(AAAS) -- Why does stress make some people reach for a drink, but not others? Variations in a key stress-response gene may be at least one reason, suggests a German study on mice. The scientists report their findings in the journal Science, published by the American Association for the Advancement of Science.

Rainer Spanagel of the University of Heidelberg and the Max Planck Institute of Psychiatry found that mice lacking the "CRH1" gene drank more alcohol than normal mice did after stressful experiences.

If humans with variations in this gene behave the same way, a fairly simple test may identify recovering alcoholics likely to relapse under stress, according to Spanagel.

"We think that, with our model, we have the neurobiological mechanism underlying a very specific phenotype of alcoholic patient," Spanagel said. "Patients with alterations in this gene may be particularly susceptible to stress, and may respond with drinking."

Psychologists might then be able to help these patients, Spanagel suggested, teaching them coping strategies for stressful situations.

"We've long known that stress is the biggest cause of relapse for many, but not all addicts, even those who've long been in recovery. This important study points both to the underlying mechanisms of this effect and potential targets for prevention and treatment efforts," said Alan Leshner, Chief Executive Officer of AAAS, and formerly Director of the National Institute on Drug Abuse.

Researchers generally agree that stress-induced alcohol drinking and the tendency to relapse seem to have a significant genetic component, but the connection is not well understood. Some studies have implicated a signaling network in the brain, called the CRH system, that regulates hormonal and behavioral responses to stress.

Signals in this system travel between neurons via two types of cell receptors, named CRH1 and CRH2. Spanagel and his colleagues investigated the CRH1 receptor, which has been linked to a variety of stress-related psychiatric disorders.

The researchers compared normal mice with a strain of mutant mice that lacked the gene responsible for making the CRH1 receptor. They first offered both groups a free choice between water and ethanol (drinking alcohol) solutions at different concentrations, as drinking fluids. Both groups opted for a solution of eight percent alcohol.

The authors exposed both groups of mice to two types of stress. One involved an attack by an unfamiliar mouse, and the other required the mice to swim on three consecutive days. After each experience, all the mice continued to drink the normal amount of alcohol.

Three weeks later, however, the mutant mice began drinking approximately three times more alcohol than the normal mice.

The mutant mice continued to drink significantly more alcohol six months later. Three months after the tests, the normal mice were still drinking the normal amount.

When the brain's CRH signaling system is not working properly, it's thought to affect other signaling systems, particularly the glutamatergic system. Neurons in this network communicate with each other using chemical signals sent via glutamate receptors.

To see if stress had affected the glutamatergic system somehow, the authors checked both groups of mice for differences in certain glutamate receptors. They found that, the neurons of the mutant mice contained more proteins called NR2B subunits, which form part of a particular type of glutamate receptor, the NMDA receptor.

The increase occurred in two brain regions that are relatively sensitive to alcohol. The nucleus accumbens is the most important part of the brain's reward system; all re-enforcing properties of alcohol come from here. And, the hippocampus is important for memory and environmental conditioning.

While the details aren't clear yet, Spanagel's team thinks that stress causes neurons to express more NR2B subunits, which are alcohol sensitive.

Spanagel isn't sure why it took three weeks for the mutant mice to increase their alcohol consumption, but he pointed out that previous studies have shown whole gamut of responses to stress, from drinking more, to drinking less, and including both delayed and immediate responses.

"We think there are different phenotypes of alcoholism, with different underlying mechanisms," Spanagel said.

While more research will be necessary to apply these findings to humans, Spanagel is fairly confident about his mouse model.

"Mimicking addictive behavior in animals is much easier than mimicking depressed or schizophrenic behavior," he said. "There is a pretty good link, in comparison to other disorders."