Type-1 diabetes not so much bad genes as good genes behaving badly, Stanford research shows

STANFORD, Calif. — Investigators combing the genome in the hope offinding genetic variants responsible for triggering early-onset diabetesmay be looking in the wrong place, new research at the StanfordUniversity School of Medicine suggests.

Early-onset diabetes, also known as type-1 diabetes, is an autoimmunedisease, caused when the immune system attacks and destroysinsulin-producing cells in a person's pancreas.

What triggers that immune response apparently has less to do with havinga distinct set of gene variants than how the behavior of genes maydiffer in people with the disease. That is the finding of a studypublished in the November issue of Clinical Immunology, by GarryFathman, MD, professor of immunology and rheumatology, and his colleagues.

The paper builds upon the knowledge that particularimmune-system-related gene variants confer type-1 diabetessusceptibility. Many people have those genes, but only a fractionactually develop the disease. This has led many investigators to conductexhaustive searches of the genome for other elusive genes that, whendefective, may predispose someone to type-1 diabetes. Fathman suggeststhey may be on the wrong track.

Fathman explained it this way: "Take a pair of identical twins, with onehaving type-1 diabetes. Although both have precisely the same genes,roughly half the time the other twin doesn't get the disease." The sameholds true for other autoimmune diseases such as multiple sclerosis andrheumatoid arthritis, he added.

The situation, Fathman said, is reminiscent of the 1988 movie "Twins,"starring Arnold Schwarzenegger and Danny DeVito. They may have startedout identical, but something diverged, somewhere. Fathman set out tofind out what it was seven years ago, in what he described, tongue-incheek, as "an interesting study that started at the dawn of history."

Rather than try to implicate a faulty gene, Fathman's team looked forgenes in the diabetic twin that act differently from the same genes inthe other twin who doesn't get diabetes.

To do this, the Stanford researchers used two types of bioengineeredmice that share a common genome, with just one key difference. So-calledNOD mice (the acronym stands for "non-obese diabetic") are extremelylikely to get type-1 diabetes and have an immune-related gene variantclosely resembling the one predisposing humans to the disease. The otherstrain of mice, known as NOD.B10, has had its chromosomal segmentcontaining the troublesome gene variant replaced with another, harmlessversion. NOD.B10 mice never get type-1 diabetes.

The Stanford team compared the activity level ("gene-expression," inscientific parlance) of each of the NOD mouse's genes - all 35,000 ofthem - with that of its counterpart in the NOD.B10 animals. To makethese comparisons as meaningful as possible, the researchers assembledthe mice into groups of three to 10 and took samples from varioustissues from each group at 10 days of age, then four, eight, 12, 16 and20 weeks, always comparing like tissues from one mouse strain to thenext at the same stage of life. This required the use of a sophisticatedbut increasingly commonplace hybrid between a microscope slide and acomputer chip - called a microarray - that can emit fluorescent signalscorresponding to the activity levels of each of the mouse's genes.

By comparing the strength of the signal from any given gene from aparticular tissue from NOD mice of a specific age to the correspondinggene in the NOD.B10 mice, it was possible to see which genes' activitylevels were turned up, or dialed down, throughout the course of diseaseprogression including the earliest stages. The NOD.B1O mice served ascontrols; by monitoring their tissues, scientists could determine anychanges in gene expression that were merely a matter of aging (20 weeksis a long time in the life of a mouse) or that merely reflectedcharacteristics of different tissues were ignored in the analysis.

The results, said Fathman, were surprising. Most genes in any giventissue of the diabetes-prone NOD mice at any given time showed about thesame activity levels as their disease-free NOD.B10 counterparts. But ineach tissue the scientists monitored, certain clusters of genes inNOD-mice - including many genes never previously identified as germaneto the disease process - seemed to participate in coordinated zigzags ofswooping and soaring expression over time, when compared with theirhealthy NOD-B10 "twins." These patterns varied from one tissue to thenext and from time to time. But in any given tissue at any given time,they were remarkably consistent.

These time-dependent gene-expression "signatures" could be observed inthe NOD mice's peripheral blood, for example, well before the mice beganto show characteristic signs of diabetes such as hyperglycemia. Fathmansaid preliminary work done in his lab indicates an exquisite similarityto the gene-expression signatures found in the blood of humans withtype-1 diabetes well before the onset of symptoms.

This finding may provide an early warning for pre-diabetics, Fathmansaid. "We need to know that people are on their way to diabetes beforethey get hyperglycemic or, better, even before their insulin-producingpancreas cells have taken a hit." Plus, the newly identified genes inthese clusters with orchestrated, disease-associated activity changesmay, in their own right, point the way to new therapies, he said.

One matter still unresolved is exactly why a diabetic identical twin'sgenes began acting differently from the non-diabetic twin's in the firstplace. Fathman believes this may be due to random differences in exactlywhich part of some invading or internal pathogen the immune systemresponds to, with one response setting off the diabetes-causinggene-activity cascade while another doesn't. His group is focused onunraveling this mystery.

Source: Stanford University Medical Center