Anandamide

In brief

What anandamide is, and why the search for it began

Start with the receptor, because the receptor came first. In 1988, in the Department of Pharmacology at St. Louis University Medical School, William Devane, Allyn Howlett and colleagues took a synthetic cannabis-like compound, attached a radioactive hydrogen atom to it, and showed that rat brain membranes bound it tightly, saturably, and in a way that was displaced by delta-9-tetrahydrocannabinol, while cannabidiol and cannabigerol displaced less than half of it even at 1 micromolar; the binding loosened in the presence of a guanine nucleotide, which is the signature of a receptor coupled to a G protein ³³. The dissociation constant was 133 picomolar. That is a specific protein, not a drug dissolving into a membrane, which had been the standing explanation for how cannabis worked.

Two years later, Lisa Matsuda and colleagues at the National Institute of Mental Health cloned the complementary DNA that encodes it: a G protein-coupled receptor that inhibits the enzyme adenylate cyclase in a dose-dependent, stereoselective, pertussis-toxin-sensitive way, whose messenger RNA appears in exactly the brain regions that bind cannabinoids ². This is now called CB1. In 1993 Sean Munro, Kerrie Thomas and Mouna Abu-Shaar at the Medical Research Council in Cambridge cloned a second one, not expressed in brain but in the macrophages of the marginal zone of the spleen, which is CB2 ⁵.

A receptor implies a key. Plants do not design molecules for mammalian brains; the brain must make something of its own that this protein evolved to read. In December 1992, in Science, William Devane, Lumir Hanus, Aviva Breuer, Roger Pertwee, Raphael Mechoulam and colleagues reported that they had found it. From pig brain they isolated an arachidonic acid derivative, settled its structure by mass spectrometry and nuclear magnetic resonance, confirmed that structure by synthesizing the compound, and showed that it displaced a radiolabeled cannabinoid from synaptosomal membranes the way a competitive ligand should and inhibited the electrically evoked twitch of the mouse vas deferens the way psychotropic cannabinoids do ¹.

They called it anandamide. The first half is the Sanskrit ananda, bliss; the second is the chemical suffix for an amide ⁴. The name has done the molecule some damage, because it promises more than the molecule delivers. Activating the same receptor can produce anxiety and panic as readily as calm, a point the people who work on it make explicitly ⁴.

The molecule itself is simple to draw and awkward to handle. Arachidonic acid is a twenty-carbon chain with four double bonds, all in the cis configuration, which bends it into a hook; attaching ethanolamine through an amide bond puts a small polar head on the end of that hook. The result is greasy, almost insoluble in water, and chemically unremarkable except for the amide bond, which is where the body cuts it.

The whole history, including every reversal

The history of anandamide runs in the opposite direction to most discoveries in pharmacology. Usually a substance is found in the body and a drug is then built to mimic or block it. Here the drug, cannabis, had been in use for thousands of years; its active ingredient had been isolated decades earlier; its receptor was characterized in 1988 and cloned in 1990; and only then did anyone go looking for what the receptor was really for.

Four claims made along the way have since had to be withdrawn or narrowed. That anandamide is the brain's retrograde messenger at synapses: it is not, 2-arachidonoylglycerol is ³⁵. That blocking its degradation would make a painkiller: in people it did not ^37,38. That the catastrophe in Rennes showed something about anandamide: the leading explanation points to one badly selective molecule, though the mechanism is unproven ^20,21. And that cannabidiol works by inhibiting the enzyme that destroys anandamide: it does so in rodent enzyme and not in the human one ³⁹.

What has survived is smaller and more solid. The molecule exists, is made on demand, is destroyed within minutes in the mouse brain ¹⁸, binds CB1 weakly, opens the capsaicin channel weakly, and matters enough that a person who breaks it down less efficiently lives a life measurably different from everyone else's ¹⁶.

How it is made, and how it is taken away

Most signaling molecules are built in advance and kept in vesicles until a stimulus releases them. Anandamide is not. It is carved out of the cell's own membrane at the moment it is wanted, which is why the enzymes that make and unmake it, rather than the molecule itself, have been the targets of the drug programs built around it.

The construction takes two steps. First, a calcium-dependent N-acyltransferase moves a fatty acyl chain from the sn-1 position of a membrane glycerophospholipid onto the nitrogen of a phosphatidylethanolamine, producing an N-acylphosphatidylethanolamine. Then a phosphodiesterase of the phospholipase D type removes the phospholipid head group, releasing the free N-acylethanolamine ¹⁰. When the acyl chain transferred is arachidonic acid, the product is anandamide; when it is a saturated chain, the products are relatives such as palmitoylethanolamide with their own activities.

That second enzyme, abbreviated NAPE-PLD in the literature, for N-acylphosphatidylethanolamine-specific phospholipase D ⁴⁷, was purified from rat heart and cloned from mouse, rat and human in 2004 by Okamoto, Morishita, Tsuboi, Tonai and Ueda. It is 393 to 396 amino acids long, bears no resemblance to any previously known phospholipase D, and belongs instead to the zinc metallohydrolase family of the beta-lactamase fold, the same structural family as the bacterial enzymes that destroy penicillin. Expressed in cells it generated anandamide and its congeners from the corresponding precursors at comparable rates and was inactive against ordinary phosphatidylcholine and phosphatidylethanolamine, and the enzyme turned out to be widely distributed, with the highest content in brain, kidney and testis ³. Purified away from everything else, it is activated by millimolar magnesium or calcium and so behaves as a constitutively active enzyme, holds a catalytically essential zinc, and loses almost all activity when any one of eight conserved histidine or aspartate residues is mutated; two naturally occurring human variants, L207F and H380R, also had low activity ³⁴.

The obvious next experiment, deleting that enzyme in a mouse, gave an unexpected answer. Calcium-dependent conversion of precursors to products fell more than five-fold. But when the brains were analyzed, only the saturated and very-long-chain members of the family had actually decreased; the polyunsaturated ones, which include anandamide, had not. A second, calcium-independent activity was found in the same brains that accepted the anandamide precursor ¹⁷. There is therefore no single gene whose loss removes anandamide, and no clean deficiency experiment to learn from.

Removal is simpler, and is where the pharmacology lives. Fatty acid amide hydrolase was isolated in 1996 from rat liver plasma membranes; it had been found as the enzyme that destroys oleamide, a lipid isolated from the spinal fluid of sleep-deprived cats, and was shown also to hydrolyze anandamide to arachidonic acid ¹¹. Its 2.8-angstrom crystal structure, published in 2002, explained how a water-soluble class of enzyme had been adapted to work on a substrate that lives in oil: a set of structural alterations lets it embed in the membrane and open a channel from the bilayer straight into its active site ¹². One enzyme does most of the work, but not all of it. A second, unrelated amidase, N-acylethanolamine acid amidase, also hydrolyzes this family of lipids, and inhibitors of it reduce inflammation and pain in animal models; it has a small drug program of its own and has never been tested against anandamide in a person ⁴⁷.

One detail of that enzyme has mattered enormously for drug development and is almost never mentioned outside specialist papers. Humans, other primates, marsupials and more distant vertebrates carry a second, distinct enzyme with the same activity, FAAH-2, which shares only 20 percent of its sequence with the first. Mice and rats do not have it at all. The two human enzymes hydrolyze simple fatty acid amides at similar rates, but the first is far more active against N-acyl ethanolamines such as anandamide ⁴⁸. Any experiment that moves from a rodent to a person is therefore not comparing like with like, a point a 2026 review of the whole program makes explicitly: rodent and human versions of the enzyme have different three-dimensional structures and interact differently with drugs, so candidates must be tested against the human enzyme or human tissue ²¹.

Where it acts, and how strongly

Anandamide has one classical target, one secondary target, and a long list of claimed targets that have not met the bar. The classical target is CB1, which is concentrated in the brain and is responsible for the psychoactive effects of cannabis ². The secondary target is CB2, concentrated in immune cells, first characterized in 1993 through its expression in splenic macrophages ⁵. The third, which gives the molecule its strangest property, is TRPV1: the channel that makes chili peppers hot ⁶.

At the cannabinoid receptor, anandamide is a real but modest ligand. In rat synaptosomal membranes it displaces a radiolabeled cannabinoid with an inhibition constant of 89 nanomolar ⁹. That is good binding. The problem is what happens next: in a living animal, anandamide produces only weak and transient cannabinoid effects, and the reason, demonstrated by the knockout mouse, is that it is broken down within minutes ^13,18. Give the same dose to a mouse that cannot destroy it and you get the full cannabinoid picture, reduced movement, analgesia, catalepsy and a fall in body temperature, all abolished by a CB1 blocker ¹³.

The TRPV1 story needs stating carefully, because it is routinely overstated. The original finding is solid: in isolated guinea pig and rat arteries, anandamide's vasodilation is capsaicin-sensitive, releases calcitonin gene-related peptide, is blocked by capsazepine and not by a CB1 blocker, is not reproduced by five other cannabinoid receptor agonists, two endogenous and three synthetic, and anandamide produces capsazepine-sensitive currents in whole cells and excised membrane patches from cells expressing the cloned channel ⁶. At the human channel, in a plate-reader calcium assay, anandamide behaved as a full agonist with a potency of 5.94 on the negative logarithmic scale against capsaicin's 7.13, and its response was blocked by capsazepine but not by two cannabinoid receptor blockers; anandamide and capsaicin cross-desensitized each other ⁷.

The best summary of what that adds up to was written in 2003 by Ruth Ross. Anandamide is a low intrinsic efficacy agonist at this channel. It looks like a partial agonist in tissues with few channels and like a full agonist in tissues with many, or in certain disease states. Its apparent efficacy is shifted by receptor reserve, phosphorylation, metabolism, uptake, CB1 activation, voltage, temperature, pH, and even by how much bovine serum albumin is in the bath. Whether it matters as a physiological agonist of this channel compared with other candidates, she wrote, has yet to be revealed ⁸. Seven years later the nomenclature committee of the International Union of Basic and Clinical Pharmacology took a similar line, and it is the last formal statement this page found: the channel possibly functions as an ionotropic cannabinoid receptor under physiological or pathological conditions, and no non-CB1, non-CB2 target then met the criteria that would make it a third cannabinoid receptor ³⁶. That position dates from 2010 and has not, as far as the searches behind this page show, been formally revisited.

There is also a cautionary tale in how the transport of anandamide was studied. In 1997 the compound AM404 was shown to block high-affinity anandamide accumulation into rat neurons and astrocytes without activating cannabinoid receptors or inhibiting hydrolysis, and to potentiate anandamide's effects in animals, which was taken as evidence for a dedicated carrier ⁴⁰. In 2000, AM404 and three chemically related uptake inhibitors, arvanil, olvanil and linvanil, were found to activate the human capsaicin channel at concentrations 400 to 10,000-fold below those needed to inhibit the transport process they were being used to study ⁴¹. Any experiment that used those compounds as transport blockers was also, unknowingly, an experiment on TRPV1.

The drug program: one futility stop, one death, and one target still standing

If anandamide is useful and short-lived, the obvious drug is one that stops it being destroyed. That logic drove twenty years of work on inhibitors of fatty acid amide hydrolase, and the results are worth laying out in order, because they are the clearest case in modern pharmacology of a target surviving its drugs.

The first clean clinical test was Pfizer's. In a crossover trial published in 2012, 74 patients with painful knee osteoarthritis were randomized to one of four treatment sequences, each pairing two weeks of either the irreversible inhibitor PF-04457845 at 4 mg daily or naproxen 500 mg twice daily with two weeks of placebo, in a design with predefined decision rules. The drug did exactly what it was built to do: enzyme activity fell by more than 96 percent and four endogenous fatty acid amides rose substantially. It also did nothing for pain. The mean difference from placebo in the WOMAC pain score was 0.04, with an 80 percent confidence interval of minus 0.63 to 0.71, while naproxen's was minus 1.13, interval minus 1.79 to minus 0.47. The study stopped at interim analysis for futility. The drug was well tolerated, with no cannabinoid-type adverse effects, and the authors wrote plainly that the absence of analgesia in humans contrasts with the animal data and that the species disconnect needs further study ³⁷. The trial was run and reported by Pfizer employees, which is worth knowing and does not change the arithmetic: a company reporting that its own compound failed is the least suspect kind of industry paper.

The second test killed someone. BIA 10-2474, developed by the Portuguese company BIAL, went into healthy volunteers in Rennes. Single doses from 0.25 to 100 mg and repeated doses of 2.5 to 20 mg for ten days had been given to 84 volunteers with no severe adverse events. A final cohort assigned six people to 50 mg daily and two to placebo. From the fifth day of dosing, an acute, rapidly progressive neurological syndrome developed in three of the four treated participants whose data were published: headache, a cerebellar syndrome, memory impairment and altered consciousness, with bilateral symmetric lesions, microhemorrhages and abnormal signal on fluid-attenuated inversion recovery and diffusion-weighted imaging, chiefly in the pons and hippocampi. One became brain dead. Two improved but were left, one with memory impairment and one with a cerebellar syndrome. One remained asymptomatic ¹⁴. The company's own later account of the full study, written largely by BIAL staff, reports 116 participants, central nervous system adverse events in five of the six who received 50 mg repeatedly, one fatal, a plasma half-life of 8 to 10 hours, and the conclusion that the available data do not provide a complete mechanistic explanation ¹⁵. A review of the regulatory animal studies, also co-authored by company staff, concluded that nothing in the preclinical package predicted it ⁵⁰.

The explanation that is available came from outside the company. Using activity-based protein profiling, a chemical method that labels active enzymes in intact cells and tissues, van Esbroeck and colleagues mapped what BIA 10-2474 actually binds in human cells and tissues. It inhibits several lipases that PF-04457845, the highly selective inhibitor already tested in people, does not touch, and BIA 10-2474, but not PF-04457845, produced substantial changes in lipid networks in human cortical neurons. Their conclusion was stated as a suggestion, not a proof: promiscuous lipase inhibitors have the potential to cause metabolic dysregulation in the nervous system ²⁰. A 2022 computational and modeling study, by BIAL's own research group, also corrected a point of the public record: the compound had been described in the clinical report as a reversible inhibitor, and the mechanism they describe is covalent and, under physiological conditions, irreversible ⁵¹.

This matters because of how the event was read at the time. The deaths of volunteers in early-phase trials are rare enough that each one reshapes a field, and the lesson many people took was that raising anandamide is dangerous. The evidence points the other way: in the osteoarthritis trial the selective inhibitor PF-04457845 cut enzyme activity by more than 96 percent and substantially raised four endogenous fatty acid amides, and produced no cannabinoid-type adverse effects at all ³⁷; a later trial of the same drug in cannabis dependence had no serious adverse events ⁵²; and a woman who has carried high anandamide all her life is described with painless injuries that heal quickly rather than with harm ¹⁶. Safe is not the same as pleasant. In healthy volunteers given that same selective inhibitor, 16 on drug against 29 on placebo, raising anandamide reduced how pleasant people found the slow, light stroking that the skin's C-tactile fibers are tuned to, and in a separate group of 101 adults higher anandamide went with the same reduced preference; ratings of emotional pictures were unaffected, so the effect was specific to social touch, and it ran opposite to what the authors had predicted ⁵³.

