A billion dollars here, a billion there, before you know it you'retalking real money.

Type II diabetes costs the U.S. economy some $85 billion dollars ayear in health care costs. Worldwide 100 million people have thisdisease, a.k.a. non-insulin-dependent diabetes mellitus, or NIDDM.

As baby boomers edge toward 60 and 70 years of age, more andmore Americans will contract NIDDM. Typically, these are thedecades of life in which Type II diabetes is diagnosed.

Research diabetologists and molecular geneticists have long huntedfor the gene or genes that cause Type II diabetes, but so far in vain.Once they localize the inherited mutations of the elusive DNAsequences responsible for the disease, they can begin to visualizedrugs for treating the disease at its root cause.

In recent years, gene hunters have hit jackpots with two familialdiseases of major impact, breast cancer and colorectal cancer. Theirlinkage analysis method was to assemble large numbers of extendedfamilies in which large numbers of members were afflicted, andscanning the genomes of their chromosomes for configurations ofDNA that occurred in individuals with the disease but not in theirnormal relatives.

This approach has motivated a Swedish clinical and researchdiabetologist named Leif Groop to spend the past six years thuscollecting data on a population of Swedish-speaking Finns whoinhabit a coastal strip of Finland on the Gulf of Bothnia (Botnia inSwedish).

That enclave's 60,000 isolated inhabitants is thought to have settledthe area over 1,000 years ago. Such a self-contained, geneticallyhomogeneous population, with a two to three percent prevalence ofType II diabetes occurring in large families, offered Groop an idealhunting ground in which to seek an NIDDM gene or genes.

Groop, at Lund University in Malmo, Sweden, is co-senior author ofa paper in the September 1996 issue of Nature Genetics. Its title:"Mapping of a gene for Type II diabetes associates with an insulinsecretion defect by a genome scan in Finnish families." Its co-seniorauthor is Eric Lander of the Whitehead Institute for BiomedicalResearch in Cambridge, Mass. They represent the two poles of anaxis _ data collection and genome scanning _ that has moved theNIDDM gene hunt off dead center.

"We had known for some time that diabetes Type II is a geneticdisease," molecular geneticist Melanie Mahtani told BioWorldToday. A post-doctoral fellow in Lander's laboratory, she is thearticle's lead author.

"That doesn't mean that NIDDM is solely genetic," she continued,"but it does mean that in families that have patients with diabetes, thedisease tends to cluster, so it will be found in certain families and notothers. So we did think we'd have a very good chance of findinggenes in Groop's collected family data."

4,185 Subjects; 400 DNA Genomic Markers

Based on Groop's four to five thousand family histories, the teamconducted oral glucose tolerance tests (OGTT) of 1,180 NIDDMpatients and 3,005 other family members.

These OGTTs were nothing like the fasting or two-hour blood testsof glucose that doctors routinely perform in their offices. In FinnishBotnia, the participants gave blood at 30, 60 and 120-minute timepoints, after swallowing a dose of glucose. Moreover, the testmeasured not only their bodies' ability to metabolize the sugar, butalso _ and critically, as it later turned out _ the actual levels ofinsulin their islet beta cells secreted.

"The other issue in our strategy," Mahtani said, "was that we hadaccess to fairly dense genetic maps of the human genome. Thatmeans we were able to scan close to 400 DNA microsatellite markersspaced evenly across the genome."

The Whitehead team contracted out a quarter of the total genotypingto Millennium Pharmaceuticals Inc. of Cambridge, Mass. "We beganan academic collaboration with Millennium about a year ago,"Mahtani recalled, "to develop our mutual methods of genotyping. Wewere two good groups working on it."

Clear Genetic Path Found

That meant basically tracking the segregation (inherited genevariation) of the chromosomes through family members who have thedisease, and correlating what chromosomes those patients share withtheir other siblings or family members who also have diabetes.

To obtain a population genetically "enriched" for NIDDM, the teamnarrowed the field down to "a core subset of 26 families that lookedvery genetic," and fit a specific set of criteria.

"We wanted families that had three or more diabetic patients,including one with an onset somewhat earlier than the mean 58 yearsin the Botnian cohort," she explained. "This was to ensure that thediabetes was going to be biased toward something genetic."

"The idea," she went on, "is that as a patient ages, his or her chancesof becoming diabetic increase. So it's not clear then whether that's anenvironmental effect, such as diet or exercise. So we chose familieswith one member who got diabetes early.

"In those 26 families," Mahtani said, "we genotyped everybody andasked: What genomic traits do all the diabetic members of a familyshare? What do they have in common?"

The answer proved disappointing.

"We had expected to see that the bulk of the families, maybe 20 or soof the 26, would all share a region of the genome or onechromosome. We didn't have it. Rather, we found," Mahtani added,"that the affected members of those families shared different regionsof the genome."

This setback led the team to re-evaluate its approach. "Maybediabetes is too complex a disease to analyze all at once," theyreasoned. "Maybe we need to subdivide it."

"So we then classified our patients by how much insulin they hadsecreted during this OGTT," Mahtani went on, "and that turned outto be the critical stage in finding this NIDDM gene. We divided themup into four quartiles: low insulin levels, medium, a little higher, veryhigh _ and it actually did work."

She continued, "It turns out that familes with the lowest levels ofinsulin secretion have a genetic basis for their diabetes; it seems to becaused by a gene on human chromsome 12."

What's more, a second gene, for an extremely rare variant form ofType II diabetes called MODY (maturity onset diabetes of the young)apparently maps to the same site on chromosome 12. "We argue,"Mahtani said, "that this may be two different mutations of the samegene."

What's next on everybody's agenda?

"I think the first question," Whitehead's Lander told BioWorldToday, "is finding the genes involved in diabetes.

"Beyond that," he continued, "will be using the knowledge of thepathways that are disrupted in diabetes to be able to definetherapeutics." n

-- David N. Leff Science Editor

(c) 1997 American Health Consultants. All rights reserved.