forms of treatment."
"Finding this gene finally allows us to go after diabetes treatment strategies that address the underlying molecular defects rather than the symptoms," said research team leader Graeme Bell, PhD, Louis Block Professor of Biochemistry & Molecular Biology and of Medicine and an investigator in the Howard Hughes Medical Institute at the University of Chicago.
"At the same time," Bell adds, "our unexpected success, and the techniques we developed along the way, should restore faith in the power of genetic approaches and provide the tools to identify the genes for other common genetically complex disorders, such as hypertension, obesity, psychiatric diseases or asthma."
Type 2 diabetes, or NIDDM, affects an estimated 135 million people worldwide, including more than 15 million Americans, almost six percent of the U.S. population and 18.4 percent of those over 65. Nearly 2,200 new cases are diagnosed every day in the U.S., for a total of 798,000 each year. That rate is steadily increasing. Diabetes is the seventh leading killer in the U.S., where the cost of the disease is estimated to exceed $98 billion each year.
In type 2 diabetes, the body either cannot produce enough insulin or does not use it effectively. If this goes untreated, glucose--a form of sugar--accumulates in the blood, slowly damaging the cardiovascular system, kidneys, eyes, and nerves. It is a leading cause of heart disease, stroke, nerve damage, and blindness, and the leading cause of amputations and kidney failure.
The disease is even more common in particular ethnic groups, affecting 10.6 percent of all Mexican Americans, 10.8 percent of African Americans and as many as 50 percent the Pima Indians, a Native American tribe based in Arizona.
Finding the genes responsible for type 2 diabetes has been unusually difficult because diabetes is not a single disease but a group of related disorders with similar symptoms. Although some unusual forms of diabetes are caused by specific single-gene mutations, the more common forms result from the interaction of multiple genetic and environmental factors, including obesity and lack of exercise.
Moreover, since the disease not only reduces life span but also is typically diagnosed after age 40, it has been difficult to gather the large multi-generation families normally used for genetic studies.
But by joining forces with Craig Hanis, PhD, from the Human Genetics Center at the University of Texas at Houston, Bell's team was able to bypass the need for multi-generation families.
Hanis and U.T. Houston colleagues have spent nearly two decades studying and working with a community of Mexican Americans in Starr County, Texas, a group with very high rates of diabetes. Hanis's team was able to provide extensive family histories, clinical data and DNA samples from 330 pairs of brothers and sisters affected with diabetes.
In 1996, a team led by Bell and Hanis was able to demonstrate linkage between an increased risk of diabetes in these Mexican Americans and an unknown gene located near one end of chromosome 2, which they called NIDDM1.
However, no one studying a complex disease had ever been able to take the next step, to identify the specific gene that had originally been localized by linkage.
"People kept telling us it was impossible," recalls Bell. But working with the Texas samples, Bell and colleague, statistical geneticist Nancy Cox, PhD, associate professor of human genetics at the University of Chicago, developed several new analytic techniques that enabled them to zero in on NIDDM1.
This statistical approach, along with increasingly detailed sequencing of multiple DNA samples to confirm or rule out candidate genes, led the authors to one minuscule variation within a previously unknown gene on chromosome 2 that neatly segregated with disease susceptibility.
The gene they tagged as NIDDM1 codes for a new protein, calpain 10.
The calpains are proteases--substances that regulate the function of other proteins by snipping off pieces, rendering the altered protein either more or less active.
Calpains are found in all human cells and throughout the animal kingdom. The researchers found some form of calpain 10 in every human adult or fetal tissue examined.
The calpain 10 gene consists of 15 functional segments, called exons. By splicing together these 15 pieces of the protein in slightly different patterns the body makes at least eight different versions of this protein. Different forms of the protein were found in different tissue types, including one form found only in the insulin-producing pancreatic islets.
The precise functions of most of the calpains remain unknown. Mutations of one of them--known as calpain 3--has been linked to a rare disorder called limb-girdle muscular dystrophy.
But Bell's team discovered a much subtler, and more surprising, variation. They found overwhelming evidence that the susceptibility-causing abnormality in the calpain 10 gene occurs not in one of the 15 exons that provide the blueprint for the different forms of calpain 10, but in an intron--a piece of non-coding DNA, sometimes referred to as 'junk DNA'--that separates the exons. This non-functional genetic matter is spliced out when the DNA is transcribed into RNA.
The tiny genetic change that can cause susceptibility to diabetes is located in intron 3, at a location dubbed UCSNP-43 (for University of Chicago Single Nucleotide Polymorphism-43). It is an alternative version of a single building block of DNA, shifting an adenine (A) to a guanine (G), changing just one of the three billion base pairs that make up the human genome. This minute alteration of the DNA acts in a recessive manner; individuals who inherit two copies of the "G" version have increased risk for diabetes.
"Finding a significant mutation in an intron is almost, but not entirely, unheard of," noted Bell. There are other examples. One was found in fruit flies, in the gene for an enzyme that breaks down alcohol. More recently a human intron variant has been discovered that is associated with decreased bone density.
Susceptibility to diabetes, however, turns out to be still more complex. In confirming the role of SNP-43, the researchers subsequently realized that two other genetic variations within the same gene--UCSNP-19 (in intron 6) and UCSNP-63 (intron 13)--act together with UCSNP-43 to affect risk. Two copies of the "G" allele at UCSNP-43 were required, but patients most at risk also had two different versions of the gene at sites 19 and 63.