The target has not been abandoned, and it has produced one positive clinical result. In a phase 2a trial published in 2019, 70 men with cannabis dependence were randomized two to one to PF-04457845 at 4 mg daily or placebo, admitted to hospital for five days to precipitate withdrawal and then treated as outpatients for three more weeks. Withdrawal symptom scores were lower on the drug on the first two days of treatment, self-reported use at four weeks was 0.40 against 1.27 joints per day, and urinary tetrahydrocannabinol metabolite concentrations were lower. There were no serious adverse events. The study was funded by the United States National Institute on Drug Abuse ⁵².

And in 2025 the other main hope was tested and failed. One hundred patients with post-traumatic stress disorder, 85 of them women, were randomized to the inhibitor JNJ-42165279 at 25 mg twice daily or placebo for twelve weeks, with internet-delivered cognitive behavioral therapy including exposure modules from week five. The drug raised anandamide as designed. It did not improve the clinician-rated primary outcome or any secondary measure, and the authors concluded that enzyme inhibition does not appear to be a suitable adjunct for enhancing exposure therapy in this disorder ⁴⁵. Two of the authors are employees of the company that makes the compound, and the result is negative, which is the combination that most deserves to be believed. A 2021 review had argued that these drugs, useless for pain, might yet be useful in post-traumatic stress disorder and cannabis use disorder ³⁸; after 2025, the post-traumatic stress disorder half rests on a negative trial and the cannabis use disorder half on a single positive phase 2a trial ^45,52.

Two 2026 reports closed the argument further. The same compound was tried in a different indication and failed there too: 153 patients with major depressive disorder with anxious distress who had not responded adequately to an antidepressant were randomized one to one to JNJ-42165279 at 25 mg daily or placebo on top of that antidepressant, and the difference on the 17-item Hamilton Depression Rating Scale at six weeks was minus 0.2 points, standard error 1.04, one-sided p=0.416, with nothing on the key secondary endpoints, while plasma fatty acid amides again rose substantially and in step with drug concentration ⁵⁴. And an imaging substudy of the post-traumatic stress disorder trial scanned 76 of the 100 participants: four weeks of inhibition raised anandamide as designed and changed neither resting-state functional connectivity nor brain activation during emotional processing, which is the mechanism the drug was supposed to engage ⁵⁵.

The most recent summary of this part of the field, a 2026 review of endocannabinoid signaling and mood regulation in the Annual Review of Pharmacology and Toxicology, puts it without decoration: the enzyme and CB1 are present in the circuits that regulate emotion, manipulating them changes anxiety, stress responses and reward in animals, inhibitors have largely failed to show efficacy in people with mood disorders, CB1 agonists have only modest effects there, and closing that gap will require understanding why the model systems and the patients disagree ⁵⁶.

The woman who feels almost no pain

A 2019 report described a 66-year-old woman in the Scottish Highlands who had come to attention because she asked for no pain relief after a trapeziectomy, an operation on the base of the thumb that is reliably painful. Beyond a single gram of intravenous acetaminophen, which the British report calls paracetamol, given in the recovery area she needed no postoperative analgesic at all, and her pain score stayed at 0 out of 10 until she went home the next day, including after the nerve block in her arm had worn off. Questioning revealed a lifetime of the same thing: frequent cuts and burns she did not notice, which healed quickly. She was reported by Habib, Okorokov, Hill, Cox and colleagues at University College London with her anesthesiologist in Inverness ¹⁶.

The genetics turned out to be unusual in a way that had not been described before. She carries two things together. The first is a microdeletion in a pseudogene the authors cloned from the region next to the fatty acid amide hydrolase gene and named FAAH-OUT, that is, the stretch of DNA lying just outside the FAAH gene, expressed in dorsal root ganglia and brain. The second is a common functional single nucleotide polymorphism in the enzyme gene itself that reduces its expression and activity. Her blood carried significantly higher concentrations of anandamide and of two related fatty acid amides than four controls, two of them carriers of the same common variant: anandamide was 70 percent higher, palmitoylethanolamide and oleoylethanolamine roughly tripled, and 2-arachidonoylglycerol, which is cleared mostly by a different enzyme, was unchanged ¹⁶.

The pattern in the family is informative. Her son carries the same roughly 8 kilobase microdeletion and reports some pain insensitivity, though less than hers, and he does not carry the common low-activity variant; her mother and daughter carry neither and perceive pain normally. One similar deletion turned up among 5,008 alleles screened in the 1000 Genomes Project, in a Colombian man whose pain phenotype is unknown, so the authors expect that other people like her exist undiagnosed. Formal sensory testing showed hyposensitivity to noxious heat in her hands and feet, and she eats Scotch bonnet chili peppers without discomfort, reporting only a brief pleasant glow in the mouth, which is worth holding beside what the section on the capsaicin channel says above ¹⁶.

A pseudogene is a stretch of DNA that looks like a gene but no longer makes a working protein, and for decades such regions were treated as debris. The 2023 follow-up showed what this one is doing. FAAH-OUT is transcribed as a long non-coding RNA, and disrupting that transcription leads to methylation of the DNA within the enzyme's promoter in a manner dependent on the methyltransferase DNMT1; the region also contains a conserved element the authors call FAAH-AMP, which acts as an enhancer for the enzyme's expression. Transcriptomic work in cells derived from the patient uncovered a network of genes whose expression is altered by disruption of this axis ²². It remains the first long non-coding RNA linked to an inherited human pain insensitivity disorder ⁴⁴.

There is a second, much more common human experiment in the same gene. The variant 385C to A changes a conserved proline at position 129 to threonine. The resulting enzyme is catalytically normal but is degraded faster, so carriers break down less anandamide. The 2002 report that described it found the homozygous form strongly associated with street drug use and with problem drug and alcohol use ⁴³. In 2015 a mouse was engineered to carry the human change, and the mouse and human carriers were compared directly: both showed reduced enzyme expression, enhanced connectivity between frontal cortex and amygdala, better extinction of learned fear and less anxiety-like behavior ⁵⁷.

What these two findings together support is a specific and limited claim: that lower activity of the enzyme that degrades anandamide goes with less pain and less anxiety in people. They do not isolate anandamide, because the patient with the microdeletion also carries raised palmitoylethanolamide and oleoylethanolamine ¹⁶, and the C385A evidence is genetic association ^43,57. It does not mean that a drug raising anandamide will reproduce the phenotype, and the clinical trials say it has not. The difference between a lifelong genetic change in a developing nervous system and twelve weeks of a pill in an adult may be the whole story. Whether that difference explains the failures is untested; the 2026 Annual Review names the gap between model systems and patients as the problem to solve ⁵⁶.

The runner's high: a textbook claim that changed hands

For four decades the pleasant, calm, mildly analgesic state that can follow sustained running was explained by endorphins, the body's opioid peptides. The evidence for it was always indirect, and it was never as thin as the reversal story suggests. Running raises blood concentrations of both beta-endorphin and anandamide in mice, so no correlation distinguishes the two ²³. And central opioid release does happen in people: in a positron emission tomography study of ten trained athletes scanned at rest and after a two-hour run averaging 21.5 kilometers, opioid receptor availability fell in prefrontal and limbic regions, euphoria rose significantly, and the degree of euphoria was inversely correlated with opioid binding in prefrontal and orbitofrontal cortex, the anterior cingulate cortex and both insulae ⁵⁸. Ten athletes is a pilot study and a correlation is not a cause, but the claim that failed was the stronger one, that opioids are necessary for the state.

The first human data pointing the other way came in 2003, when Sparling, Giuffrida, Piomelli, Rosskopf and Dietrich had trained male college students run on a treadmill or cycle for 50 minutes at 70 to 80 percent of maximum heart rate and found that moderate-intensity exercise activates the endocannabinoid system; they offered it as a new candidate mechanism for exercise-induced analgesia ⁵⁹.

The decisive animal experiment came in 2015. Fuss, Steinle, Bindila, Auer, Kirchherr, Lutz and Gass combined pharmacological blockade, genetic deletion and behavioral testing in mice, and separated the components of the state. Reduced anxiety after running depended on intact CB1 receptors on forebrain GABAergic neurons. Reduced pain depended on peripheral CB1 and CB2 receptors. Sedation was not affected by blocking either cannabinoid or opioid receptors. Euphoria, as they noted, cannot be studied in a mouse ²³.

The human version followed in 2021. Siebers, Biedermann, Bindila, Lutz and Fuss randomized 63 participants, double-blind, to the opioid receptor blocker naltrexone or placebo, and had them run on a treadmill for 45 minutes at moderate intensity or walk as a control. Running produced greater euphoria and lower anxiety than walking, and raised plasma anandamide and 2-arachidonoylglycerol. Blocking opioid receptors prevented none of this, and the proportion of participants reporting a subjective runner's high was the same with naltrexone as with placebo. The authors concluded that the development of a runner's high does not depend on opioid signaling in humans, and that endocannabinoids are strong candidates ²⁴.

That is the reversal, and it is a real one: a claim that was in textbooks is not supported, and the burden has moved. What it is not is proof that anandamide causes the feeling. Both studies in people are correlational on the endocannabinoid side; naltrexone experiments can rule an explanation out, not rule one in. No one has yet given a person a cannabinoid receptor blocker before exercise and shown that the high disappears, which is the experiment the field still owes.

Two 2026 studies show how hard the measurement is. In a randomized crossover trial, 30 healthy adults cycled for 60 minutes at 30 percent and at 50 percent of heart rate reserve. Plasma endocannabinoid concentrations did not change uniformly and varied widely between individuals. Low-intensity exercise significantly reduced negative mood, but that change was not correlated with the cumulative plasma anandamide or 2-arachidonoylglycerol during exercise; only a within-person analysis across time points found a significant negative correlation between plasma anandamide and negative affect in the low-intensity condition ⁶⁰. And in a 12-week randomized trial of combined exercise and internet-delivered cognitive behavioral therapy in 80 adults with mild to moderate depression, the intervention group walked 81.6 meters farther than controls on a six-minute walk test at 12 weeks, a difference exceeding the minimal clinically important one, while none of nine endocannabinoid system biomarkers, anandamide included, showed any interaction between group and time ⁶¹. Exercise helps. The lipid in the tube is not yet a convincing account of why.

The newest human data come from the group that ran the naltrexone experiment, and they are field studies rather than experiments. Nineteen trained runners completed both a marathon and a duration-matched walk, with blood taken every 14 kilometers and again after a 45-minute break, and 36 ultramarathon runners were sampled before and after races of 100, 160 or 230 kilometers. Plasma anandamide rose progressively through the marathon and was still raised after the break, and it was raised after every ultramarathon distance, while 2-arachidonoylglycerol rose only in the later stages of the marathon and in early recovery. The marathon produced more euphoria and less anxiety than walking; the ultramarathons raised pain, lowered anxiety and did not significantly change euphoria, which is a useful warning against reading a rising lipid and a good feeling as the same event ⁶².

Anandamide in chocolate, and what the amounts mean

In August 1996 a one-page letter in Nature by Emmanuelle di Tomaso, Massimiliano Beltramo and Daniele Piomelli, titled Brain cannabinoids in chocolate, reported that cocoa and chocolate contain anandamide and two related compounds ⁶³. The quantities it reported are not quoted on this page.

Within two years the significance was publicly disputed in the same journal, in a letter from Vincenzo Di Marzo, Luciano De Petrocellis, Ester Fride, Raphael Mechoulam and colleagues titled Trick or treat from food endocannabinoids? ⁶⁴. It was published as a comment on the chocolate letter. The exchange is the origin of the standing position in the field: these compounds are present in several foods, and presence is not the same as effect.

There is one place where the chocolate claim was tested against reality, and the result is instructive. In a Belgian case reported in 2000, a court heard a defense argument that a man's positive urine screen for cannabis came from eating a large quantity of chocolate, on the grounds that chocolate contains anandamide-related lipids that inhibit anandamide breakdown and act as cannabinoid mimics. The toxicologists synthesized N-oleoylethanolamide and N-linoleoylethanolamide, spiked them together with anandamide into urine, and ran the immunoassay. None of the samples tested positive: there is no cross-reactivity with the cannabis assay. The defense claim was refuted and the accused was convicted ⁶⁵.

The general point is pharmacological rather than legal. Anandamide eaten in food has to survive the gut and the liver, both of which are rich in the enzyme that destroys fatty acid amides ^11,48, and then reach a receptor at a concentration somewhere near its inhibition constant of 89 nanomolar ⁹. Nothing in the record suggests chocolate does this. What chocolate does to mood is not in dispute; the explanation on offer here is not supported.

Pregnancy, implantation and the strangest concentrations in the body

When this was measured in 1997, the highest concentration of anandamide yet found in a mammalian tissue was not in the brain. Schmid, Paria, Krebsbach, Schmid and Dey measured it in the mouse uterus around the time of implantation and reported 142 to 1,345 picomoles per micromole of lipid phosphorus, which works out to roughly 1 to 7 micrograms per gram of wet tissue, the highest found anywhere at that time ³¹.

What makes this interesting is the pattern rather than the number. The concentration tracks the uterus's willingness to accept an embryo, and it tracks it backward. Levels are low when the uterus is receptive and high when it is refractory; they are highest in the non-receptive pseudopregnant uterus and between implantation sites, and lowest exactly at the site where an embryo is implanting. In the same study, anandamide slowed blastocysts hatching from their outer shell in culture, and a synthetic cannabinoid agonist given systemically inhibited implantation, an effect reversed by a CB1 blocker ³¹.

A 2001 follow-up tied the embryo into the same regulation. Mice deficient in CB1, CB2 or both showed asynchronous preimplantation development. Uterine anandamide and blastocyst CB1 are normally turned down together as the uterus becomes receptive and the blastocyst becomes competent; both stayed high in the non-receptive uterus, in dormant blastocysts during delayed implantation, and in pregnant mice lacking leukemia inhibitory factor, which fail to implant. And a sustained dose of a natural cannabinoid blocked implantation in normal mice but not in mice lacking both receptors, which is the control that makes the effect receptor-mediated rather than toxic ⁶⁶.

