Genetic Protections

Diabetes does not have a simple, single genetic basis in the Mendelian sense (tall plants or short, blue eyes or brown, diabetic or not). Rather, it is a complex, polygenic disease. That is to say, in almost everybody who develops diabetes, several genes act together, with input from environmental factors, to bring it about.

New tools are enabling the systematic study of the genes that underlie the disease, and producing surprising findings. Comparing the genomes of people with and without diabetes, scientists at the Harvard-affiliated Massachusetts General Hospital and the Broad Institute of Harvard and MIT have identified 17 specific genetic variants associated with type 2 diabetes, says professor of genetics and of medicine David M. Altshuler, an endocrinologist and human geneticist who heads the Broad’s program in medical and population genetics. Before conducting the analysis, researchers at the Broad and other prominent programs that study the disease made a list of more than 500 “suspect” parts of the genome where they expected to find a correlation with diabetes based on earlier research. Not a single one came back positive. “What this tells me,” says Altshuler, “is there’s a lot of the biology of the disease that we don’t yet understand.”

The fact that type 2 diabetes has a far stronger genetic concordance in identical twins than type 1 does is not widely recognized. Someone whose twin has type 1 diabetes has a 30 percent chance of developing it himself; for someone whose twin has type 2 diabetes, the probability is “upwards of 80 percent,” says Altshuler.

He hopes genetic analysis will ultimately lead to therapies. It has already yielded intriguing hints about how the disease might work. Of the 17 diabetes “hot spots” identified, 11 were associated with decreases in insulin secretion—and not one was associated with insulin resistance. Type 2 diabetics have both insulin resistance and impaired insulin secretion; this finding implies that the latter plays a stronger role than the former in the progression to disease. In other words, the people who get diabetes are those whose beta cells cannot compensate by pumping out immense doses of insulin to compensate for insulin resistance—and individual genetic makeup strongly influences the body’s capacity to generate more insulin. This, too, indicates a closer similarity between type 1 and type 2 diabetes than previously recognized, since type 1 diabetes is characterized by complete breakdown of insulin production when pancreatic beta cells are destroyed.

Another group of researchers is looking beyond the genome to epigenetics: changes in the expression of genes independent of changes in the underlying DNA. Changes in the way DNA is packaged encourage or discourage gene expression, and here the intrauterine environment has powerful influence. Assistant professor of medicine Mary-Elizabeth Patti studies low-birthweight mice as a way to understand the strong correlation between low birthweight and obesity later in life, in both humans and mice. Patti was surprised to find that the offspring of the low-birthweight mice—that is, the grandchildren of the mice that were underfed during pregnancy—were also predisposed to diabetes, even though nutrition during their gestation, and their entire lifetime, had been normal.

Although this study brought about low birthweight through gestational caloric restriction, low birthweight can result when the mother’s health suffers in other ways—including hypertension and diabetes. So Patti’s findings mean that the current epidemic of metabolic disease could result, at least in part, from our grandparents’ life experiences. More unsettling is the potential impact on future generations.

But Patti’s research has yielded one bit of happy news. Restricting food intake for the low-birthweight mice, so that they eat no more than the baby mice whose weight at birth was normal, keeps the former from gaining weight as rapidly—and from becoming obese and diabetic later in life. “There is a period of plasticity when the organism is still sensitive to manipulation,” says Patti. “That’s the key point.” This research may have revealed a strategy for breaking this metabolic vicious cycle. But, she says, “We clearly need to test this in humans.”