The canonical clinical-stage FAAH inhibitor, strong probe value but modest clinical payoff so far
Look, if you’re working anywhere near the endocannabinoid system, PF-04457845 is the compound you reach for. Not because it’s the flashiest thing in the library, but because it’s the cleanest, most human-validated FAAH inhibitor we have. Pfizer’s covalent piperidine-urea scaffold gives you irreversible, highly selective FAAH inhibition with a ~14-hour plasma half-life, verified target engagement in humans, and a pharmacokinetic profile that actually translates from preclinical models. Thats a short list of compounds that can claim all three.
The mechanism is worth thinking through carefully. FAAH inhibition doesn’t force receptor agonism the way a direct CB1 agonist does. Instead, it elevates the endogenous pool of N-acylethanolamides: anandamide hitting CB1/CB2 and TRPV1, OEA and PEA acting primarily through PPARs and GPR119. You get a diffuse, physiological-range modulation of multiple pathways from a single enzyme knockout. On paper, that’s elegant. The CNS penetration is real, the peripheral bioavailability is solid, and the selectivity over other serine hydrolases is among the best documented in the class.
Here’s the thing though: the clinical results have been disappointing in the indications where FAAH inhibition seemed most obvious. The 2012 Huggins et al. phase 2 in osteoarthritis pain showed clear target engagement and almost no signal on the primary endpoint. Fibromyalgia data fared similarly. What we don’t know is whether that reflects the mechanism, the compound, the indication selection, or something about how endocannabinoid tone works in chronic pain states versus acute models. That ambiguity is scientifically interesting even if its commercially discouraging.
The shadow of the 2016 Bial tragedy is worth addressing directly. BIA-10-2474, the FAAH inhibitor that killed one volunteer and injured five others in Rennes, was a different compound with different off-target liabilities. Post-incident analysis placed the blame on BIA-10-2474’s promiscuous serine hydrolase inhibition profile, not the FAAH inhibition mechanism itself. PF-04457845’s selectivity data held up under scrutiny. The tragedy did, however, cast a long institutional shadow over the class and almost certainly slowed clinical development timelines industry-wide.
Where we think PF-04457845 earns its place in a research library is as a translational probe. The recent cannabis use disorder work is genuinely more interesting than the pain programs, and the rationale for FAAH inhibition in addiction biology is mechanistically distinct enough to be worth watching. For investigators studying the ECS, comparing FAAH inhibitor selectivity profiles, or benchmarking against other endocannabinoid-elevating strategies, this is the reference compound. Its not a frontier molecule, and the bench scientist who wants the next big thing will find more novelty elsewhere. But if you’re asking whether FAAH inhibition does what you think it does in a human being, PF-04457845 is how you find out.