The human counterpart is a single study with an extraordinary result. In 2000 Maccarrone and colleagues in Rome measured anandamide hydrolase activity in peripheral lymphocytes of pregnant women. In an initial group of 50 women at 6 to 11 weeks, the 7 who went on to miscarry had a mean activity of 60.43 picomoles per minute per milligram of protein against 169.60 in the 43 who did not. They then defined a threshold on an antibody-based assay and applied it to 120 women at 7 to 8 weeks: all 15 below the threshold miscarried, against 1 of the 105 above it ⁴².

That is a near-perfect separation, and near-perfect separations in a single cohort are the kind of finding that most often shrinks on repetition. The same group added a third cohort of 100 healthy women in which low lymphocyte enzyme activity again went with spontaneous abortion, and showed how the enzyme is regulated: physiological concentrations of progesterone raise it, partly through the Th2 cytokines interleukin-4 and interleukin-10, while the Th1 cytokines interleukin-12 and interferon gamma lower it ⁶⁷. That is a further cohort from the same laboratory, not an independent replication, and twenty-six years later no test of this kind is part of routine obstetric care. Human tissue work points the same way as the mouse work: anandamide was higher in decidua from 18 miscarriages than from 8 elective terminations, and conditioned medium from the uterine natural killer cells of those miscarriage samples carried high tumor necrosis factor alpha and blocked the differentiation of endometrial stromal cells in culture ⁶⁸. The lymphocyte result should be read as a striking observation awaiting independent confirmation, not as a discarded one: nothing has shown it to be wrong either.

One further animal result shows the system is not only about implantation. When newborn mouse pups were given a CB1 blocker, either once on the first day of life or daily for a week from the second, milk ingestion and growth collapsed and the pups died within four to eight days; co-administering tetrahydrocannabinol in the week-long regimen almost fully reversed it ⁶⁹. In newborn mice, at least, suckling depends on it.

Obesity, psychosis, bipolar disorder, stress and cannabinoid hyperemesis

Circulating anandamide has been measured in a long list of conditions. Five of those measurements are worth setting out, because together they show both what the number can do and what it cannot.

Obesity. In 2005 Engeli and colleagues compared 20 obese and 20 lean women. Circulating anandamide was 35 percent higher and arachidonoylglycerol, measured as the combined 1- and 2-isomers because the assay cannot separate them, 52 percent higher in the obese group. In fat tissue, messenger RNA for CB1 was 34 percent lower and for the degrading enzyme 59 percent lower, and enzyme expression in fat correlated strongly and negatively with circulating endocannabinoids. A 5 percent weight loss in a separate group of 17 women changed none of it ²⁹. The reading is that the peripheral endocannabinoid system is upregulated in obesity, probably by making less of the enzyme that clears it, and that it does not reverse with modest weight loss.

Psychosis. In 2004 Giuffrida, Leweke, Piomelli and colleagues found cerebrospinal fluid anandamide eight-fold higher in 47 antipsychotic-naive patients in a first episode of paranoid schizophrenia than in 84 healthy controls, 13 patients with dementia or 22 with affective disorders. The elevation was absent in 37 patients treated with typical antipsychotics but not in 34 treated with atypical ones. Within the unmedicated acute group, higher anandamide went with fewer psychotic symptoms, a correlation of minus 0.452 ³⁰. The authors' interpretation, that the elevation is a compensatory adaptation to the disease rather than a cause of it, is the reading the inverse correlation favors, though a cross-sectional correlation cannot by itself establish it. A 2026 study in 58 antipsychotic-naive patients with a first episode of psychosis pointed the same way: higher baseline plasma anandamide predicted lower general psychopathology a year later, with no association with total symptom scores or with the positive and negative subscales ⁷⁰.

Bipolar disorder. A 2026 longitudinal study followed 23 patients with bipolar I disorder from acute mania to clinical stabilization, measuring plasma anandamide and 2-arachidonoylglycerol by liquid chromatography with tandem mass spectrometry. In patients who did not use cannabis, both fell significantly as they recovered. In those who did use cannabis, there was no significant reduction despite clinical improvement, with interaction p values of 0.016 for anandamide and 0.029 for 2-arachidonoylglycerol. Neither lipid tracked symptom severity on the mania rating scale ²⁸. Twenty-three patients is small, and the finding is an association in a single cohort.

Stress. The clean human experiment was done in 2012. Seventy-one adults attended two sessions, one containing a standardized psychosocial stressor, the Trier Social Stress Test, and one a control task. Serum anandamide and the related N-acylethanolamines rose immediately after the stress period and not after the control task; the rise in palmitoylethanolamide tracked the rise in cortisol; and before any stressor, the people with higher anxiety ratings had lower anandamide. The responses differed by sex, by menstrual cycle phase and between ethnic groups, which is a warning about every single-sample study in this field ⁷¹. The animal literature reads the same way from the other end: chronic stress turns endocannabinoid signaling down, and raising endocannabinoid tone blunts some of the consequences of chronic stress, which is part of why the failure of the drugs in people is the puzzle it is ⁵⁶.

Cannabinoid hyperemesis syndrome. The syndrome was described in 2004 by Allen, de Moore, Heddle and Twartz in South Australia: nineteen heavy cannabis users with a cyclical vomiting illness, of whom nine plus one previously published case were followed. Cannabis use preceded the vomiting in every case; stopping it stopped the illness in seven; three, including the published case, never stopped and kept vomiting, and three who restarted after abstaining relapsed, two of whom recovered on stopping again; and nine of the ten showed a compulsive washing behavior during attacks, which was the novel observation that made the syndrome recognizable ⁷². The obvious hypothesis is that the endocannabinoid system is deranged. The nearest direct test was done in cyclic vomiting syndrome, not in the cannabis syndrome itself: in 22 patients with cyclic vomiting syndrome, mean age 40, 17 of them women and some of them cannabis users, against 12 matched controls, serum endocannabinoids did not differ between groups at all. What did rise in the sick phase were two related lipids, N-oleoylethanolamine and N-palmitoylethanolamide, which correlated with nausea and with poor sleep; salivary cortisol and alpha amylase were higher in the cannabis users than the non-users during attacks ²⁷. No equivalent measurement has been published in patients defined by cannabinoid hyperemesis itself, and a 2025 review of that syndrome treats its mechanism as unsettled while noting that the average gap between the first episode and diagnosis in published cases is three to six years ⁷³. A broader 2025 review places the endocannabinoid system among several interacting pathways in cyclic vomiting and its cannabinoid-associated form, alongside the substance P receptor NK1, CB1, the stress axis, TRPV1 and others ⁷⁴.

The thread running through all five is set out in Cecilia Hillard's 2018 review of circulating endocannabinoids. Blood concentrations respond to energy intake and storage, to exercise, to sleep and its disruption, to inflammation, to pain and to stress; 2-arachidonoylglycerol has a marked daily rhythm. The sheer number of physiological inputs makes a circulating endocannabinoid concentration poor evidence on its own for any particular disorder, however useful it is for generating hypotheses ³².

Cannabidiol's claimed enzyme inhibition, and the deficiency hypothesis

Two widely repeated claims attach anandamide to things people can buy. Both are worth stating at their real strength.

The first is that cannabidiol works by inhibiting the enzyme that destroys anandamide. Its source is a good study. In 2012, Leweke, Piomelli, Koethe and colleagues ran a double-blind randomized trial of cannabidiol against amisulpride, an established antipsychotic, in acute schizophrenia. Both treatments were safe and led to significant clinical improvement; cannabidiol had a markedly better side-effect profile; and cannabidiol treatment was accompanied by a significant rise in serum anandamide, which was significantly associated with the clinical improvement. The authors proposed that inhibiting anandamide breakdown contributes to the antipsychotic effect, describing cannabidiol as a compound that moderately inhibits that breakdown ⁷⁵.

The mechanism did not survive. In 2015 Elmes, Kaczocha, Deutsch and colleagues showed that cannabidiol does not inhibit the human enzyme, in contrast to the rodent one, and used docking and site-directed mutagenesis to identify the residues in the active site that account for the species difference. What cannabidiol and tetrahydrocannabinol do instead is bind fatty acid-binding proteins, the intracellular carriers that ferry anandamide to the enzyme, and competing for those carriers slows anandamide's uptake and breakdown ³⁹. The clinical observation stands; the explanation printed on a thousand product pages does not. It is the same lesson the whole field has learned twice over, in the FAAH-2 species difference ⁴⁸ and in the 2026 review's insistence that candidate drugs be tested on the human enzyme ²¹.

The clinical observation has since been tested directly in people, with results that depend on dose. At the doses sold over the counter it does not reproduce: in a triple-blind randomized crossover trial, 70 healthy occasional cannabis users took oral cannabidiol at 20, 50, 100 and 200 mg and placebo, and dose had no effect on the cumulative rise in plasma anandamide, while cumulative palmitoylethanolamide and oleoylethanolamine were actually higher on placebo than on the 50 mg dose, a difference confined to the women ⁷⁶. At 800 mg daily for four weeks in 78 patients with cocaine use disorder there was no sustained effect on plasma anandamide either, with Bayes factors near 0.19 favoring no difference ⁷⁷. At 225 to 450 mg daily for five weeks in 66 patients with hypertension, plasma anandamide did rise, from 36.7 to 47.9 nanograms per milliliter, with no change in 2-arachidonoylglycerol, but the size of that rise had no relation to the fall in blood pressure ⁷⁸. So cannabidiol can raise anandamide at pharmaceutical doses, by some mechanism other than inhibition of the human enzyme, and that rise has not yet been shown to carry any clinical effect.

The second claim is the hypothesis of clinical endocannabinoid deficiency, proposed by Ethan Russo in 2001 and restated in a 2016 review. Its premise is that every person has an underlying endocannabinoid tone, set by the amounts of anandamide and 2-arachidonoylglycerol, their production and metabolism, and the abundance and state of the receptors; its claim is that in certain conditions that tone becomes deficient and produces syndromes. The conditions proposed are migraine, fibromyalgia and irritable bowel syndrome, chosen for their shared hyperalgesia, central sensitization, overlapping populations and history of being dismissed as psychosomatic ²⁵.

The review is honest about the original evidence base, which was genetic overlap, comorbidity, patterns of symptoms that the system could plausibly mediate, and the observation that cannabinoids often help. It argues that newer work strengthens the case, citing differences in cerebrospinal fluid anandamide in migraine and imaging evidence of reduced endocannabinoid function in post-traumatic stress disorder ²⁵. The human measurement that second claim leans on deserves to be shown: in 25 untreated patients with non-combat post-traumatic stress disorder, 12 trauma-exposed controls and 23 healthy controls, positron emission tomography with the CB1-selective tracer carbon-11 OMAR found receptor availability 19.5 percent higher than in the healthy controls and 14.5 percent higher than in the trauma-exposed controls, most pronounced in women, while circulating anandamide was 53.1 and 58.2 percent lower ⁷⁹. More receptor and less ligand is what a deficient tone would look like, and it sits awkwardly beside the 2026 finding of less of the degrading enzyme in the amygdala ⁴⁶, which should raise anandamide; the two have not been reconciled in the same patients. Two further things should be said alongside the hypothesis. The author's listed affiliation is PHYTECS, a company working on the endocannabinoid system, which is a commercial interest readers should weigh. And direct measurements are mixed: lower cerebrospinal fluid anandamide was reported in chronic migraine ²⁶, but endocannabinoids were not lower in cyclic vomiting syndrome patients than controls ²⁷, and the 2026 imaging study in post-traumatic stress disorder found less of the degrading enzyme in the amygdala, which would raise anandamide rather than lower it, the opposite of the animal prediction ⁴⁶.

The fair position is that an endocannabinoid tone exists, that it varies, and that no threshold below which a person is deficient has been defined for any tissue, any condition or any assay.

2-arachidonoylglycerol, briefly, and why it matters here

Anandamide is not the body's main endocannabinoid, and saying so is necessary to keep the rest honest. In 1995 two groups independently described 2-arachidonoylglycerol: Mechoulam and colleagues isolated it from canine intestine, measured binding constants of 472 nanomolar at CB1 and 1,400 at CB2, showed it inhibited adenylate cyclase in mouse spleen cells at about the potency of tetrahydrocannabinol, and found that given intravenously it produced the characteristic set of four cannabinoid effects in mice ¹⁹; Sugiura and colleagues found it in brain, where its apparent binding affinity rose markedly when an esterase inhibitor was present or the assay was run on ice, because the compound is destroyed during the measurement ⁹.

The structural difference is small and the functional difference is not. Both carry the same twenty-carbon arachidonic acid chain. Anandamide attaches it to ethanolamine through an amide bond; 2-arachidonoylglycerol attaches it to glycerol through an ester bond. Esters are easier to break than amides, and a different enzyme, monoacylglycerol lipase, does most of that work ^16,80.

The decisive comparison came in 2010. Tanimura and colleagues made mice lacking each of the two enzymes that synthesize 2-arachidonoylglycerol. In mice lacking diacylglycerol lipase alpha, endocannabinoid-mediated retrograde suppression of synaptic transmission was completely absent in cerebellum, hippocampus and striatum, basal 2-arachidonoylglycerol was markedly reduced, and stimulus-evoked increases were abolished; in mice lacking the beta form everything was normal ³⁵. Retrograde signaling at central synapses, the function most people have in mind when they say endocannabinoid, is not anandamide's job.

A 2026 result divides the labor more finely rather than giving anandamide its old job back. In adult mouse hippocampus, Noriega-Prieto and colleagues found that 2-arachidonoylglycerol signals selectively to neurons, producing the familiar presynaptic depression, while anandamide signals to astrocytes, which then potentiate neighboring synapses; anandamide, and not 2-arachidonoylglycerol, was required for spike-timing-dependent long-term potentiation ⁸¹. If that holds up elsewhere, anandamide is not a weaker copy of the other endocannabinoid but a messenger with a different address. It is work in mouse brain slices, with no human counterpart.

What anandamide appears to do instead is set a background. The phrase used by the authors of the knockout mouse study is an endogenous cannabinoid tone that modulates pain perception ¹³, and that framing has held up: it is present at low concentrations much of the time, is raised by stress, exercise and injury, and is cleared fast. The two lipids are frequently measured together and should not be read as interchangeable; in the bipolar study above, for example, they moved together, while in cyclic vomiting syndrome neither moved and two cousins did ^27,28.

The other enzyme has had its own drug program, and it ended the same way. Blocking monoacylglycerol lipase raises 2-arachidonoylglycerol, and the inhibitor Lu AG06466, formerly ABX-1431, was taken into Tourette syndrome, a neurodevelopmental disorder of chronic motor and vocal tics. A phase 1b crossover study in 20 adults already on standard treatment gave a single 40 mg dose or placebo: two of the three tic scales, including the total tic score of the Yale Global Tic Severity Scale, improved significantly, as did premonitory urges ⁸². The 12-week randomized placebo-controlled trial that followed, in 49 adults at two dose levels, found no efficacy at all; the difference on that same scale at week 8 was 3.0 points, 95 percent confidence interval 0.1 to 5.9, p=0.043, in favor of placebo, and the authors concluded there was no evidence of efficacy in suppressing tics ⁸³. Both studies were company-sponsored. A single-dose signal that disappears over twelve weeks is the same shape as the anandamide story, and it is the reason to treat raising either endocannabinoid as an unproven strategy rather than an untried one.

