Oral GLP-1 drugs reach the brain. Fifteen trials are now targeting addiction.

Beyond the gut: where GLP-1 receptors actually live

Glucagon-like peptide-1 (GLP-1) is an incretin hormone released primarily from L-cells in the intestinal mucosa in response to food intake. Its peripheral actions are well characterized: it stimulates pancreatic insulin secretion in a glucose-dependent manner, suppresses glucagon, slows gastric emptying, and signals satiety to the hypothalamus via the vagus nerve and bloodstream [1].

What gets less attention in the public discussion of semaglutide, liraglutide, and the new oral agents is the brain distribution of GLP-1 receptors (GLP-1Rs). They are not confined to the hypothalamus. GLP-1Rs are expressed in the ventral tegmental area (VTA), the nucleus accumbens, the prefrontal cortex, the amygdala, and the hippocampus [2]. These are not peripheral metabolic regulators. These are the core nodes of the mesolimbic reward pathway — the circuitry that determines whether something feels rewarding enough to seek out again.

The presence of GLP-1Rs in reward circuits was documented as early as the 1990s in animal histological studies. The functional significance — that pharmacological agonism of these receptors could modulate addictive behavior — took longer to establish. The timeline from receptor mapping to clinical trial is now closing rapidly.

The dopamine mechanism — how GLP-1 suppresses craving

The mesolimbic dopamine pathway runs from the VTA to the nucleus accumbens. Activation of this pathway by drugs of abuse, alcohol, or rewarding stimuli releases dopamine in the nucleus accumbens — the neurochemical signal that the brain encodes as "this was good, do it again." Repeated activation of this circuit by addictive substances progressively sensitizes the dopamine release response to drug cues, which is what craving is at a neurochemical level.

GLP-1 receptors in the VTA and nucleus accumbens appear to modulate this dopamine release signal. When activated, they reduce the magnitude of drug-cue-induced dopamine release — not by blocking dopamine receptors directly, but by reducing presynaptic dopamine output from VTA neurons [3]. The pharmacological effect is a dampening of the craving signal, not an elimination of dopamine function.

Preclinical data established this mechanism across multiple models. Erreger and colleagues demonstrated that the GLP-1 receptor agonist exendin-4 reduced amphetamine-induced locomotor activity in rats — a standard proxy for reward sensitization [4]. Subsequent work from multiple independent groups confirmed that GLP-1R agonism reduces self-administration of alcohol, cocaine, and nicotine in animal models, with effects mediated specifically through central GLP-1Rs rather than peripheral satiety signaling [5].

The translation question — does the mechanism hold in humans — is now being answered.

GLP-1 drugs are not just changing what people eat. They appear to be changing what people want — and the science behind that statement is more solid than most people realize.

The alcohol data — RCTs and observational findings

The human evidence for GLP-1 RAs and alcohol is the most developed of the addiction verticals. The early signal came from observational reports: patients prescribed semaglutide or liraglutide for weight loss or diabetes noted, unprompted, that they were drinking less. This is the same pattern that preceded the weight-loss discovery — patient-level observation before any trial was designed to detect it.

Klausen and colleagues published the first dedicated randomized controlled trial of a GLP-1 RA in alcohol use disorder (AUD) in 2022 in JCI Insight [6]. Participants with AUD were randomized to once-weekly exenatide or placebo over 26 weeks alongside cognitive-behavioral therapy. The trial did not meet its primary endpoint — reduction in heavy drinking days in the overall population was not statistically significant. However, exenatide significantly attenuated fMRI alcohol cue reactivity in the ventral striatum, and in a pre-specified exploratory subgroup of participants with obesity (BMI > 30), exenatide significantly reduced both heavy drinking days and total alcohol intake — suggesting the reward-circuit effect may be most pronounced in that phenotype.

A 2024 retrospective analysis of insurance claims data published in Nature Communications found that patients prescribed semaglutide had significantly lower rates of new AUD diagnoses and AUD-related hospitalizations compared to matched controls on other anti-obesity medications [7]. This is observational and subject to confounding, but the effect size was large enough to be clinically meaningful and consistent with the mechanistic hypothesis.

Leggio and colleagues at NIH's National Institute on Alcohol Abuse and Alcoholism (NIAAA) have been among the most active investigators in this space, publishing a series of reviews and mechanistic papers arguing that GLP-1 RAs represent a genuinely novel class of anti-craving pharmacology — distinct in mechanism from existing AUD pharmacotherapy (naltrexone, acamprosate, disulfiram) [8].

Nicotine and opioids: the emerging evidence

The alcohol signal is the most mature. Nicotine and opioids are following on slightly shorter timelines, with results beginning to emerge.

