Researchers led by a BBRF grantee report results of experiments that have potentially valuable precision-medicine implications for treating people with alcohol-use disorder (AUD).

AUD involves drinking behaviors marked by a loss of control over drinking, including an inability to stop drinking despite evidence of adverse impacts on one’s health and/or interpersonal relationships, or other consequences such as an inability to meet responsibilities at work, school, or in the family. The National Institute on Alcohol Abuse and Alcoholism estimates that about 1 in 10 American adults (27 million) had AUD in the past year.

The FDA has approved three medications for treating AUD, although for various reasons they are prescribed relatively infrequently. Naltrexone, acamprosate, and disulfiram can block the rewarding effects of alcohol, reduce cravings, or cause unpleasant reactions if alcohol is consumed.

Most adults with AUD do not receive treatment. Of those who are treated, many relapse. This fact makes the search for more effective AUD treatments an active area of research. At Vanderbilt University, a team led by Cody A. Siciliano, Ph.D., a 2017 BBRF Young Investigator, used mouse models of AUD to shed light on the effectiveness of one of the three U.S.-approved AUD treatments, naltrexone, compared with another quite similar compound called nalmefene which is approved for use in AUD in the European Union but not in the U.S. Nalmefene is approved for other indications in the U.S., and is sometimes prescribed for AUD off-label.

The relative effectiveness of the two drugs is the subject of intense debate, and has influenced the decision about whether nalmefene should be approved for AUD in the U.S. As approved by the European Medicines Agency in 2013, nalmefene treatment aims to reduce the amount of alcohol consumed, as opposed to naltrexone, which is used to maintain abstinence from alcohol. Naltrexone is taken daily, while nalmefene is typically taken “as needed” on days when alcohol is expected to be consumed or when craving is high.

While they are prescribed in Europe for different purposes, the two drugs have similar chemical structures and both target multiple receptors for opioids. They differ in how they act at different opioid receptor subtypes (there are 4 primary cellular receptors subtypes for opioid molecules), as well as in their pharmacodynamic action. At the kappa opioid receptor subtype, naltrexone acts as an antagonist (blocking activity there), while nalmefene acts as a partial agonist, meaning that it will increase activity when receptor activation is low and decrease activity if receptor activation is high.

The two drugs have not been compared in a head-to-head clinical trial in humans, although each has been tested separately. These trials have yielded widely disparate outcomes, fueling controversy about which is more effective. Dr. Siciliano and his lab, including co-first authors Zarha Farahbakhsh and Alex Brown, as well as Suzanne O. Nolan, Ph.D., 2024 BBRF Young Investigator, and Snigdha Mukerjee, Ph.D., 2022 BBRF Young Investigator, used mouse models of AUD to do a head-to-head comparison of naltrexone and nalmefene. They did this to test a hypothesis that the two drugs might be effective in different AUD subpopulations—in mice, and perhaps also in people. The results were reported in Communications Medicine.

A cohort of 56 male mice was tested. All were conditioned to seek ethanol, effectively modeling AUD in people. Each animal was tested behaviorally under four conditions: after treatment with either naltrexone, nalmefene, or a third drug called U50,488, or a saline placebo. “Wash-out” periods separated each of these tests. In each “block” of tests, each animal, 30 minutes after receiving one of the three drugs or placebo, was allowed to self-administer ethanol freely in one session, and in two subsequent sessions was offered the opportunity to self-administer ethanol that had been mixed with an unpleasant substance (quinine). The latter tests the degree to which an animal accustomed to drinking and seeking ethanol will drink despite adverse or unpleasant circumstances.

The researchers found that naltrexone and nalmefene were about equally effective in reducing ethanol consumption, when the entire population of tested animals was taken into account. (The third drug, U50,488, actually increased ethanol consumption; placebo had no impact on drinking). What stood out was the fact that some animals (~60%) responded only to naltrexone, while the others (~40%) responded only to nalmefene.

“Importantly, these are two distinct subpopulations of individuals,” the team reported. “Nalmefene responders show no response to naltrexone and vice-versa. Our results suggest that if prescribed using a personalized approach, utilizing both compounds may have an additive effect on overall treatment outcomes, even if there is no clear difference in efficacy when each is assessed in isolation.”

This assumes, of course, that results in mice translate to people with AUD. With this in mind, the team used two approaches in the mice to find a way to predict whether any given individual would respond to nalmefene or naltrexone.

One method was a biomarker test based on a simple blood draw; an accuracy of 80% was attained in determining preferred treatment group status on the basis of compounds detected in the samples. In people, a blood-based screening approach, the team said, would require creation of a database including sample constituents and clinical outcomes to train a predictive model.

The other method was based on drinking behavior. This is based on observations in the AUD mouse models of individuals who were classified as “compulsive drinkers” (for whom nalmefene was effective) vs. “low drinkers” (who responded well to naltrexone). This method could potentially be implemented based on patient self-reports of drinking habits, the team said.

They speculated that each behavior type might reflect differences in underlying biology and specifically in levels of activity at kappa opioid receptors, which may vary with these behaviors in different individuals. They cited previous research suggesting that chronic ethanol intake induced upregulated (increased) kappa opioid receptor activity, which drives an escalation of intake.

In the team’s view, “both methods [blood test and behavioral classification] provide a tractable starting point” for future precision treatment strategies.

Their study, they wrote, “provides preliminary evidence that differences in kappa opioid receptor signaling are associated with individual differences in AUD-relevant behavior and may contribute to treatment response.” In the team’s view, this provides a framework for other animal research to contribute to a mechanistic understanding of differing responses to available AUD treatments.

In future research, there is a pressing need to conduct similar tests in female mice, the team noted. They chose male mice for this preliminary research because the response of the kappa opioid receptor is well understood in males but not in females. They also cautioned that neither their AUD mouse model nor any other model can fully recapitulate the full spectrum of AUD-like behaviors in people. Yet if their research is validated, it would be an example of comparatively low-cost animal research generating findings that can directly help accelerate precision-medicine approaches in people.