What changed, and what is still unknown

Five things that were believed are not. Anandamide is not the retrograde messenger at central synapses; 2-arachidonoylglycerol is ³⁵. Blocking its breakdown does not relieve pain in people, whatever it does in animals ^37,38. The Rennes catastrophe was not a demonstration that raising anandamide is dangerous; the leading chemical evidence, from an independent group, points to off-target lipase inhibition by one promiscuous compound ²⁰, although the mechanism has never been established ^14,21. Cannabidiol does not raise anandamide by inhibiting the human degrading enzyme ³⁹. And the runner's high is not an endorphin phenomenon, at least not a necessary one, because blocking opioid receptors does not abolish it ²⁴.

Several things remain genuinely open, and the most important is the species gap. Humans carry a second degrading enzyme that mice and rats lack ⁴⁸, and the human and rodent forms of the first enzyme differ structurally enough that a 2026 review calls for all candidates to be tested on human enzyme or human tissue ²¹. Whether that difference explains why strong analgesia in animals did not appear in the osteoarthritis trial is a hypothesis, not a finding ³⁷.

Nobody knows whether anandamide reaches TRPV1 at physiologically relevant concentrations in an intact person. The isolated-vessel and transfected-cell evidence is strong, the measured potency at the human channel is roughly fifteen-fold weaker than capsaicin's, the apparent efficacy depends on conditions that vary between tissues, and the tools used for years to probe anandamide transport turned out to activate the channel themselves at far lower concentrations than they blocked transport ^6,7,8,41. There is now a tool for the question. In rat sensory neurons the immediate precursor of anandamide, 20:4-NAPE, lowered excitability at 10 nanomolar to 1 micromolar through CB1 and raised it at 10 micromolar through TRPV1, and the NAPE-PLD inhibitor LEI-401 abolished both effects; given into the spinal canal, the low concentration made the animals less sensitive and the high concentration produced mechanical allodynia ⁸⁴. If it works that way in people, the same molecule would relieve and cause pain depending only on how much of it is made, which would explain a good deal of the inconsistency in this literature. It is rat work, in male animals.

How anandamide is carried into and out of cells is still unsettled. The 1997 transport evidence stands as a phenomenon, no dedicated membrane transporter has been confirmed, a partly cytosolic candidate called FLAT was proposed in 2012 ⁸⁵, and fatty acid-binding proteins appear to serve as intracellular carriers, which is also the best current explanation for why cannabis constituents raise circulating anandamide in people ^39,40.

The clinical questions that could be answered and have not been: whether a cannabinoid receptor blocker given before exercise abolishes the runner's high in a person, which would convert the 2021 result from an exclusion into a demonstration ²⁴; whether the 2000 lymphocyte enzyme result in early pregnancy survives a second cohort ⁴²; why inhibiting the degrading enzyme helps cannabis withdrawal but not pain or post-traumatic stress disorder ^37,45,52; and why the amygdala in post-traumatic stress disorder holds less of that enzyme when the animal literature predicted more ⁴⁶.

The deepest unanswered question is the one the field names itself. A lifelong genetic reduction in the degrading enzyme produces a person who is nearly painless, barely anxious, quick to heal, and forgetful ^16,22. Twelve weeks of a drug that inhibits the same enzyme raises anandamide but produces no clinical benefit in post-traumatic stress disorder ⁴⁵. Until someone can explain that gap, the endocannabinoid system remains a target with genetic support in people and inhibitors that have largely failed in trials ⁵⁶.

The tests, one by one

• Plasma or serum anandamide by liquid chromatography with tandem mass spectrometryThe lipid is extracted from blood and measured against a labeled internal standard. It is the method behind almost every human number on this page, and its chief limitation is that circulating concentrations answer to food, exercise, sleep, stress, inflammation and time of day at once, so a single value is hard to attach to any one condition ³². A 2026 crossover trial in 30 healthy adults found that plasma concentrations did not change uniformly with exercise and varied enormously between individuals, so that only within-person change carried any signal ⁶⁰.Research only: no clinical decision anywhere rests on a blood anandamide resultAnalytically reliable; biologically noisyIt cannot tell you what is happening at a synapse or a nerve ending, which is where the molecule actually acts ³²
• Cerebrospinal fluid anandamideA lumbar puncture sample measured the same way. It is the only human measurement that comes from inside the central nervous system, and it produced the clearest finding of its kind: eight-fold higher concentrations in 47 antipsychotic-naive people in a first episode of paranoid schizophrenia than in 84 healthy controls, 13 patients with dementia or 22 with affective disorders, with a negative correlation against psychotic symptom severity of minus 0.452 ³⁰.Research into psychosis, pain and headacheGood within a study; no reference range exists for clinical useFluid around the brain is not the tissue of the brain, and the direction of causation cannot be read from a concentration
• Brain imaging of the degrading enzyme with carbon-11 CURBA positron-emitting tracer that binds fatty acid amide hydrolase irreversibly, so the rate at which it is trapped reports how much active enzyme a brain region holds; quantifying it requires an arterial input function and a two-tissue compartment model ⁸⁶. It is the nearest thing to seeing anandamide handling in a living human brain. In a 2026 study of 86 people, 35 with post-traumatic stress disorder and 51 controls, amygdala enzyme levels were 10 percent lower in the disorder, the opposite of what animal work predicted, and were not associated with symptom severity ⁴⁶.Research imaging in psychiatry and addictionA quantitative measure of the enzyme, validated as a methodIt measures the enzyme, not the lipid; less enzyme implies more anandamide only indirectly
• Fatty acid amide hydrolase activity and protein in peripheral lymphocytesThe enzyme is measured in circulating white cells by a radiochromatographic assay and by an antibody-based assay. It produced the single most dramatic clinical claim in this field: in 120 women at 7 to 8 weeks of pregnancy, all 15 whose enzyme concentration fell below a threshold miscarried, against 1 of the 105 above it ⁴².A research assay. No such test is offered in routine obstetric careWithin that single-center study the separation was near-total; the same laboratory later reported a third cohort pointing the same way, but no independent replication of it was found in the searches behind this page ⁶⁷A blood cell is not a uterus, and a marker that separates groups perfectly in one cohort usually does not do so in the next
• Genotyping of the FAAH C385A variantA single base change, 385C to A, replaces a conserved proline at position 129 with threonine. The resulting enzyme works normally but is broken down faster, so carriers degrade less anandamide; in homozygous form it was associated with street drug use and problem drug and alcohol use in the 2002 report that described it ⁴³. A knock-in mouse carrying the human change, compared with human carriers, showed the same biochemistry, the same stronger connection between frontal cortex and amygdala, better fear extinction and less anxious behavior ⁵⁷.Research into anxiety, fear extinction and substance use; no clinical roleThe genotype call is exactThe step from genotype to any individual's experience is not established, and the original drug-use association is a single association study

Where it connects

Sources

86 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Devane WA, Hanus L, Breuer A, Pertwee RG, Stevenson LA, Griffin G, Gibson D, Mandelbaum A, Etinger A, Mechoulam R. Isolation and structure of a brain constituent that binds to the cannabinoid receptor.Science · 1992 · 258(5090):1946-9doi:10.1126/science.1470919 · PMID 1470919

   The discovery paper. Arachidonylethanolamide was identified in porcine brain in a screen for endogenous cannabinoid receptor ligands, its structure determined by mass spectrometry and nuclear magnetic resonance and confirmed by synthesis, and it inhibited binding of a radiolabeled cannabinoid probe to synaptosomal membranes as a competitive ligand and inhibited the electrically evoked mouse vas deferens twitch. It names the compound; it does not establish a physiological role.

2. 2
   Matsuda LA, Lolait SJ, Brownstein MJ, Young AC, Bonner TI. Structure of a cannabinoid receptor and functional expression of the cloned cDNA.Nature · 1990 · 346(6284):561-4doi:10.1038/346561a0 · PMID 2165569

   The cloning of CB1: a complementary DNA encoding a G protein-coupled receptor that inhibits adenylate cyclase dose-dependently, stereoselectively and in a pertussis toxin-sensitive way, whose messenger RNA is found in the cell lines and brain regions that bind cannabinoids. It establishes the receptor; it says nothing about an endogenous ligand.

3. 3
   Okamoto Y, Morishita J, Tsuboi K, Tonai T, Ueda N. Molecular characterization of a phospholipase D generating anandamide and its congeners.J Biol Chem · 2004 · 279(7):5298-305doi:10.1074/jbc.M306642200 · PMID 14634025

   Purification of the enzyme from rat heart and cloning of its complementary DNA from mouse, rat and human: 393 to 396 amino acids, no homology with known phospholipase D enzymes, a member of the zinc metallohydrolase family of the beta-lactamase fold, generating anandamide and other N-acylethanolamines at comparable rates in COS-7 cells, inactive with phosphatidylcholine and phosphatidylethanolamine, and most abundant in brain, kidney and testis.

4. 4
   Moreira FA, Lutz B. The endocannabinoid system: emotion, learning and addiction.Addict Biol · 2008 · 13(2):196-212doi:10.1111/j.1369-1600.2008.00104.x · PMID 18422832

   A review whose abstract states that N-arachidonoyl ethanolamine was named anandamide after the Sanskrit word ananda, meaning bliss, and that emotional responses to cannabinoids are not always pleasant: anxiety and panic may also follow CB1 activation. Used here for the etymology and for that caveat.

5. 5
   Munro S, Thomas KL, Abu-Shaar M. Molecular characterization of a peripheral receptor for cannabinoids.Nature · 1993 · 365(6441):61-5doi:10.1038/365061a0 · PMID 7689702

   The cloning of CB2, a cannabinoid receptor expressed in macrophages of the marginal zone of the spleen rather than in brain, with the observation that the previously cloned receptor is expressed in brain but not in the periphery except at a low level in testes. The macrophages are where it was found to be expressed, not the material it was cloned from.

6. 6
   Zygmunt PM, Petersson J, Andersson DA, Chuang H, Sorgard M, Di Marzo V, Julius D, Hogestatt ED. Vanilloid receptors on sensory nerves mediate the vasodilator action of anandamide.Nature · 1999 · 400(6743):452-7doi:10.1038/22761 · PMID 10440374

   Anandamide-induced vasodilation in isolated arteries is capsaicin-sensitive and accompanied by calcitonin gene-related peptide release, blocked by a CGRP receptor antagonist and by capsazepine but not by the CB1 blocker SR141716A, and not reproduced by 2-arachidonoylglycerol, palmitylethanolamide, HU 210, WIN 55,212-2 or CP 55,940; anandamide produced capsazepine-sensitive currents in cells expressing the cloned vanilloid receptor. Isolated guinea pig and rat tissue.

7. 7
   Smart D, Gunthorpe MJ, Jerman JC, Nasir S, Gray J, Muir AI, Chambers JK, Randall AD, Davis JB. The endogenous lipid anandamide is a full agonist at the human vanilloid receptor (hVR1).Br J Pharmacol · 2000 · 129(2):227-30doi:10.1038/sj.bjp.0703050 · PMID 10694225

   In transfected human cells, anandamide and capsaicin were full agonists with pEC50 values, the negative logarithm of the concentration producing half the maximal response, of 5.94 plus or minus 0.06 and 7.13 plus or minus 0.11; the anandamide response was inhibited by capsazepine with a pKB of 7.40 but not by the cannabinoid antagonists AM630 or AM281, and the two agonists cross-desensitized. Work from SmithKline Beecham Pharmaceuticals; a recombinant-cell result, not a tissue or human one.

8. 8
   Ross RA. Anandamide and vanilloid TRPV1 receptors.Br J Pharmacol · 2003 · 140(5):790-801doi:10.1038/sj.bjp.0705467 · PMID 14517174

   The careful review: anandamide is a low intrinsic efficacy TRPV1 agonist, behaving as a partial agonist where receptor reserve is low and a full agonist where it is high or in certain disease states, with efficacy modified by phosphorylation, metabolism, uptake, CB1 activation, voltage, temperature, pH and bovine serum albumin; it states explicitly that its importance as a physiological TRPV1 agonist has yet to be revealed.

9. 9
   Sugiura T, Kondo S, Sukagawa A, Nakane S, Shinoda A, Itoh K, Yamashita A, Waku K. 2-Arachidonoylglycerol: a possible endogenous cannabinoid receptor ligand in brain.Biochem Biophys Res Commun · 1995 · 215(1):89-97doi:10.1006/bbrc.1995.2437 · PMID 7575630

   Competition against tritiated CP-55,940 in rat synaptosomal membranes gave an inhibition constant of 89 nanomolar for anandamide; 2-arachidonoylglycerol also bound, with apparent affinity much higher when an esterase inhibitor was present or the assay run at 0 degrees Celsius. Binding data in membranes, not a functional or in vivo measure.

10. 10
    Ueda N, Okamoto Y, Tsuboi K. Endocannabinoid-related enzymes as drug targets with special reference to N-acylphosphatidylethanolamine-hydrolyzing phospholipase D.Curr Med Chem · 2005 · 12(12):1413-22doi:10.2174/0929867054020918 · PMID 15974992

    A review stating the two-step biosynthesis: transfer of a fatty acyl chain from the sn-1 position of a glycerophospholipid onto phosphatidylethanolamine by a calcium-dependent N-acyltransferase to make N-acylphosphatidylethanolamine, then release of the N-acylethanolamine by a phospholipase D-type phosphodiesterase. A review, not primary data.

11. 11
    Cravatt BF, Giang DK, Mayfield SP, Boger DL, Lerner RA, Gilula NB. Molecular characterization of an enzyme that degrades neuromodulatory fatty-acid amides.Nature · 1996 · 384(6604):83-7doi:10.1038/384083a0 · PMID 8900284

    Isolation, cloning and expression from rat liver plasma membranes of the enzyme originally called oleamide hydrolase, with the demonstration that it converts anandamide to arachidonic acid, and the renaming to fatty acid amide hydrolase.

12. 12
    Bracey MH, Hanson MA, Masuda KR, Stevens RC, Cravatt BF. Structural adaptations in a membrane enzyme that terminates endocannabinoid signaling.Science · 2002 · 298(5599):1793-6doi:10.1126/science.1076535 · PMID 12459591

    The 2.8-angstrom crystal structure of the enzyme complexed with an arachidonyl inhibitor, showing the structural alterations that let it integrate into the membrane and reach its substrate directly from the bilayer. Rat enzyme.

