Endocannabinoid Metabolism Pathways

The body doesn't wait for a plant to make cannabinoids

Somewhere around 600 million years of evolution, vertebrates developed the machinery to manufacture their own lipid signalling molecules , molecules that bear a striking structural resemblance to the compounds in Cannabis sativa. That's not a coincidence, but it is regularly misunderstood. The plant didn't give us the system; we had the system long before the plant mattered to anyone. Understanding how the endocannabinoid system actually metabolises its key messengers is, in my view, one of the most underexplored corners of cannabinoid science , at least in the popular press, which tends to focus on receptors and skip over the enzymatic machinery that controls everything upstream and downstream.

So. Let's look at the metabolism: how endocannabinoids are synthesised on demand, how they signal, and how they're inactivated. The two primary endocannabinoids are anandamide (N-arachidonoylethanolamine, or AEA) and 2-arachidonoylglycerol (2-AG). They're chemically distinct, they're synthesised by different enzymes, and they're broken down by different enzymes. Treating them as interchangeable is a mistake I made early on , one I've since corrected.

Anandamide: synthesis and the NAPE-PLD pathway

Anandamide is an N-acylethanolamine (NAE), part of the broader N-acylethanolamines family. Its synthesis begins in the phospholipid bilayer of the cell membrane. The principal route runs through a two-step process: first, a calcium-dependent transacylase transfers arachidonic acid from the sn-1 position of phosphatidylcholine to phosphatidylethanolamine, producing N-arachidonoyl phosphatidylethanolamine (NAPE). Then, a phospholipase D enzyme; specifically NAPE-PLD, cleaves that precursor to release anandamide directly.

That's the canonical route. But, and this is where it gets genuinely interesting; multiple alternative pathways exist. NAPE can be processed by ABHD4 into lyso-NAPE and then by glycerophosphodiesterase 1 (GDE1) into anandamide. There's also a route through phospholipase C followed by a phosphatase. The existence of redundant synthesis pathways suggests the system places high value on being able to produce anandamide under varying cellular conditions; NAPE-PLD knockout mice still generate anandamide, which confirmed the alternatives aren't trivial.

Once synthesised, anandamide acts as a partial agonist at CB1 receptors and, to a lesser extent, CB2 receptors. It also activates TRPV1 receptors, which is a detail that matters for understanding its broader signalling profile beyond the classical cannabinoid receptor axis.

2-AG: a different lipid, a different enzyme set

2-AG is structurally a monoacylglycerol rather than an ethanolamine. It's present in brain tissue at concentrations roughly 170-fold higher than anandamide, a fact that surprises most people when they first encounter it. The dominant synthesis route for 2-AG starts with phosphatidylinositol (PI), which is hydrolysed by phospholipase C-beta (PLCβ) to produce diacylglycerol (DAG). Diacylglycerol lipases alpha and beta (DAGLα and DAGLβ) then convert DAG to 2-AG.

DAGLα is the primary isoform in neurons; DAGLβ dominates in microglia and macrophages. That tissue-specific distribution is relevant when thinking about where 2-AG is produced and which downstream receptor populations it's most likely to reach. 2-AG is a full agonist at both CB1 and CB2 receptors, a key distinction from anandamide's partial agonism at CB1.

Both endocannabinoids are synthesised "on demand," meaning they're produced postsynaptically in response to cellular depolarisation or receptor activation, rather than stored in vesicles and released like classical neurotransmitters. This retrograde signalling logic; where the postsynaptic neuron sends a lipid messenger back across the synapse to modulate the presynaptic terminal, is a defining feature of the endocannabinoid system and worth keeping clearly in mind.

Degradation: FAAH, MAGL, and why enzyme inhibition matters to researchers

Synthesis is only half the story. What limits how long and how intensely an endocannabinoid signals is its degradation, and both anandamide and 2-AG have dedicated enzymatic machinery for exactly that.

Anandamide is primarily inactivated by fatty acid amide hydrolase, universally shortened to FAAH. FAAH is a serine hydrolase located on the intracellular face of the endoplasmic reticulum membrane. It hydrolyses anandamide into arachidonic acid and ethanolamine. FAAH also processes other N-acylethanolamines, including palmitoylethanolamide (PEA) and oleoylethanolamide (OEA), which is relevant context when considering selectivity in research models.