For nicotine, the early data comes primarily from the same observational channel: patients on GLP-1 RAs reporting reduced cigarette cravings. A 2024 analysis using the TriNetX health research network — covering millions of electronic health records — found that patients newly prescribed semaglutide had significantly reduced rates of smoking cessation failure compared to patients on other weight management medications, and reduced rates of tobacco use disorder diagnoses at one-year follow-up [9]. This is not trial data, but the effect size suggests a pharmacologically meaningful signal.

For opioids, the data is primarily preclinical, but the mechanism is consistent. GLP-1R agonism in rodent models reduces conditioned place preference for morphine (a standard measure of opioid reward) and attenuates naloxone-precipitated withdrawal signs in morphine-dependent animals [5]. The human trial data for opioid use disorder (OUD) is lagging the alcohol data by roughly two years — with multiple trials now recruiting that should produce phase 2 results by 2026–2027.

Why oral formulation matters for CNS access

Injectable GLP-1 RAs (semaglutide injection, liraglutide) achieve therapeutic plasma levels that cross the blood-brain barrier to a degree sufficient to produce central effects. But they do so as large, peptide-based molecules with relatively limited CNS penetration compared to small molecules designed for oral administration.

Orforglipron is an oral, non-peptide GLP-1 receptor agonist developed by Eli Lilly — the first in its class to complete phase 3 trials. As a small molecule, orforglipron crosses the blood-brain barrier more readily than the injectable peptide-based formulations. NIH-backed preclinical work from the University of Virginia demonstrated that orforglipron achieves meaningful CNS concentrations at therapeutic plasma levels — producing GLP-1R activation in the VTA, nucleus accumbens, and prefrontal cortex at exposures consistent with human dosing [10].

This matters for addiction pharmacology because the efficacy of GLP-1 RAs in suppressing craving is likely dependent on direct central receptor engagement, not merely peripheral metabolic modulation. If oral small-molecule GLP-1 RAs have better CNS penetration than injectable peptides, they may prove more effective for addiction applications specifically — even though injectable formulations have a larger existing evidence base.

The current trial landscape: what's recruiting now

As of early 2026, the ClinicalTrials.gov registry lists over fifteen registered trials investigating GLP-1 receptor agonists specifically for substance use disorders. The breakdown by indication:

• Alcohol use disorder (AUD): Six to eight trials, including phase 2 and phase 3 designs, testing semaglutide (injectable and oral), exenatide, and combination approaches alongside behavioral therapy. Sites in the US, Denmark, and the UK. Several are NIH-funded through NIAAA.
• Nicotine use disorder / tobacco: Three to four trials, primarily in early phase, testing semaglutide and liraglutide as smoking cessation adjuncts alongside standard NRT (nicotine replacement therapy) or varenicline.
• Opioid use disorder (OUD): Three to five trials, currently mostly phase 1 and 2 safety and dose-finding, testing compatibility with buprenorphine/naloxone maintenance treatment and measuring craving and relapse outcomes at 12 and 24 weeks.
• Cannabis and stimulants: Two trials in very early phases — these indications are further behind but the mechanistic rationale from animal models is similar.

The most significant finding across early-phase alcohol trials is that the GLP-1 RA effect appears largest on heavy drinking days and on cue-induced craving — the specific phenotype that existing pharmacotherapy (naltrexone, for example) targets but fails many patients with. If phase 3 data holds up, GLP-1 RAs would represent the first genuinely novel mechanism of action in addiction pharmacotherapy in decades.

What this doesn't mean — important caveats

The excitement around GLP-1 RAs in addiction medicine is warranted, but several caveats belong in the clinical conversation.

First, the existing completed trials are small. The Klausen AUD trial enrolled 127 participants. The observational studies, while large, are retrospective and subject to confounding — patients prescribed semaglutide may differ from controls in ways not fully captured by matched variables [6]. Phase 3 trials are underway but results are 2–3 years out.

Second, the side-effect profile of GLP-1 RAs — nausea, vomiting, gastroparesis, pancreatitis (rare), thyroid C-cell effects (rare, mechanistic concern in rodents, not confirmed in humans at therapeutic doses) — is relevant in an addiction context where medication adherence and tolerability are already significant challenges. GLP-1 RA-induced nausea may be misinterpreted as withdrawal symptoms [11].

Third, the mechanism — reduced dopaminergic signaling in reward circuits — is not specific to addictive substances. If sustained GLP-1 RA use blunts the reward signal broadly, the long-term neurological consequences for motivation, anhedonia, and general quality of life are not fully characterized in humans. Animal data on long-term GLP-1R agonism in CNS reward circuits is reassuring but not definitive.

The clinical value of these drugs in addiction medicine may prove transformative. But the honest summary for early 2026 is: mechanism is strong, early-phase data is encouraging, phase 3 confirmation is pending. GLP-1 RAs are not yet evidence-based standard of care for addiction by any established guideline — they are an active and serious area of investigation [12]. For context on how GLP-1 response varies at the receptor level, see our analysis of GLP-1 non-responders.