13. 13
    Cravatt BF, Demarest K, Patricelli MP, Bracey MH, Giang DK, Martin BR, Lichtman AH. Supersensitivity to anandamide and enhanced endogenous cannabinoid signaling in mice lacking fatty acid amide hydrolase.Proc Natl Acad Sci U S A · 2001 · 98(16):9371-6doi:10.1073/pnas.161191698 · PMID 11470906

    States that anandamide binds and activates CB1 in vitro but induces only weak and transient cannabinoid effects in vivo, probably through rapid catabolism; knockout mice carry fifteen-fold higher brain anandamide, show reduced pain sensation reversed by the CB1 antagonist SR141716A, and respond to administered anandamide with hypomotility, analgesia, catalepsy and hypothermia. Mouse data.

14. 14
    Kerbrat A, Ferre JC, Fillatre P, Ronziere T, Vannier S, Carsin-Nicol B, Lavoue S, Verin M, Gauvrit JY, Le Tulzo Y, Edan G. Acute Neurologic Disorder from an Inhibitor of Fatty Acid Amide Hydrolase.N Engl J Med · 2016 · 375(18):1717-1725doi:10.1056/NEJMoa1604221 · PMID 27806235

    The clinical report of the Rennes phase 1 accident with BIA 10-2474. Eighty-four volunteers had received single doses of 0.25 to 100 mg and repeated doses of 2.5 to 20 mg for ten days without severe events; in the final cohort of six on 50 mg daily and two on placebo, three of the four who consented developed from day five a syndrome of headache, cerebellar signs, memory impairment and altered consciousness with symmetric pontine and hippocampal lesions, one becoming brain dead. The paper describes the drug as a reversible inhibitor and states the mechanism of toxicity is unknown.

15. 15
    Rocha JF, Santos A, Gama H, Moser P, Falcao A, Pressman P, Wallace Hayes A, Soares-da-Silva P. Safety, Tolerability, and Pharmacokinetics of FAAH Inhibitor BIA 10-2474: A Double-Blind, Randomized, Placebo-Controlled Study in Healthy Volunteers.Clin Pharmacol Ther · 2022 · 111(2):391-403doi:10.1002/cpt.2290 · PMID 33998672

    The sponsor's full account of the trial: 116 participants, well tolerated to 100 mg as a single dose and 20 mg once daily for ten days, central nervous system adverse events in five of six receiving 50 mg repeatedly with one fatal outcome, linear pharmacokinetics, terminal half-life 8 to 10 hours, reversible dose-related increases in plasma fatty acid amides, and the conclusion that the data do not provide a complete mechanistic explanation. Five of the eight authors are employees of BIAL, the manufacturer; the others are at Coimbra, the Daedalus Foundation and the University of South Florida.

16. 16
    Habib AM, Okorokov AL, Hill MN, Bras JT, Lee MC, Li S, Gossage SJ, van Drimmelen M, Morena M, Houlden H, Ramirez JD, Bennett DLH, Srivastava D, Cox JJ. Microdeletion in a FAAH pseudogene identified in a patient with high anandamide concentrations and pain insensitivity.Br J Anaesth · 2019 · 123(2):e249-e253doi:10.1016/j.bja.2019.02.019 · PMID 30929760

    A 66-year-old woman who required no postoperative analgesic other than one gram of intravenous acetaminophen after trapeziectomy, with pain scores of 0 out of 10 until discharge and a lifelong history of painless cuts and burns that healed quickly. She carries a microdeletion of about 8 kilobases in the FAAH-OUT pseudogene together with a common hypomorphic variant of the enzyme gene. Against four controls, two of them carriers of that common variant, circulating anandamide was 70 percent higher and palmitoylethanolamide and oleoylethanolamine roughly tripled, while 2-arachidonoylglycerol was largely unchanged. The authors describe the disorder as one of hypoalgesia, altered fear and memory symptoms and a non-anxious disposition; the memory symptoms are long-standing lapses the patient reports herself. A single case report.

17. 17
    Leung D, Saghatelian A, Simon GM, Cravatt BF. Inactivation of N-acyl phosphatidylethanolamine phospholipase D reveals multiple mechanisms for the biosynthesis of endocannabinoids.Biochemistry · 2006 · 45(15):4720-6doi:10.1021/bi060163l · PMID 16605240

    Knockout mice showed more than a five-fold reduction in calcium-dependent conversion of precursors bearing both saturated and polyunsaturated chains, but only the saturated N-acylethanolamines fell in level; a calcium-independent activity accepting the anandamide precursor was found in the knockout brains. The key negative result behind the statement that there is no single biosynthetic switch for anandamide.

18. 18
    Willoughby KA, Moore SF, Martin BR, Ellis EF. The biodisposition and metabolism of anandamide in mice.J Pharmacol Exp Ther · 1997 · 282(1):243-7PMID 9223560

    Male mice received tritiated anandamide with 50 mg/kg of unlabeled anandamide intravenously. Anandamide was detectable in brain within 1 minute; at 1 and 5 minutes most of the brain label had already been converted to arachidonic acid and other polar metabolites, and there were almost no detectable brain concentrations of anandamide at 15 and 30 minutes. The source of the time course quoted on this page. Mouse data at a high dose; the PubMed record carries no DOI.

19. 19
    Mechoulam R, Ben-Shabat S, Hanus L, Ligumsky M, Kaminski NE, Schatz AR, Gopher A, Almog S, Martin BR, Compton DR. Identification of an endogenous 2-monoglyceride, present in canine gut, that binds to cannabinoid receptors.Biochem Pharmacol · 1995 · 50(1):83-90doi:10.1016/0006-2952(95)00109-d · PMID 7605349

    Isolation of 2-arachidonoylglycerol from canine intestine, with inhibition constants of 472 plus or minus 55 nanomolar at CB1 and 1,400 plus or minus 172 at CB2, inhibition of adenylate cyclase in mouse spleen cells at about the potency of tetrahydrocannabinol, and the characteristic tetrad of effects in mice after intravenous administration.

20. 20
    van Esbroeck ACM, Janssen APA, Cognetta AB, Ogasawara D, Shpak G, van der Kroeg M, Kantae V, Baggelaar MP, de Vrij FMS, Deng H, Allara M, Fezza F, Lin Z, van der Wel T, Soethoudt M, Mock ED, den Dulk H, Baak IL, Florea BI, Hendriks G, De Petrocellis L, Overkleeft HS, Hankemeier T, De Zeeuw CI, Di Marzo V, Maccarrone M, Cravatt BF, Kushner SA, van der Stelt M. Activity-based protein profiling reveals off-target proteins of the FAAH inhibitor BIA 10-2474.Science · 2017 · 356(6342):1084-1087doi:10.1126/science.aaf7497 · PMID 28596366

    Activity-based proteomics in human cells and tissues found that BIA 10-2474 inhibits several lipases not targeted by the selective, clinically tested inhibitor PF-04457845, and that BIA 10-2474 but not PF-04457845 substantially altered lipid networks in human cortical neurons. The authors state this suggests, rather than proves, that promiscuous lipase inhibition can cause metabolic dysregulation in the nervous system.

21. 21
    Catalano A, Fezza F, Duranti A, Cavalluzzi MM, Lentini G, Mangiatordi GF, Rocchi MBL, Maccarrone M. Fatty acid amide hydrolase (FAAH) inhibitor design and preclinical studies: lessons learned from BIA 10-2474 and the challenges ahead.Expert Opin Drug Discov · 2026 · 21(9):1019-1036doi:10.1080/17460441.2026.2716127 · PMID 42615850

    A 2026 review of the whole program. It states that the exact mechanism of BIA 10-2474's toxicity has not been fully established, and that rodent and human forms of the enzyme have different three-dimensional structures and interact differently with exogenous molecules, so candidate drugs should be tested on the human enzyme or on human specimens. A narrative review built from a literature search, not new data.

22. 22
    Mikaeili H, Habib AM, Yeung CW, Santana-Varela S, Luiz AP, Panteleeva K, Zuberi S, Athanasiou-Fragkouli A, Houlden H, Wood JN, Okorokov AL, Cox JJ. Molecular basis of FAAH-OUT-associated human pain insensitivity.Brain · 2023 · 146(9):3851-3865doi:10.1093/brain/awad098 · PMID 37222214

    The mechanism behind the case: disruption of FAAH-OUT transcription leads to DNMT1-dependent DNA methylation within the FAAH promoter; FAAH-OUT also contains a conserved regulatory element, FAAH-AMP, acting as an enhancer for FAAH expression; transcriptomic analysis of patient-derived cells uncovered a dysregulated gene network. Cell and molecular work built on one patient.

23. 23
    Fuss J, Steinle J, Bindila L, Auer MK, Kirchherr H, Lutz B, Gass P. A runner's high depends on cannabinoid receptors in mice.Proc Natl Acad Sci U S A · 2015 · 112(42):13105-8doi:10.1073/pnas.1514996112 · PMID 26438875

    Running raises blood beta-endorphin and anandamide in mice; pharmacological, genetic and behavioral work showed that post-exercise anxiolysis depends on CB1 receptors on forebrain GABAergic neurons and analgesia on peripheral CB1 and CB2, while sedation was unaffected by cannabinoid or opioid blockade. The authors note euphoria cannot be studied in a mouse.

24. 24
    Siebers M, Biedermann SV, Bindila L, Lutz B, Fuss J. Exercise-induced euphoria and anxiolysis do not depend on endogenous opioids in humans.Psychoneuroendocrinology · 2021 · 126:105173doi:10.1016/j.psyneuen.2021.105173 · PMID 33582575

    Sixty-three participants, double-blind randomized placebo-controlled, naltrexone or placebo before 45 minutes of treadmill running at moderate intensity or walking. Running raised euphoria, lowered anxiety and raised plasma anandamide and 2-arachidonoylglycerol; opioid blockade prevented none of it, and the fraction reporting a runner's high was the same in both arms. An exclusion of opioids, not a demonstration that endocannabinoids cause the state.

25. 25
    Russo EB. Clinical Endocannabinoid Deficiency Reconsidered: Current Research Supports the Theory in Migraine, Fibromyalgia, Irritable Bowel, and Other Treatment-Resistant Syndromes.Cannabis Cannabinoid Res · 2016 · 1(1):154-165doi:10.1089/can.2016.0009 · PMID 28861491

    The fullest statement of the clinical endocannabinoid deficiency hypothesis, first proposed in 2001: that an underlying endocannabinoid tone exists and that congenital or acquired deficiency of it produces migraine, fibromyalgia, irritable bowel syndrome and related conditions. The author acknowledges the original basis was genetic overlap, comorbidity, symptom patterns and the benefit of exogenous cannabinoids, without objective proof or formal trial data. His listed affiliation is PHYTECS, a commercial endocannabinoid company, which is the relevant conflict of interest.

26. 26
    Sarchielli P, Pini LA, Coppola F, Rossi C, Baldi A, Mancini ML, Calabresi P. Endocannabinoids in chronic migraine: CSF findings suggest a system failure.Neuropsychopharmacology · 2007 · 32(6):1384-90doi:10.1038/sj.npp.1301246 · PMID 17119542

    Cerebrospinal fluid anandamide was significantly lower, and palmitoylethanolamide slightly but significantly higher, in patients with chronic migraine and in patients with probable chronic migraine and probable analgesic-overuse headache than in non-migraine controls (p<0.01 and p<0.02). Anandamide correlated negatively with calcitonin gene-related peptide in both patient groups. This is the measured reduction the deficiency hypothesis rests on; it is a cross-sectional comparison and establishes no direction.

27. 27
    Venkatesan T, Zadvornova Y, Raff H, Hillard CJ. Endocannabinoid-related lipids are increased during an episode of cyclic vomiting syndrome.Neurogastroenterol Motil · 2016 · 28(9):1409-18doi:10.1111/nmo.12843 · PMID 27098832

    Twenty-two patients with cyclic vomiting syndrome in well and sick phases against 12 matched controls. Contrary to the authors' own hypothesis, serum endocannabinoids did not differ between groups; N-oleoylethanolamine and N-palmitoylethanolamide were higher in the sick phase (p=0.001 and p=0.04) and correlated with nausea and poor sleep; salivary cortisol and alpha amylase were higher in marijuana users than non-users during the sick phase. A direct negative test of the deficiency idea in this condition.

28. 28
    Ochandiano I, Powlowski P, Andreu H, Olivier L, Cabeza G, Gimenez-Palomo A, Guitart-Mampel M, Garrabou G, Escelsior A, Bioque M, Valenti M, Berk M, Andreazza AC, Vieta E, Ross RA, Pacchiarotti I. Plasma endocannabinoid levels in bipolar disorder: Differences between manic and euthymic states and the impact of cannabis use.J Affect Disord · 2026 · 415:122483doi:10.1016/j.jad.2026.122483 · PMID 42735776

    Twenty-three patients with bipolar I disorder followed from acute mania to stabilization. Non-cannabis users showed a significant decline in both anandamide and 2-arachidonoylglycerol with recovery (p<0.05); cannabis users showed no significant reduction despite clinical improvement (interaction p=0.016 and p=0.029). Neither lipid was associated with symptom severity on the mania rating scale. A small longitudinal cohort.

29. 29
    Engeli S, Bohnke J, Feldpausch M, Gorzelniak K, Janke J, Batkai S, Pacher P, Harvey-White J, Luft FC, Sharma AM, Jordan J. Activation of the peripheral endocannabinoid system in human obesity.Diabetes · 2005 · 54(10):2838-43doi:10.2337/diabetes.54.10.2838 · PMID 16186383

    Circulating anandamide and 1/2-arachidonoylglycerol were 35 and 52 percent higher in 20 obese than 20 lean women (p<0.05); adipose tissue messenger RNA was 34 percent lower for CB1 and 59 percent lower for the degrading enzyme, with a strong negative correlation between enzyme expression and circulating endocannabinoids; 5 percent weight loss in 17 further women changed none of it. Cross-sectional and associative.

30. 30
    Giuffrida A, Leweke FM, Gerth CW, Schreiber D, Koethe D, Faulhaber J, Klosterkotter J, Piomelli D. Cerebrospinal anandamide levels are elevated in acute schizophrenia and are inversely correlated with psychotic symptoms.Neuropsychopharmacology · 2004 · 29(11):2108-14doi:10.1038/sj.npp.1300558 · PMID 15354183

    Cerebrospinal fluid anandamide was eight-fold higher in 47 antipsychotic-naive first-episode paranoid schizophrenia patients than in 84 healthy controls, 13 dementia patients or 22 affective disorder patients; the elevation was absent in 37 patients on typical antipsychotics but not in 34 on atypical ones, and within the unmedicated acute group it correlated negatively with psychotic symptoms (rS = -0.452, p=0.001). The authors read it as a compensatory adaptation.