With two slightly different versions of the gene at three sites, there are eight possible combinations; the researchers found only four in patients tested. If the two versions at each site are labeled "1" and "2," the most common combination in Mexican Americans was 112 on one chromosome and 121 on the other. The 112/121 combination was associated with a three-fold increased risk of diabetes. The 112/111 combination had no effect on risk. The 112/221 combination decreased risk.
The researchers suspect that these genetic variations work together to alter the way the gene is expressed in different tissues. They propose a "two-hit" model, where one allele of the high-risk version alters calpain 10 expression in the insulin-producing Beta cells of the pancreas, and the other allele alters expression in muscle or fat cells, tissues that either use or store glucose.
Because the discovery was so surprising and the techniques so innovative, the referees for Nature Genetics requested a series of supporting studies to help verify the finding, delaying publication by nearly 18 months.
One of those studies, a related paper to be published in the October issue of the Journal of Clinical Investigation, demonstrates the effect of the gene variation in a different ethnic group at extremely high risk for diabetes--the Pima Indians. Working with Bell's team, Leslie Baier, PhD, of the Phoenix Epidemiology and Clinical Research Branch of the National Institute of Diabetes and Digestive and Kidney Diseases, looked at the effects of calpain 10 in Pima Indians, who have the world's highest prevalence of type 2 diabetes.
They found that Pima Indians, without diabetes but with two copies of the "G" version of UCSNP-43, produced 53 percent less messenger RNA for calpain 10 in muscle cells. They also found that these not-yet-diabetic subjects have a series of metabolic abnormalities, including lower metabolism and increased insulin resistance, that resemble mild diabetes.
These metabolic effects are strikingly consistent with a 30-year-old theory that certain ethnic groups have evolved "thrifty" genes, which have historically protected them in times of famine but may endanger them in the current environment of nutritional overabundance. Pimas with the "G" version of the calpain 10 gene demonstrate many of the characteristics associated with a more frugal energy balance, such as decreased sleeping metabolic rate and a tendency to hoard rather than 'burn' glucose.
The genetic variation does not cause diabetes by itself, the authors emphasize. The calpain 10 gene interacts with lifestyle factors such as diet, exercise and other genes, particularly one on chromosome 15, to cause diabetes. The combination, "is not the whole story of genetic liability," said Bell, "but we think it accounts for about 14 percent of the 'population attributable risk' in Mexican Americans and about four percent in Europeans." If the Mexican American population lacked this diabetes gene, the prevalence of diabetes would be 14 percent lower.
Calpain 10 is just one member of this large family of similar genes. Bell's team is currently searching for other calpains that may play roles in disease susceptibility.
Understanding the mechanism of these subtle genetic tendencies should eventually provide targets for new medications, suggests Bell. Until then, the genetics of diabetes risk should help physicians discover which healthy young adults are most susceptible and begin to intervene.
"As with all disease genes," added NIDDK's Dr. Spiegel, "identification offers the prospect of genetic diagnosis, so those at increased risk can be identified early and prevention measures, such as appropriate diet and exercise, instituted."
Additional authors of the Nature Genetics paper include, from the University of Chicago: co-first authors (with Nancy Cox) Yukio Horikawa and Naohisa Oda, as well as Xiangquan Li, Manami Hara, Yoshinori Hinokio, Tom Lindner, Hirosata Mashima, Peter Schwartz, Laura del Bosque-Plata, Yohko Horikawa, Yukie Oda, Issei Yoshiuchi, Susan Colilla, and Kenneth Polonsky (now at Washington University); from the Virginia Mason Research Center in Seattle, Washington: Pat Concannon and Shan Wei; from Tokyo Women's Medical University: Naoka Iwasaki; from NIDDK, Phoenix: Leslie Baier and Clifton Bogardus; from the University of Lund: Sweden, Marju Orho-Melander and Leif Groop; from Technical University, Germany: Jan Schulze; and from the University of Texas Health Science Center at Houston: Eric Boerwinkle.
Authors of the Pima-phenotype paper include Bell, Cox, Horikawa and N. Oda from Chicago, and Leslie Baier, Paskasari Permana, Gong-Qing Shen, Xiaolin Yang, William Knowler, Richard Pratley, Robert Hanson, David Mott and Clifton Bogardus from NIDDK.
The Bell lab has a long history of finding diabetes genes. In 1990, they mapped MODY1, the gene responsible for an unusual form of early-onset diabetes, to a small region on chromosome 20. In 1992 they found MODY2, demonstrating that mutations of the gene for the enzyme glucokinase caused a different form of this early-onset diabetes. In 1997, the Bell lab found that patients from MODY3 families had one of several different mutations in the gene for hepatocyte nuclear factor 1alpha (HNF-1alpha). Finding MODY3 led to the rapid discovery of MODY1, a functionally related gene known as HNF-4alpha.
The NIDDK, the American Diabetes Association, the Juvenile Diabetes Foundation, the State of Texas, the Japanese Ministry of Health and Welfare, the Howard Hughes Medical Institute, the Blum-Kovler Foundation, an unrestricted grant for cardiovascular and metabolic research from Bristol-Meyers Squibb, and others provided support for this project.
For the first time, a genome-wide approach has identified a susceptibility gene for a common, genetically complex disorder. The researchers have identified the major susceptibility gene for type 2 diabetes in Mexican Americans, revealing an unexpected biochemical pathway leading to diabetes. This suggests novel approaches to prevention, diagnosis and treatment.