FAAH — Fatty Acid Amide Hydrolase

What it is and what it hydrolyses in vitro

FAAH was cloned and characterised by Cravatt and colleagues in 1996 as the enzyme degrading neuromodulatory fatty-acid amides. Its substrate list is broader than anandamide, and that breadth matters for reading any inhibitor result. Oleoylethanolamide and palmitoylethanolamide are FAAH substrates with their own pharmacology that is not CB1-mediated at all: PEA acts substantially through PPAR-alpha and is an anti-inflammatory and analgesic agent in its own right, and OEA is a satiety signal. So a FAAH inhibitor raises at least three biologically active lipids simultaneously, only one of which is a cannabinoid receptor agonist. Attributing the whole effect of FAAH inhibition to anandamide at CB1 is a common error and is not supported by the substrate profile.

Sources: Cravatt BF 1996 · Howlett AC 2002

The C385A polymorphism and anandamide tone in humans human data

This is the cleanest human genetics on the shelf. Sipe and colleagues identified a missense single-nucleotide polymorphism in human FAAH, C385A, producing a proline-to-threonine substitution at residue 129. Chiang and colleagues then showed the mechanism: the P129T protein is not catalytically dead but is expressed at reduced levels and shows reduced cellular activity, because the variant is subject to increased proteolytic degradation. The functional consequence is higher anandamide tone in carriers. Dincheva and colleagues brought a knock-in mouse and human carriers together in 2015 and reported that the variant is associated with enhanced fronto-amygdala connectivity, faster habituation to a threat cue, and lower anxiety-related measures. That convergence of a defined molecular lesion, a matching mouse, a matching imaging phenotype and a matching behavioural phenotype is what makes the FAAH-anxiety link one of the better-supported gene-to-behaviour stories in this area.

• C385A (rs324420) is common enough to be studied in ordinary population samples.
• The mechanism is reduced protein stability, not loss of catalytic machinery.
• The original association literature was with problem drug use; the anxiety and threat-habituation phenotype came later and is the better-replicated part.

Sources: Sipe JC 2002 · Chiang KP 2004 · Dincheva I 2015

The animal pharmacology, which was excellent animal

Kathuria and colleagues reported in Nature Medicine that the carbamate FAAH inhibitor URB597 raised brain anandamide and produced anxiolytic-like effects in rats in the elevated zero-maze and the isolation-induced vocalisation test, without the motor and hypothermic effects of a direct agonist and without producing the classical cannabinoid tetrad. That result was the founding promise of the field: the tissue-selective, ceiling-limited anxiolysis and analgesia that direct agonism cannot give you. Preclinical FAAH work of this kind is extensive and largely positive, which is exactly why the human record that follows is worth stating plainly.

Sources: Kathuria S 2003

The human record: a real, important negative result human data

FAAH inhibition has repeatedly failed to reproduce its animal analgesia in people, and this shelf says so rather than skipping it. The most informative trial is Huggins and colleagues 2012 with PF-04457845, an irreversible FAAH-1 inhibitor, in osteoarthritis of the knee. The trial was well designed for exactly this question: it confirmed target engagement, showing that the compound raised circulating fatty acid amides as intended, and it still produced no effective analgesia relative to placebo, while naproxen as an active comparator did. The mechanism worked; the medicine did not. Similar disappointment followed FAAH inhibitors in other pain indications. The interpretations on offer are all instructive — that raised anandamide tone is not sufficient for analgesia in established human pain, that the animal models predicted the biomarker rather than the endpoint, or that FAAH-2, present in humans and absent in rodents, changes the target biology between species. What is not available is a reading in which FAAH inhibition is a demonstrated human analgesic. Anyone building a "raise your own anandamide" narrative has to sit with that.

• Target engagement confirmed; endpoint not met. Those are separable, and only the second is the medicine.
• The active comparator worked in the same trial, so the failure is not an insensitive study.
• Humans have FAAH-2; rodents do not. Species disanalogy is a live explanation, not an excuse.

Sources: Huggins JP 2012

BIA 10-2474: why "it is only an enzyme inhibitor" is not a safety argument human data

In January 2016 a phase 1 trial in Rennes, France, of the FAAH inhibitor BIA 10-2474, developed by the Portuguese company Bial, produced an acute and severe neurological syndrome in volunteers in the highest repeated-dose cohort. Kerbrat and colleagues reported the clinical picture in the New England Journal of Medicine: headache progressing to a rapidly evolving encephalopathy with cerebellar and brainstem involvement and symmetrical pontine and hippocampal lesions on imaging. One volunteer died; others were left with residual neurological deficits. Van Esbroeck and colleagues established the most likely reason in 2017 using activity-based protein profiling: BIA 10-2474 is not a selective FAAH inhibitor. At the concentrations reached it covalently inhibited several additional lipases, including enzymes of brain lipid metabolism that other FAAH inhibitors leave alone, and disrupted lipid networks in human cells. The regulatory reading afterwards was that this was a compound-specific off-target toxicity rather than a class effect, and selective FAAH inhibitors including PF-04457845 have not shown the same syndrome. Both halves of that conclusion have to be held at once. The class is not condemned, and the episode is nonetheless the standing answer to the idea that inhibiting an endogenous enzyme is inherently gentle. An enzyme inhibitor is a reactive molecule meeting a proteome, selectivity is an empirical property that has to be measured rather than assumed, and the price of assuming it here was a death in a healthy-volunteer study.

• The injury pattern was central and structural: brainstem, cerebellum, pons, hippocampus.
• Mechanism attributed to off-target covalent inhibition of other lipases, not to FAAH inhibition as such.
• Selectivity is measured, not inferred from a target label. "Just an enzyme inhibitor" is not a risk assessment.

Sources: Kerbrat A 2016 · van Esbroeck ACM 2017

FAAH is already in your medicine cabinet animal

One under-discussed fact keeps the enzyme in perspective. Paracetamol (acetaminophen) is deacetylated to p-aminophenol, which crosses into the central nervous system and is there conjugated with arachidonic acid to form N-arachidonoylphenolamine, AM404. Högestätt and colleagues showed that this conjugation is FAAH-dependent and does not occur in FAAH-null mice. AM404 is a TRPV1 agonist, a weak cannabinoid receptor ligand and an inhibitor of endocannabinoid cellular uptake, and is considered responsible for at least part of paracetamol analgesia. So the most widely consumed analgesic in the world depends on this enzyme to generate an active metabolite that acts on this system. The corpus already holds a longer treatment of the AM404 pathway; it is recorded here because it is the cleanest demonstration that the endocannabinoid system is not an exotic target but an ordinary one that mainstream pharmacology has been using without saying so.

Sources: Högestätt ED 2005

Natural FAAH inhibition, in proportion contested in vitro

Several dietary compounds inhibit FAAH in vitro: kaempferol competitively with a Ki in the low micromolar range, biochanin A with activity reported at submicromolar concentrations, and the macamides of maca, which are structurally fatty-acid amides themselves and inhibit both FAAH and anandamide cellular uptake. Those findings are real and are tabulated on the natural-inhibitors page with their assay levels. They do not license the inference that eating the plant inhibits the enzyme in a person, and the two clinical facts above are the reason to be especially careful with that inference here: even a purpose-built, potent, selective, target-engaging FAAH inhibitor failed to produce the intended human effect.

Contested — caveat. All of the natural-product FAAH potencies are isolated-enzyme or cell-assay values. No human pharmacokinetic study establishes that dietary intake of any of these compounds achieves the necessary concentration at the enzyme.

Sources: Thors L 2008 · Thors L 2010 · Wu H 2013 · Huggins JP 2012