31. 31
    Schmid PC, Paria BC, Krebsbach RJ, Schmid HH, Dey SK. Changes in anandamide levels in mouse uterus are associated with uterine receptivity for embryo implantation.Proc Natl Acad Sci U S A · 1997 · 94(8):4188-92doi:10.1073/pnas.94.8.4188 · PMID 9108127

    The periimplantation mouse uterus contained 142 to 1,345 picomoles per micromole of lipid phosphorus, roughly 1 to 7 micrograms per gram wet weight, the highest levels then found in any mammalian tissue. Levels were highest in the non-receptive pseudopregnant uterus and between implantation sites and lowest at implantation sites; anandamide reduced zona-hatching in vitro and the synthetic agonist CP 55,940 inhibited implantation, both reversed by a CB1 antagonist. Mouse data.

32. 32
    Hillard CJ. Circulating Endocannabinoids: From Whence Do They Come and Where are They Going?Neuropsychopharmacology · 2018 · 43(1):155-172doi:10.1038/npp.2017.130 · PMID 28653665

    A review of human studies of circulating endocannabinoids. Concentrations track energy intake and storage, exercise, sleep and its disruption, inflammation, pain and stress, and 2-arachidonoylglycerol is strongly circadian; the conclusion is that the number of contributing physiological factors makes circulating concentrations difficult to use in isolation as a biomarker for any one disorder.

33. 33
    Devane WA, Dysarz FA, Johnson MR, Melvin LS, Howlett AC. Determination and characterization of a cannabinoid receptor in rat brain.Mol Pharmacol · 1988 · 34(5):605-13PMID 2848184

    The demonstration that a cannabinoid receptor exists. Tritiated CP-55,940 bound rat brain P2 membranes saturably with a dissociation constant of 133 picomolar and a maximum of 1.85 picomoles per milligram of protein; binding was displaced by delta-9-tetrahydrocannabinol and by less than 50 percent by cannabidiol or cannabigerol at 1 micromolar, was stereoselective, and was loosened by a non-hydrolyzable GTP analog. Rat tissue; the PubMed record carries no DOI.

34. 34
    Wang J, Okamoto Y, Morishita J, Tsuboi K, Miyatake A, Ueda N. Functional analysis of the purified anandamide-generating phospholipase D as a member of the metallo-beta-lactamase family.J Biol Chem · 2006 · 281(18):12325-35doi:10.1074/jbc.M512359200 · PMID 16527816

    The purified recombinant enzyme is activated by millimolar magnesium or calcium and so appears constitutively active, is highly specific for N-acylphosphatidylethanolamines without selectivity among acyl species, carries a catalytically important zinc, and loses activity on mutation of any of eight conserved residues; the natural variants L207F and H380R were also low in activity.

35. 35
    Tanimura A, Yamazaki M, Hashimotodani Y, Uchigashima M, Kawata S, Abe M, Kita Y, Hashimoto K, Shimizu T, Watanabe M, Sakimura K, Kano M. The endocannabinoid 2-arachidonoylglycerol produced by diacylglycerol lipase alpha mediates retrograde suppression of synaptic transmission.Neuron · 2010 · 65(3):320-7doi:10.1016/j.neuron.2010.01.021 · PMID 20159446

    Retrograde synaptic suppression was totally absent in cerebellum, hippocampus and striatum of mice lacking diacylglycerol lipase alpha and intact in mice lacking the beta form; basal 2-arachidonoylglycerol was markedly reduced and stimulus-evoked elevation absent only in the alpha knockouts. The clean assignment of retrograde signaling to 2-arachidonoylglycerol rather than anandamide, in mouse brain slices.

36. 36
    Pertwee RG, Howlett AC, Abood ME, Alexander SPH, Di Marzo V, Elphick MR, Greasley PJ, Hansen HS, Kunos G, Mackie K, Mechoulam R, Ross RA. International Union of Basic and Clinical Pharmacology. LXXIX. Cannabinoid receptors and their ligands: beyond CB1 and CB2.Pharmacol Rev · 2010 · 62(4):588-631doi:10.1124/pr.110.003004 · PMID 21079038

    The nomenclature committee's review. It sets criteria any proposed third cannabinoid receptor should meet, concludes that no non-CB1, non-CB2 target currently meets them, and identifies TRPV1 as a target that possibly functions as an ionotropic cannabinoid receptor under physiological or pathological conditions and should be investigated further.

37. 37
    Huggins JP, Smart TS, Langman S, Taylor L, Young T. An efficient randomised, placebo-controlled clinical trial with the irreversible fatty acid amide hydrolase-1 inhibitor PF-04457845, which modulates endocannabinoids but fails to induce effective analgesia in patients with pain due to osteoarthritis of the knee.Pain · 2012 · 153(9):1837-1846doi:10.1016/j.pain.2012.04.020 · PMID 22727500

    Seventy-four patients, crossover, 4 mg daily of the inhibitor against naproxen 500 mg twice daily and placebo. Mean differences from placebo in WOMAC pain score were 0.04 (80 percent confidence interval minus 0.63 to 0.71) for the inhibitor and minus 1.13 (minus 1.79 to minus 0.47) for naproxen; stopped at interim analysis for futility despite more than 96 percent enzyme inhibition and substantial increases in four fatty acid amides. Conducted and reported by Pfizer employees.

38. 38
    Fowler CJ. The endocannabinoid system - current implications for drug development.J Intern Med · 2021 · 290(1):2-26doi:10.1111/joim.13229 · PMID 33348434

    A review stating that fatty acid amide hydrolase inhibitors were, in themselves, not useful for the treatment of pain, while suggesting they may be useful for post-traumatic stress disorder and cannabis use disorder, and noting the gap between patients' perceived benefit from cannabis in chronic non-cancer pain and the at best modest effect in meta-analyses of randomized trials. Written before the 2025 post-traumatic stress disorder trial.

39. 39
    Elmes MW, Kaczocha M, Berger WT, Leung K, Ralph BP, Wang L, Sweeney JM, Miyauchi JT, Tsirka SE, Ojima I, Deutsch DG. Fatty acid-binding proteins (FABPs) are intracellular carriers for delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD).J Biol Chem · 2015 · 290(14):8711-21doi:10.1074/jbc.M114.618447 · PMID 25666611

    At least three human fatty acid-binding proteins bind tetrahydrocannabinol and cannabidiol, and both compounds inhibit cellular uptake and catabolism of anandamide by targeting them. Critically, in contrast to the rodent enzyme, cannabidiol does not inhibit human fatty acid amide hydrolase, so enzyme inhibition cannot explain the rise in circulating anandamide after cannabidiol in people; site-directed mutagenesis identified the residues conferring the species difference.

40. 40
    Beltramo M, Stella N, Calignano A, Lin SY, Makriyannis A, Piomelli D. Functional role of high-affinity anandamide transport, as revealed by selective inhibition.Science · 1997 · 277(5329):1094-7doi:10.1126/science.277.5329.1094 · PMID 9262477

    AM404 inhibited high-affinity anandamide accumulation in rat neurons and astrocytes without activating cannabinoid receptors or inhibiting hydrolysis, and enhanced anandamide responses in vitro and in vivo, taken as evidence for carrier-mediated transport. A partly cytosolic variant of the degrading enzyme, FLAT, was proposed as an anandamide carrier in 2012 [67]; no dedicated membrane transporter has been confirmed.

41. 41
    De Petrocellis L, Bisogno T, Davis JB, Pertwee RG, Di Marzo V. Overlap between the ligand recognition properties of the anandamide transporter and the VR1 vanilloid receptor: inhibitors of anandamide uptake with negligible capsaicin-like activity.FEBS Lett · 2000 · 483(1):52-6doi:10.1016/s0014-5793(00)02082-2 · PMID 11033355

    Four established uptake inhibitors, AM404, arvanil, olvanil and linvanil, activate the human vanilloid receptor at concentrations 400 to 10,000-fold lower than those needed to inhibit anandamide transport; three new compounds inhibit transport without that activity. The paper that undermines the interpretation of the transport experiments done with these compounds since 1997.

42. 42
    Maccarrone M, Valensise H, Bari M, Lazzarin N, Romanini C, Finazzi-Agro A. Relation between decreased anandamide hydrolase concentrations in human lymphocytes and miscarriage.Lancet · 2000 · 355(9212):1326-9doi:10.1016/S0140-6736(00)02115-2 · PMID 10776746

    In 50 pregnant women at 6 to 11 weeks, the 7 who miscarried had lymphocyte anandamide hydrolase activity of 60.43 (SD 29.34) against 169.60 (30.20) picomoles per minute per milligram of protein in the 43 who did not (p<0.0001). In a second group of 120 women at 7 to 8 weeks, all 15 below a defined threshold miscarried against 1 of 105 above it (p<0.0001). A single-center study. The same laboratory later reported a third cohort of 100 women in which low lymphocyte enzyme activity again went with spontaneous abortion [76], but no independent replication of this separation was found in the searches behind this page, and no such test is in routine obstetric use.

43. 43
    Sipe JC, Chiang K, Gerber AL, Beutler E, Cravatt BF. A missense mutation in human fatty acid amide hydrolase associated with problem drug use.Proc Natl Acad Sci U S A · 2002 · 99(12):8394-9doi:10.1073/pnas.082235799 · PMID 12060782

    The 385C to A polymorphism converts proline 129 to threonine, producing an enzyme with normal catalytic properties but enhanced sensitivity to proteolytic degradation; in homozygous form it was strongly associated with street drug use and problem drug and alcohol use. A single genetic association study from 2002, with all the replication caveats that implies.

44. 44
    Habib AM, Cox JJ, Okorokov AL. Out of the dark: the emerging roles of lncRNAs in pain.Trends Genet · 2024 · 40(8):694-705doi:10.1016/j.tig.2024.04.009 · PMID 38926010

    A review placing FAAH-OUT as the first long non-coding RNA associated with an inherited human pain insensitivity disorder and surveying others implicated in chronic pain. Used here for that framing claim only.

45. 45
    Mayo LM, Gauffin E, Petrie GN, Tansey R, Mazurka R, Haggarty CJ, Jones MR, Engelbrektsson H, Aminoff V, Huhne-Landgraf A, Schmidt ME, Pemberton DJ, Fredlund C, Ostman L, Karlsson H, Lofberg A, Pietrzak M, Andersson G, Capusan AJ, Hill MN, Heilig M. The efficacy of elevating anandamide via inhibition of fatty acid amide hydrolase (FAAH) combined with internet-delivered cognitive behavioral therapy in the treatment of post-traumatic stress disorder: a randomized, placebo-controlled clinical trial.Neuropsychopharmacology · 2025 · 50(10):1564-1572doi:10.1038/s41386-025-02128-w · PMID 40382500

    One hundred patients with post-traumatic stress disorder, 85 women, randomized to JNJ-42165279 25 mg twice daily or placebo for twelve weeks with internet-delivered cognitive behavioral therapy from week five. The drug raised anandamide and had no effect on the clinician-assessed primary outcome or on any secondary measure. A negative trial, with two authors employed by Janssen Pharmaceutica.

46. 46
    Gaudette EV, Green DGJ, Watling SE, Dheda SS, Lora SJ, Zilberman M, Hill MN, Tyndale RF, Richardson JD, McCluskey T, Narciso L, Rusjan PM, Kloiber S, Houle S, Warsh JJ, Boileau I. In Vivo Characterization of Fatty Acid Amide Hydrolase in Posttraumatic Stress Disorder: A [11C]CURB Positron Emission Tomography Study.Biol Psychiatry · 2026 · in pressdoi:10.1016/j.biopsych.2026.09.008 · PMID 42772587

    Eighty-six participants, 35 with post-traumatic stress disorder and 51 controls. Amygdala enzyme levels were 10 percent lower in the disorder (p=0.02), with no differences in other corticolimbic regions and no association with clinical symptoms; in exploratory analyses childhood trauma was associated with 11.6 percent higher corticolimbic levels regardless of diagnosis (p=0.014, 28 versus 24 participants). The direction is opposite to the preclinical prediction, and the authors note it may inform the limited efficacy of inhibitor trials.

47. 47
    Hanske A, Nazare M, Grether U. Chemical Probes for Investigating the Endocannabinoid System.Curr Top Behav Neurosci · 2026 · 76:69-144doi:10.1007/7854_2024_563 · PMID 39747798

    A 2026 review of the chemical tools used to study this system, and the current statement of which enzymes matter: anandamide is degraded mainly by fatty acid amide hydrolase or by N-acylethanolamine acid amidase into ethanolamine and arachidonic acid, inhibitors of the second reduce inflammation and pain in animal models, 2-arachidonoylglycerol is made by diacylglycerol lipase and degraded by monoacylglycerol lipase, and the biosynthetic enzyme for anandamide is named N-acylphosphatidylethanolamine-specific phospholipase D, abbreviated NAPE-PLD. A narrative review; one of the three authors is employed by Roche.

48. 48
    Wei BQ, Mikkelsen TS, McKinney MK, Lander ES, Cravatt BF. A second fatty acid amide hydrolase with variable distribution among placental mammals.J Biol Chem · 2006 · 281(48):36569-78doi:10.1074/jbc.M606646200 · PMID 17015445

    Discovery of FAAH-2 in humans, sharing 20 percent sequence identity with FAAH-1, present in multiple primate genomes, marsupials and more distant vertebrates but absent from mouse and rat; FAAH-1 is much more active against N-acyl ethanolamines such as anandamide. The paper that states the species caveat for extrapolating this system across mammals.

49. 49
    Felder CC, Joyce KE, Briley EM, Mansouri J, Mackie K, Blond O, Lai Y, Ma AL, Mitchell RL. Comparison of the pharmacology and signal transduction of the human cannabinoid CB1 and CB2 receptors.Mol Pharmacol · 1995 · 48(3):443-50PMID 7565624

    In cells stably transfected with the human CB1 or CB2 receptor, anandamide was essentially equipotent at the two subtypes in competition binding, and the rank order of potency in binding was reproduced in functional inhibition of cyclic AMP accumulation. CB1, but not CB2, coupled to the modulation of Q-type calcium and inwardly rectifying potassium channels. Recombinant cells rather than tissue; the PubMed record carries no DOI.

50. 50
    Hayes AW, Weber K, Moser P, Soares-da-Silva P. Non-clinical toxicology evaluation of BIA 10-2474.Crit Rev Toxicol · 2021 · 51(1):65-75doi:10.1080/10408444.2020.1867821 · PMID 33528291

    A review of the regulatory animal studies concluding that the central nervous system effects seen in rodents were of a kind common in such studies and did not constitute a signal, and that the fatality was unpredictable from a non-clinical toxicological perspective. Two of the four authors are BIAL staff, which is the relevant conflict of interest.