Anandamide can also be oxidised by cyclooxygenase-2 (COX-2) to produce prostaglandin-ethanolamides (prostamides), and by lipoxygenases and cytochrome P450 enzymes, though these routes are considered secondary under most physiological conditions. The COX-2 pathway is one I find particularly interesting because it places endocannabinoid metabolism directly within the arachidonic acid cascade; a system most people associate with classical inflammation biology, not cannabinoids.

2-AG has a different primary degradation enzyme: monoacylglycerol lipase (MAGL). MAGL hydrolyses 2-AG to arachidonic acid and glycerol. Because MAGL-derived arachidonic acid is a substantial source of prostaglandin precursors in the brain, MAGL inhibition has attracted significant attention in preclinical research, the downstream implications extend well beyond the endocannabinoid system itself. ABHD6 and ABHD12 also hydrolyse 2-AG, though at substantially lower rates than MAGL under normal conditions.

What happens when plant cannabinoids enter this picture

Plant-derived cannabinoids don't sit entirely outside this metabolic machinery, they interact with it in ways that complicate simple receptor-binding narratives. Cannabidiol (CBD), for instance, is a known inhibitor of FAAH in vitro; whether this translates to physiologically meaningful FAAH inhibition at concentrations achievable in humans after oral dosing is a genuinely open question, and I'd argue some popular writing on the topic has run ahead of the evidence.

THC is a partial agonist at CB1 and CB2 receptors; the same receptors anandamide and 2-AG act on, but its pharmacokinetics differ substantially from any endocannabinoid. It's lipophilic, it accumulates in adipose tissue, and it undergoes extensive hepatic metabolism including first-pass metabolism when taken orally, producing 11-hydroxy-THC and then 11-nor-9-carboxy-THC (THC-COOH) as major metabolites. The endocannabinoids, by contrast, are locally synthesised and locally degraded; they don't circulate systemically in the same sense.

I spent an afternoon last winter going back through Raphael Mechoulam's 1992 paper describing the isolation of anandamide, the one that coined the name from the Sanskrit ānanda, meaning bliss; and was struck again by how carefully the authors framed their receptor binding data. Very much a "here is a finding; here are the limitations" approach. That scientific caution has sometimes been lost in subsequent decades of popular writing about the ECS.

The endocannabinoidome: a wider metabolic context

Anandamide and 2-AG are the most studied endocannabinoids, but they sit within a much larger network of lipid mediators. This broader system has been termed the endocannabinoidome, a term that encompasses dozens of structurally related N-acylethanolamines, monoacylglycerols, and their biosynthetic enzymes and receptor targets.

PEA (palmitoylethanolamide) and OEA (oleoylethanolamide), for example, are produced by the same NAPE-PLD pathway as anandamide and are both FAAH substrates, yet they show minimal affinity for CB1 and CB2 receptors. They act primarily at peroxisome proliferator-activated receptor-alpha (PPARα) and GPR119. Including them in a discussion of endocannabinoid metabolism is not a digression, it's necessary context for understanding why enzyme inhibition studies produce effects that extend beyond what receptor binding assays would predict.

The hypothesis that disruptions to endocannabinoid tone; through altered synthesis, altered degradation, or altered receptor density, might underlie certain physiological states is sometimes discussed under the framework of clinical endocannabinoid deficiency. That remains a hypothesis, and a contested one. The mechanistic groundwork for it runs directly through the metabolic pathways described above, which is why getting this enzymatic picture right matters before evaluating the broader theoretical claims.

A note on research limitations

Almost everything detailed above was established first in animal models, predominantly rodent. Translating enzymatic kinetics from murine tissue to human biology is not automatic. Human FAAH, for instance, has a documented polymorphism (C385A) that reduces enzyme activity; carriers of the minor allele show elevated anandamide levels in some assays. That's a real finding, but the downstream significance of that variation in living humans remains an active area of inquiry, not a settled question.

If you're reading primary literature in this space, pay close attention to whether studies measured endocannabinoid levels in plasma, cerebrospinal fluid, or tissue homogenate, the numbers are not comparable across matrices, and a lot of confusion in the secondary literature traces back to exactly that conflation.

Sources

• Endocannabinoid metabolism and the enzymatic degradation of anandamide and 2-AG; NCBI / PMC
• Physiology, Endocannabinoids, StatPearls / NCBI Bookshelf
• Cannabinoid scheduling and regulatory framework, Therapeutic Goods Administration (TGA)
• The endocannabinoidome: the world of endocannabinoids and related mediators; NCBI / PMC

, Elise Tran, Science writer, cannabinoids & the ECS