Back-to-back papers in today's issue of Nature offerexperimental evidence on the onset of juvenile (insulin-dependent) diabetes PP and its possible prevention or arrest.

The first paper, titled "Spontaneous loss of T-cell tolerance toglutamic acid decarboxylase (GAD) in murine insulin-dependent diabetes," is by Daniel Kaufman et al of theUniversity of California at Los Angeles (UCLA).

The second, "Immune response to glutamic acid decarboylase(GAD) in non-obese diabetic mice" by Roland Tisch et al is fromthe laboratory of Hugh McDevitt of Stanford University.

UCLA has already licensed its pending patent on the potentialrole of the GAD enzyme in diabetes to Syva Co. of Palo Alto,Calif., for a diagnostic to detect the presence of GAD antibodies.The university is now seeking a licensee for GAD's potentialtherapeutic applications. Thus far, Stanford has not filed apatent application or made contact with industry.

The GAD enzyme plays a double role in cell-to-cellcommunication. In the brain, it synthesizes g-aminobutyric acid(GABA), the main inhibitory neurotransmitter of the centralnervous system. In the pancreas, GAD's GABA is thought to actas a signaling molecule in the islets of Langerhans. Theseharbor the b-cells, which secrete insulin on demand. GABAsupposedly lets the b-cells communicate with neighboring cellsinvolved in regulating the body's glucose levels.

When certain of the immune system's T cells mistakenly attackGAD proteins as foreign intruders, they destroy the b-cells andbring on insulitis PP inflammation of the islets. Eventually, withinsulin production shut down, insulin-dependent diabetesmellitus (IDDM) ensues.

But what causes those T cells to attack the body's own protein?Actually, GAD is only one of a whole platoon of antigens thatmarauding T cells erroneously perceive as autoimmune targetsin youngsters predestined by their genetic inheritance as beingprone to contract juvenile-onset diabetes. What causes this Tcell onslaught on "self" proteins?

UCLA's Kaufman speculates that because GAD is made mainlyin the brain, which the blood-brain barrier walls off from thebody's immune system, the immune defenses may not befamiliar enough with the enzyme to recognize it as "self," not"stranger."

Two years ago Kaufman and his group were the first to clonethe genes of GAD. They noticed that its molecule had a smallregion similar in structure to a sequence of the commonCoxsackievirus, which is linked epidemiologically with IDDM."Ten percent of normal children have Coxsackievirus infections,which cause flu-like symptoms, whereas 35 percent ofyoungsters at the onset of diabetes have the virus," Kaufmantold BioWorld. "Other researchers have injected Coxsackie fromIDDM patients into mice and monkeys, which promptly camedown with diabetes."

But this "very damning etiology," Kaufman added, doesn'tconvince all diabetologists. "Some argue that "people withdiabetes just have a higher rate of Coxsackie infection becausetheir immune systems are screwed up," he said. "And thosemice and monkeys caught diabetes from the virus because itreplicates in the very b-cells that the disease destroys."

Kaufman hastened to add that he doesn't mean to say thatCoxsackie is responsible for all diabetes, but "I would certainlybet my dinner it causes a percentage of cases."

To pin down the part played by GAD and other autoimmuneantigens, Kaufman's team recruited a strain of mice called NOD(non-obese diabetic). These rodents are also geneticallyprogrammed to incur IDDM early in life and are a commonanimal model for human IDDM.

Kaufman and his team tested NOD mice from birth to 28 weeksof age for T cell reactivity to b-cell antigens targeted by IDDMautoantibodies. Besides the two isoforms of GAD itself, theyincluded three other autoantigens PP carboxypepsidase H,insulin and an epitope of heat-shock protein.

As they reported in today's Nature, "proliferative T-cellresponses to these antigens developed spontaneously in asequential order. First, a response to GAD arose at four weeksof age, concurrent with the onset of insulitis." The cascade of Tcell antigen reactions followed at intervals till the NOD micewere 12 to 15 weeks old.

The UCLA experiments demonstrated that a potentiallypathogenic helper T-cell population is spontaneously set up toattack GAD early in the development of NOD diabetes, negatingthe immune system's innate tolerance to the antigen.

In an attempt to preserve this tolerance, the team "mademilligram quantities of GAD by recombinant techniques,"Kaufman recalled. They inoculated naive 3-week-old NODanimals intravenously with GAD protein (an injection routeknown to inactivate T cells that could recognize the protein)and control mice with other b-cell antigens. At 12 weeks of age,when the previously tested mice were showing diabeticreactions, "100 percent of the GAD-injected ones (but none ofthe controls) displayed no T-cell reactivity to GAD or the otherantigens and were completely free of insulitis."

The tolerized mice never developed diabetes, according to thereport. "All of the 17 GAD-treated animals, at 37 weeks of age,still had normal glucose levels, whereas 70 percent of the 20mice receiving control antigens developed hyperglycemia by19 weeks," it said.

Nor did it escape Kaufman's notice that "as a similarautoimmune progression may also occur in human IDDM and inother T-cell-mediated autoimmune diseases (involvingdifferent autoantigens), these findings should be useful in thedesign of immunotherapies."

Looking to the future, Kaufman surmised that tolerization bythe oral route will arrive. "For instance, if our antibody testdetected the presence of autoantibodies, and you've already gotthe process going, you might be able to give GAD by mouth andtry to suppress the immune response to it, just as Harvardrheumatologists did recently with collagen." (See BioWorld,Sept. 27).

Comparing UCLA's Nature paper with that of Stanford's RolandTisch, Kaufman said: "The bottom line that GAD is the importantantigen is the same. There are some subtle differences." Forinstance, he said that "they injected the thymus, where T cellscan also be tolerized, but that type of therapy is not veryapplicable to humans. We also show a greater degree ofprotection. Their therapy protected 70 percent of their NODmice, whereas 100 percent of ours did not develop diabetes."

Molecular immunologist Tisch, a post-doctoral fellow inMcDevitt's lab at Stanford, stressed, as did Kaufman, that muchwork remains to be done to determine if tolerizing healthy butgenetically susceptible human patients to autoantigens is afeasible possibility.

In a salient variation from the UCLA research, Tisch and hisgroup made NOD mice transgenic by transforming one of thegenes that endow them with the murine equivalent of humanjuvenile-onset diabetes. This gene encodes one of thehistocompatibility Class II molecules, which mediate cellularinteractions of the immune system. It consists of two chains,alpha and beta.

"Specifically, it's the beta that's unique to the NOD mouse,making it susceptible to diabetes," Tisch told BioWorld. He andhis colleagues swapped histidine and serine codons for prolineand aspartic acid in the beta chain gene and raised transgenicNOD mice that are free of diabetes.

Tisch appraised the Stanford and UCLA work as "actually verycomplementary. They carried out a lot of the same experimentswe did, and basically came out with the same results as we did.They did a few other things than what we did, and we didadditional experiments that they didn't do. So our workconfirms as well as complements one another."

-- David N. Leff Science Editor

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