51. 51
    Cerqueira NMFSA, Neves M, Rocha J, Soares-da Silva P, Palma PN. Inactivation Mechanism of the Fatty Acid Amide Hydrolase Inhibitor BIA 10-2474.Chembiochem · 2022 · 23(20):e202200166doi:10.1002/cbic.202200166 · PMID 35843872

    Quantum mechanical and molecular modeling work describing the inhibition as target-specific covalent binding to the catalytic nucleophile, with hydrolysis of the adduct unlikely under physiological conditions and therefore irreversible inactivation; it identifies a role for the FAAH-specific residue Thr236. This corrects the description of the drug as reversible in the clinical report. Four of five authors are at BIAL.

52. 52
    D'Souza DC, Cortes-Briones J, Creatura G, Bluez G, Thurnauer H, Deaso E, Bielen K, Surti T, Radhakrishnan R, Gupta A, Gupta S, Cahill J, Sherif MA, Makriyannis A, Morgan PT, Ranganathan M, Skosnik PD. Efficacy and safety of a fatty acid amide hydrolase inhibitor (PF-04457845) in the treatment of cannabis withdrawal and dependence in men: a double-blind, placebo-controlled, parallel group, phase 2a single-site randomised controlled trial.Lancet Psychiatry · 2019 · 6(1):35-45doi:10.1016/S2215-0366(18)30427-9 · PMID 30528676

    Seventy men with cannabis dependence, 46 randomized to 4 mg daily and 24 to placebo. Withdrawal symptom scores were lower on the drug than on placebo on the first two days of treatment (6.04 against 11.00, p=0.048; 6.02 against 11.74, p=0.035), self-reported use at four weeks was 0.40 against 1.27 joints per day (p=0.0003), and urinary THC-COOH, the carboxylic acid metabolite of tetrahydrocannabinol that drug screens detect, was 265.55 against 657.92 nanograms per milliliter (p=0.009). No serious adverse events. Single site, men only, funded by the United States National Institute on Drug Abuse.

53. 53
    Jones MR, Haggarty CJ, Petrie GN, Lunge AR, Morrison I, Hill MN, Heilig M, Mayo LM. Endocannabinoid contributions to the perception of socially relevant, affective touch in humans.Neuropsychopharmacology · 2025 · 50(5):849-855doi:10.1038/s41386-025-02053-y · PMID 39843850

    Two studies. In 101 adults recruited for the presence (52) or absence (49) of documented childhood maltreatment, higher anandamide was associated with a reduced preference for the slow, light stroking that C-tactile afferents respond to. In a randomized experiment, 16 healthy participants received the inhibitor PF-04457845 and 29 received placebo; raising anandamide reduced the rated pleasantness of that touch, with no effect on ratings of emotional images. Both results ran contrary to the authors' stated hypothesis. Small samples, and the first study is an association.

54. 54
    Schmidt ME, Gargano C, Zhou X, Palmer JA, Saad ZS, Vieta E, Drevets WC, Stuyckens K, Pandina G. Efficacy and safety of adjunctive treatment with the fatty acid amide hydrolase inhibitor JNJ-42165279 in participants with major depressive disorder with anxious distress: A double-blind, placebo-controlled, randomised study.Eur Neuropsychopharmacol · 2026 · 106:112771doi:10.1016/j.euroneuro.2026.112771 · PMID 41616640

    A phase 2a double-blind randomized placebo-controlled trial of JNJ-42165279 at 25 mg daily added to an existing antidepressant in 153 participants aged 18 to 64 with major depressive disorder with anxious distress and an inadequate response to a serotonin or serotonin-noradrenaline reuptake inhibitor. There was no treatment effect on the primary endpoint, change in the 17-item Hamilton Depression Rating Scale at week 6 (least-squares mean difference minus 0.2, standard error 1.04, one-sided p=0.416), and none on the key secondary endpoints, although plasma fatty acid amides rose substantially and correlated strongly with drug concentration. Eight of the nine authors are employees of Johnson and Johnson, which makes the compound, and the result is negative.

55. 55
    Tansey R, Perini I, Petrie GN, Haggarty CJ, Mazurka R, Yngve A, Mina S, Jones MR, Engelbrektsson H, Capusan AJ, Hill MN, Heilig M, Mayo LM. Functional neuroimaging of fatty acid amide hydrolase inhibition in posttraumatic stress disorder: a randomized clinical trial.Transl Psychiatry · 2026 · 16(1)doi:10.1038/s41398-026-03864-3 · PMID 41651806

    The imaging substudy of the 100-patient post-traumatic stress disorder trial, with 76 participants scanned. Four weeks of enzyme inhibition significantly raised anandamide, but there were no differences between treatment groups in resting-state functional connectivity or in brain activation during emotional processing. Greater symptom improvement, irrespective of drug, went with lower connectivity between ventromedial and right dorsolateral prefrontal cortex and lower right dorsolateral prefrontal activation. A negative mechanistic result reported by the trial's own investigators.

56. 56
    Rayasam A, Hillard CJ. Endocannabinoid Signaling and Mood Regulation.Annu Rev Pharmacol Toxicol · 2026 · online ahead of printdoi:10.1146/annurev-pharmtox-052825-085837 · PMID 42862946

    The current state of the field in one sentence of its own abstract: fatty acid amide hydrolase inhibitors have largely failed to demonstrate efficacy and CB1 agonists have only modest effects in people with mood disorders, and progress will require understanding the inconsistencies between model systems and patients. A review.

57. 57
    Dincheva I, Drysdale AT, Hartley CA, Johnson DC, Jing D, King EC, Ra S, Gray JM, Yang R, DeGruccio AM, Huang C, Cravatt BF, Glatt CE, Hill MN, Casey BJ, Lee FS. FAAH genetic variation enhances fronto-amygdala function in mouse and human.Nat Commun · 2015 · 6:6395doi:10.1038/ncomms7395 · PMID 25731744

    A knock-in mouse recapitulating the human C385A variant, compared directly with human carriers. Both show reduced enzyme expression and higher anandamide, selectively enhanced fronto-amygdala connectivity, better fear extinction learning and less anxiety-like behavior. The strongest cross-species evidence that anandamide concentration changes human fear regulation.

58. 58
    Boecker H, Sprenger T, Spilker ME, Henriksen G, Koppenhoefer M, Wagner KJ, Valet M, Berthele A, Tolle TR. The runner's high: opioidergic mechanisms in the human brain.Cereb Cortex · 2008 · 18(11):2523-31doi:10.1093/cercor/bhn013 · PMID 18296435

    Ten trained athletes were scanned with the non-selective opioid ligand [18F]FDPN at rest and after two hours of endurance running, 21.5 plus or minus 4.7 kilometers. Opioid receptor availability fell preferentially in prefrontal and limbic or paralimbic structures, euphoria rose significantly after running, and euphoria correlated inversely with opioid binding in prefrontal and orbitofrontal cortex, the anterior cingulate cortex, both insulae and temporoparietal regions. A ten-person pilot study in men, and a correlation rather than a demonstration of cause, but it is the reason the endorphin account cannot be called evidence-free.

59. 59
    Sparling PB, Giuffrida A, Piomelli D, Rosskopf L, Dietrich A. Exercise activates the endocannabinoid system.Neuroreport · 2003 · 14(17):2209-11doi:10.1097/00001756-200312020-00015 · PMID 14625449

    Trained male college students running on a treadmill or cycling for 50 minutes at 70 to 80 percent of maximum heart rate; the first human evidence that moderate exercise activates the endocannabinoid system, offered as a candidate mechanism for exercise-induced analgesia. Small, male, trained, and uncontrolled for the many other things exercise changes.

60. 60
    Yabe T, Ando D, Koyama K. Low-intensity exercise reduces negative affect via circulating endocannabinoid dynamics in healthy young adults.Eur J Appl Physiol · 2026 · 126(10):5655-5663doi:10.1007/s00421-026-06364-w · PMID 42489728

    A randomized crossover trial in 30 healthy adults cycling 60 minutes at 30 and at 50 percent of heart rate reserve. Plasma endocannabinoids did not change uniformly and varied widely between individuals; low-intensity exercise reduced negative affect, but that reduction was not correlated with cumulative plasma anandamide or 2-arachidonoylglycerol, and only a within-person repeated-measures correlation found a significant negative association between anandamide and negative affect.

61. 61
    Ruiz-Munoz M, Jimenez-Pavon D, Garrido-Palomino I, Escudier-Vazquez JM, Gomez de Lara EJV, Espana-Romero V. Effects of an internet-based combined exercise and cognitive-behavioral therapy intervention on endocannabinoid system biomarkers and physical fitness in adults with mild-to-moderate depression: a SONRIE randomized controlled trial.Eur Arch Psychiatry Clin Neurosci · 2026 · 276(7):3701-3714doi:10.1007/s00406-026-02330-x · PMID 42550192

    Eighty adults randomized to a 12-week combined exercise and internet-delivered cognitive behavioral therapy program or control. No significant time-by-group interaction for any of nine endocannabinoid system biomarkers, anandamide included (F(2,88)=1.20, p=0.306), while the intervention group walked 81.6 meters farther than controls on a six-minute walk test at 12 weeks, a between-group difference exceeding the minimal clinically important one.

62. 62
    Siebers M, Huvermann D, Siebers C, Canales-Romero D, Florea-Ghile A, Keite L, John L, Munk M, Bizjak R, Witzel J, Schulz S, Kirsten J, Bindila L, Hinney A, Grau M, Bizjak DA, Engler H, Fuss J. Endocannabinoid dynamics across marathon and ultramarathon running: evidence from two field studies.BMC Med · 2026 · 24(1)doi:10.1186/s12916-026-05186-z · PMID 42681668

    Two field studies from the group that ran the naltrexone experiment. Nineteen trained runners completed a marathon and a duration-matched walk, with blood sampled every 14 kilometers and after a 45-minute break, and 36 ultramarathon runners were sampled before and after races of 100, 160 or 230 kilometers. Plasma anandamide rose progressively through the marathon and remained raised after the break, and was raised after every ultramarathon distance, while 2-arachidonoylglycerol rose only in the later stages of the marathon and in early recovery; walking produced modest increases as well. Marathon running produced more euphoria and less anxiety than walking, while ultramarathon running raised pain, lowered anxiety and did not significantly change euphoria. Field studies with no blockade arm, so still correlational.

63. 63
    di Tomaso E, Beltramo M, Piomelli D. Brain cannabinoids in chocolate.Nature · 1996 · 382(6593):677-8doi:10.1038/382677a0 · PMID 8751435

    The original one-page letter reporting anandamide and related compounds in cocoa and chocolate. The PubMed record carries no abstract, so what it fixes here is the date, authorship, journal and scope of the claim; its quantities are not reproduced in the indexed record and are not quoted on this page.

64. 64
    Di Marzo V, Sepe N, De Petrocellis L, Berger A, Crozier G, Fride E, Mechoulam R. Trick or treat from food endocannabinoids?Nature · 1998 · 396(6712):636-7doi:10.1038/25267 · PMID 9872309

    A letter published as a comment on the chocolate report, indexed under cacao, milk and human milk, gas chromatography-mass spectrometry and food analysis. It carries no abstract; it is cited here to establish that the significance of food endocannabinoids was publicly disputed within two years by a group including Mechoulam.

65. 65
    Tytgat J, Van Boven M, Daenens P. Cannabinoid mimics in chocolate utilized as an argument in court.Int J Legal Med · 2000 · 113(3):137-9doi:10.1007/s004140050284 · PMID 10876983

    A defense lawyer argued that a positive urine cannabis immunoassay came from eating a large quantity of chocolate containing anandamide-related lipids. The toxicologists synthesized N-oleoyl- and N-linoleoylethanolamide, spiked them with anandamide into urine, and found no sample positive: no cross-reactivity with the cannabinoid assay. The claim was refuted and the accused convicted.

66. 66
    Paria BC, Song H, Wang X, Schmid PC, Krebsbach RJ, Schmid HH, Bonner TI, Zimmer A, Dey SK. Dysregulated cannabinoid signaling disrupts uterine receptivity for embryo implantation.J Biol Chem · 2001 · 276(23):20523-8doi:10.1074/jbc.M100679200 · PMID 11279117

    Preimplantation embryo development became asynchronous in CB1 and/or CB2 deficient mice; uterine anandamide and blastocyst CB1 normally fall together with the onset of receptivity but stayed high in the non-receptive uterus, in dormant blastocysts during delayed implantation, and in leukemia inhibitory factor-deficient mice; a sustained exogenous cannabinoid blocked implantation in wild-type but not in double mutant mice.

67. 67
    Maccarrone M, Valensise H, Bari M, Lazzarin N, Romanini C, Finazzi-Agro A. Progesterone up-regulates anandamide hydrolase in human lymphocytes: role of cytokines and implications for fertility.J Immunol · 2001 · 166(12):7183-9doi:10.4049/jimmunol.166.12.7183 · PMID 11390466

    Physiological concentrations of progesterone raised lymphocyte fatty acid amide hydrolase through increased gene expression at the transcriptional and translational level, partly by way of the Th2 cytokines interleukin-4 and interleukin-10, while the Th1 cytokines interleukin-12 and interferon gamma lowered the enzyme; human chorionic gonadotropin and cortisol had no effect. In a clinical study of 100 healthy women, low lymphocyte enzyme activity correlated with spontaneous abortion. The same laboratory as the Lancet study, so a further cohort rather than independent confirmation.

68. 68
    Fonseca BM, Cunha SC, Goncalves D, Mendes A, Braga J, Correia-da-Silva G, Teixeira NA. Decidual NK cell-derived conditioned medium from miscarriages affects endometrial stromal cell decidualisation: endocannabinoid anandamide and tumour necrosis factor-alpha crosstalk.Hum Reprod · 2020 · 35(2):265-274doi:10.1093/humrep/dez260 · PMID 31990346

    First-trimester human decidua from 18 miscarriages and 8 elective terminations. Anandamide, measured by ultra-performance liquid chromatography with tandem mass spectrometry, was higher in the miscarriage tissue, and conditioned medium from the uterine natural killer cells of those samples carried high tumor necrosis factor alpha and interfered with the differentiation of endometrial stromal cells in culture. An in vitro model on human tissue in which the inflammation of miscarriage was already present before sampling, a limitation the authors name themselves. The authors declare no conflicts of interest.

69. 69
    Fride E, Ginzburg Y, Breuer A, Bisogno T, Di Marzo V, Mechoulam R. Critical role of the endogenous cannabinoid system in mouse pup suckling and growth.Eur J Pharmacol · 2001 · 419(2-3):207-14doi:10.1016/s0014-2999(01)00953-0 · PMID 11426843

    Giving the CB1 antagonist SR141716A to newborn mouse pups, once on day 1 or daily for a week from day 2, devastated milk ingestion and growth, with death in four to eight days; co-administered tetrahydrocannabinol almost fully reversed it in the week-long regimen. Mouse data, with no human counterpart.

70. 70
    Martinez-Sadurni L, Toll A, Olivares-Berjaga D, Rodriguez N, Martinez-Pinteno A, Haro JM, Trabsa A, Berge D, Perez Sola V, Mane A. Endocannabinoid plasma concentrations and their relationship with clinical outcome and cannabis use in antipsychotic-naive first episode psychosis.Psychiatry Res · 2026 · 366:117479doi:10.1016/j.psychres.2026.117479 · PMID 42826464

    Plasma anandamide and 2-arachidonoylglycerol in 58 antipsychotic-naive patients with a first episode of psychosis. Higher baseline anandamide was associated with lower general psychopathology at one year; no significant associations with total symptom scores or with the positive and negative subscales, and 2-arachidonoylglycerol showed only a non-significant trend toward later cannabis use. An observational cohort.

71. 71
    Dlugos A, Childs E, Stuhr KL, Hillard CJ, de Wit H. Acute stress increases circulating anandamide and other N-acylethanolamines in healthy humans.Neuropsychopharmacology · 2012 · 37(11):2416-27doi:10.1038/npp.2012.100 · PMID 22763622

    Seventy-one adults attended two sessions, one containing the Trier Social Stress Test and one a control task. Stress raised serum anandamide and the related N-acylethanolamines immediately after the stress period; the rise in palmitoylethanolamide correlated positively with the rise in cortisol; and before any stressor, higher anxiety ratings went with lower anandamide. Sex, menstrual cycle status and race modified the responses. A within-person human experiment in healthy volunteers, not a disease cohort.

72. 72
    Allen JH, de Moore GM, Heddle R, Twartz JC. Cannabinoid hyperemesis: cyclical hyperemesis in association with chronic cannabis abuse.Gut · 2004 · 53(11):1566-70doi:10.1136/gut.2003.036350 · PMID 15479672

    The case series that named the syndrome. Nineteen patients identified in South Australia, of whom nine plus one previously published case were analyzed: chronic cannabis use predated the vomiting illness in every case, cessation stopped it in seven, three who never abstained stayed ill and three who rechallenged after a period of abstinence relapsed, two of those three recovering when they stopped again, and nine of ten displayed an abnormal washing behavior during active illness. An uncontrolled case series, and the washing behavior is its most durable contribution.

73. 73
    Jimenez-Castillo RA, Arumugam S, Remes-Troche JM, Venkatesan T. Cannabinoid hyperemesis syndrome: A review.Rev Gastroenterol Mex (Engl Ed) · 2025 · 90(2):214-226doi:10.1016/j.rgmxen.2025.02.002 · PMID 40517066

    A narrative review of cannabinoid hyperemesis syndrome: its overlap with cyclic vomiting syndrome, the expectation that sustained cessation of cannabis resolves the episodes, the average delay of three to six years between the first episode and diagnosis in published cases, the role proposed for the endocannabinoid system, the diagnostic criteria and current management. Co-authored by the investigator who ran the negative endocannabinoid measurement study. A review, not new data.

74. 74
    Li BUK. Landscape of Cyclic Vomiting Syndrome: From Bedside to Bench, Past to Present.Neurogastroenterol Motil · 2025 · 37(3):e14990doi:10.1111/nmo.14990 · PMID 39789960

    A 2025 review placing cannabinoid hyperemesis syndrome within cyclic vomiting syndrome and listing the pathways with some evidence behind them, including NK1, CB1, the hypothalamic-pituitary-adrenal axis, the PPM1D gene and estrogen, with TRPV1, CGRP, GDF-15 and mitochondrial dysfunction as possible. It treats the mechanism as unsettled.

75. 75
    Leweke FM, Piomelli D, Pahlisch F, Muhl D, Gerth CW, Hoyer C, Klosterkotter J, Hellmich M, Koethe D. Cannabidiol enhances anandamide signaling and alleviates psychotic symptoms of schizophrenia.Transl Psychiatry · 2012 · 2(3):e94doi:10.1038/tp.2012.15 · PMID 22832859

    A double-blind randomized trial of cannabidiol against the antipsychotic amisulpride in acute schizophrenia. Both were safe and led to significant clinical improvement, cannabidiol with a markedly better side-effect profile, and cannabidiol treatment was accompanied by a significant rise in serum anandamide that was significantly associated with improvement. The proposed mechanism, inhibition of anandamide deactivation, is stated as a suggestion. The indexed abstract gives neither the number of patients nor the size of the clinical improvement, so no figures of that kind are quoted on this page.

76. 76
    Abboud A, Chester LA, Hebert FO, Jutras-Aswad D. Sex-specific association between low oral doses of cannabidiol (CBD) and plasma concentration of anandamide (AEA), N-palmitoylethanolamine (PEA) and N-oleoylethanolamine (OEA) in healthy occasional cannabis users.J Cannabis Res · 2026 · 8(1):20doi:10.1186/s42238-025-00356-x · PMID 41514385

    A triple-blind placebo-controlled randomized crossover trial in 70 healthy occasional cannabis users given oral cannabidiol at 20, 50, 100 and 200 mg and placebo, with lipids measured by liquid chromatography with mass spectrometry. Dose affected the cumulative rise in palmitoylethanolamide and oleoylethanolamine but not in anandamide; placebo gave significantly higher cumulative exposure than the 50 mg dose, and those reductions were confined to the women. Plasma concentrations of all three lipids fell over the course of every visit, on drug and on placebo alike. ClinicalTrials.gov NCT05407285. Single over-the-counter doses in healthy volunteers, not a therapeutic regimen.

77. 77
    Hebert FO, Mongeau-Perusse V, Rizkallah E, Mahroug A, Bakouni H, Morissette F, Brissette S, Bruneau J, Dubreucq S, Jutras-Aswad D. Absence of Evidence for Sustained Effects of Daily Cannabidiol Administration on Anandamide Plasma Concentration in Individuals with Cocaine Use Disorder: Exploratory Findings from a Randomized Controlled Trial.Cannabis Cannabinoid Res · 2024 · 10(2):e341-e352doi:10.1089/can.2023.0273 · PMID 38770686

    An exploratory analysis of a randomized double-blind placebo-controlled trial of 800 mg of cannabidiol daily in cocaine use disorder: 78 participants randomized, 40 to drug and 38 to placebo, 64 analyzed. Mean plasma anandamide was 0.26 against 0.29 nanograms per milliliter at day 8 and 0.27 against 0.30 at week 4, with Bayes factors of 0.190 and 0.194 favoring no effect. ClinicalTrials.gov NCT02559167. An exploratory endpoint, and acute or short-term effects were not excluded.

78. 78
    Kumric M, Dujic G, Vrdoljak J, Supe Domic D, Dujic Z, Bozic J. Can Endocannabinoids Explain CBD-Mediated Reduction in Blood Pressure? Insights from a Randomized, Placebo-Controlled, Crossover Trial.Cannabis Cannabinoid Res · 2025 · online ahead of printdoi:10.1177/25785125251392773 · PMID 41203237

    A secondary analysis of the HYPER-H21-4 crossover trial in 66 patients with primary hypertension, each given five weeks of cannabidiol, 225 to 300 mg daily rising to 375 to 450 mg daily, and five weeks of placebo. Plasma anandamide rose significantly on cannabidiol, from 36.7 plus or minus 18.0 to 47.9 plus or minus 24.6 nanograms per milliliter (p=0.042), with no rise on placebo and no change in 2-arachidonoylglycerol. The change in anandamide did not correlate with the change in systolic pressure (r = minus 0.106, p=0.428), and body mass index, current antihypertensive treatment and fasting glucose predicted the rise. The authors conclude there is no conclusive evidence that the lipid explains the blood pressure effect.

79. 79
    Neumeister A, Normandin MD, Pietrzak RH, Piomelli D, Zheng MQ, Gujarro-Anton A, Potenza MN, Bailey CR, Lin SF, Najafzadeh S, Ropchan J, Henry S, Corsi-Travali S, Carson RE, Huang Y. Elevated brain cannabinoid CB1 receptor availability in post-traumatic stress disorder: a positron emission tomography study.Mol Psychiatry · 2013 · 18(9):1034-40doi:10.1038/mp.2013.61 · PMID 23670490

    Twenty-five untreated patients with non-combat post-traumatic stress disorder, 12 trauma-exposed controls and 23 healthy controls were scanned with the CB1-selective tracer carbon-11 OMAR. Brain-wide receptor availability was 19.5 percent higher than in the healthy controls and 14.5 percent higher than in the trauma-exposed controls (p=0.001), most pronounced in women; peripheral anandamide was 53.1 and 58.2 percent lower; cortisol was lower in both trauma groups; and the three markers together classified nearly 85 percent of cases. Cross-sectional and associative, and the combination of more receptor with less ligand is an inference about tone, not a measurement of it.

80. 80
    Dinh TP, Carpenter D, Leslie FM, Freund TF, Katona I, Sensi SL, Kathuria S, Piomelli D. Brain monoglyceride lipase participating in endocannabinoid inactivation.Proc Natl Acad Sci U S A · 2002 · 99(16):10819-24doi:10.1073/pnas.152334899 · PMID 12136125

    Monoglyceride lipase was cloned from a rat brain library and found in the regions where CB1 is present, with a laminar distribution in hippocampus suggesting a presynaptic location. Overexpressing it in rat cortical neurons reduced stimulus-evoked accumulation of 2-arachidonoylglycerol but not of anandamide, which points to hydrolysis by this enzyme as a primary route of 2-arachidonoylglycerol inactivation. Rat tissue and cultured neurons.

81. 81
    Noriega-Prieto JA, Falcon-Moya R, Noeker JA, Cai R, Fundazuri UB, Eraso-Pichot A, Cai S, Guttipatti P, Belisle L, Rodriguez-Moreno A, van der Stelt M, Li Y, Cheer JF, Marsicano G, Kofuji P, Araque A. Distinct endocannabinoids specifically signal to astrocytes or neurons in the adult mouse hippocampus.Nat Neurosci · 2026 · 29(2):445-454doi:10.1038/s41593-025-02148-1 · PMID 41469443

    In adult mouse hippocampus, 2-arachidonoylglycerol signaled selectively to neurons, producing the presynaptic depression by which the system is usually described, while anandamide signaled to astrocytes, which then potentiated neighboring synapses; anandamide, and not 2-arachidonoylglycerol, together with astrocyte signaling, was required for spike-timing-dependent long-term potentiation. The first assignment of a specific synaptic job to anandamide rather than a background role. Mouse brain slices, with no human counterpart.

82. 82
    Muller-Vahl KR, Fremer C, Beals C, Ivkovic J, Loft H, Schindler C. Endocannabinoid Modulation Using Monoacylglycerol Lipase Inhibition in Tourette Syndrome: A Phase 1 Randomized, Placebo-Controlled Study.Pharmacopsychiatry · 2022 · 55(3):148-156doi:10.1055/a-1675-3494 · PMID 34847610

    An exploratory phase 1b crossover study in 20 adults with Tourette syndrome on standard-of-care medication, comparing a single fasted 40 mg dose of the monoacylglycerol lipase inhibitor Lu AG06466, formerly ABX-1431, with placebo. All scales trended toward tic reduction and two of three reached significance at various timepoints, including the total tic score of the Yale Global Tic Severity Scale; premonitory urges fell significantly. Headache, somnolence and fatigue were the common adverse events. One author is employed by Abide Therapeutics, which developed the compound, and two by Lundbeck.

83. 83
    Muller-Vahl KR, Fremer C, Beals C, Ivkovic J, Loft H, Schindler C. Monoacylglycerol Lipase Inhibition in Tourette Syndrome: A 12-Week, Randomized, Controlled Study.Mov Disord · 2021 · 36(10):2413-2418doi:10.1002/mds.28681 · PMID 34117788

    A 12-week multicenter randomized double-blind placebo-controlled trial of two dose levels of Lu AG06466 in 49 adults with Tourette syndrome. Both treatment groups improved on the total tic score of the Yale Global Tic Severity Scale, but the treatment difference at week 8 of 3.0 points (95 percent confidence interval 0.1 to 5.9, p=0.043) favored placebo, and no secondary endpoint favored the drug; the authors conclude there was no evidence of efficacy in suppressing tics. Company-sponsored and negative, from the same author group as the single-dose study.

84. 84
    Bhattacharyya A, Vasconcelos D, Spicarova D, Palecek J. 20:4-NAPE induced changes of mechanical sensitivity and DRG neurons excitability are concentration dependent and mediated via NAPE-PLD.Sci Rep · 2025 · 15(1):14131doi:10.1038/s41598-025-98567-y · PMID 40269193

    Rat dorsal root ganglion neurons with intrathecal behavioral testing. The anandamide precursor 20:4-NAPE reduced neuronal excitability at 10 nanomolar to 1 micromolar and increased it at 10 micromolar; the NAPE-PLD inhibitor LEI-401 blocked both effects; CB1 blockade prevented the inhibition and TRPV1 blockade the excitation; given into the spinal canal, the low concentration produced hyposensitivity and the high concentration mechanical allodynia. Rat data in male animals.

85. 85
    Fu J, Bottegoni G, Sasso O, Bertorelli R, Rocchia W, Masetti M, Guijarro A, Lodola A, Armirotti A, Garau G, Bandiera T, Reggiani A, Mor M, Cavalli A, Piomelli D. A catalytically silent FAAH-1 variant drives anandamide transport in neurons.Nat Neurosci · 2012 · 15(1):64-9doi:10.1038/nn.2986 · PMID 22101642

    Describes FLAT, a partly cytosolic variant of fatty acid amide hydrolase-1 with no amidase activity that binds anandamide with low micromolar affinity and helps move it into cells; the established transport inhibitors AM404 and OMDM-1 block the effect, and a new antagonist, ARN272, interrupted anandamide deactivation and was analgesic in rodent pain models through CB1. One group's proposal for the molecular identity of anandamide transport, in neural cells and rodents.

86. 86
    Rusjan PM, Knezevic D, Boileau I, Tong J, Mizrahi R, Wilson AA, Houle S. Voxel level quantification of [11C]CURB, a radioligand for Fatty Acid Amide Hydrolase, using high resolution positron emission tomography.PLoS One · 2018 · 13(2):e0192410doi:10.1371/journal.pone.0192410 · PMID 29444138

    Methodological validation of the irreversible radioligand used to image the degrading enzyme in the human brain, including the modeling assumptions and the reduction in variability compared with the standard graphical method. A methods paper; it establishes how the measurement is made, not what it shows in disease.