The Entourage Effect, Honestly

What the term originally meant in vitro

The phrase entourage effect was introduced by Ben-Shabat, Mechoulam and colleagues in 1998, and the claim was narrow and specific: endogenous 2-acyl-glycerols that are themselves inactive at cannabinoid receptors enhanced the activity of 2-AG. It was a statement about endogenous lipid mixtures, not about plant terpenes, and it was made in a pharmacological assay. Everything commonly labelled entourage effect today is an extension of that original observation to a different set of molecules, a different concentration regime and, usually, a different kind of evidence. Keeping the original meaning in view is the first step to reading the literature honestly, because the strength of the 1998 finding does not transfer to the claims made in its name.

Sources: Ben-Shabat S 1998

The strong form contested in vitro

The strong form of the modern claim is that the terpene fraction of cannabis flower meaningfully shapes the human effect of the cannabinoids present — that myrcene sedates and shifts blood-brain-barrier permeability, that limonene lifts mood, that pinene offsets memory impairment, and that the cultivar-to-cultivar differences users report are terpene-driven. Russo laid out the case in 2011 in the most-cited paper on the subject, assembling mechanistic and preclinical evidence for phytocannabinoid-terpenoid synergy and proposing specific pairings. The paper is a genuine synthesis and it is a hypothesis-generating review, and the pharmacology it assembles is drawn largely from studies using terpene concentrations far above those a person inhales. This shelf marks the strong form contested, not refuted — and the distinction between those two words is the content of the next two sections.

Contested — caveat. The strong form is a hypothesis supported largely by mechanistic and preclinical work at supraphysiological terpene concentrations. Two subsequent receptor-level studies failed to find the predicted modulation. It is contested, and the burden of proof sits with the claim.

Sources: Russo EB 2011

The negative evidence, which is good in vitro

Two independent groups tested the receptor-level prediction directly and did not find it. Santiago, Sachdev, Arnold, McGregor and Connor tested the terpenoids commonly found in Cannabis sativa — including α-pinene, β-pinene, β-caryophyllene, linalool, limonene and myrcene — for effects on Δ9-THC signalling at human CB1 and CB2 in cellular assays, and titled the paper Absence of Entourage: the terpenoids did not activate the receptors and did not modulate the functional activity of Δ9-THC at them. Finlay, Sircombe, Nimick, Jones and Glass reached the same conclusion independently the following year in Frontiers in Pharmacology, again finding no terpenoid-mediated entourage effect at cannabinoid receptors. Two groups, different assay systems, same result, both explicitly designed to test the claim. That is a substantial negative result and any treatment of the entourage effect that omits it is not reporting the literature.

• Terpenoids tested did not activate CB1 or CB2 themselves.
• Terpenoids tested did not shift the potency or efficacy of Δ9-THC at either receptor.
• Replicated independently in a second laboratory with a different assay platform.
• What is excluded is specifically receptor-level modulation. Other routes to a real interaction remain open, and are covered below.

Sources: Santiago M 2019 · Finlay DB 2020

The concentration argument, explicitly in vitro

This is the quantitative core and it deserves stating in plain numbers rather than gestured at. Terpene effects in the supporting literature are typically demonstrated at concentrations in the tens to hundreds of micromolar, and sometimes millimolar, in a dish or at a dose in a rodent that corresponds to a substantial fraction of a gram per kilogram. The terpene content of cannabis inflorescence is on the order of one to a few percent by dry weight for the whole terpene fraction, spread across dozens of compounds, of which a substantial part is lost to pyrolysis and to the sidestream during smoking or does not volatilise in a vaporiser at the temperature used. What arrives in a person from a typical inhaled dose is, for any individual terpene, in the microgram-to-low-milligram range, distributing into a large lipophilic volume. The resulting plasma and brain concentrations are orders of magnitude below the in-vitro concentrations at which the effects are demonstrated. That gap — not laboratory technique, not commercial bias — is the central reason to doubt the strong form. It is the same argument the natural-inhibitors page makes about micromolar enzyme IC50 values and dietary polyphenols, and it applies here with the same force.

Sources: Santiago M 2019 · Finlay DB 2020 · Russo EB 2011

What IS reasonably supported human data

Rejecting the strong form does not leave nothing. Four things stand up. First, cannabinoid-cannabinoid interaction: cannabidiol does modulate the effects of Δ9-THC, and Boggs and colleagues reviewed the clinical and preclinical evidence for those functional interactions — the effect is dose-dependent and ratio-dependent rather than uniformly protective, which is itself an important qualification, but the interaction is real and is measured in humans. Second, pharmacokinetic entourage: cannabidiol and other constituents affect the metabolism and disposition of Δ9-THC, which changes exposure without any receptor-level synergy being required. A pharmacokinetic interaction is a perfectly respectable entourage mechanism and it is the one with the most human data. Third, aroma and expectancy: subjective cannabis effect is demonstrably shaped by smell, set and expectation, and terpenes are what a cultivar smells like. That is a psychological mechanism rather than a pharmacodynamic one, and it is not a dismissal — it is a real determinant of reported effect that happens to be mediated by the nose rather than by CB1. Fourth, receptor-independent effects at high local concentration: terpenes do have pharmacology, and at the concentrations actually achieved on skin in a topical, or at the airway epithelium during inhalation, local concentration can be high even when systemic concentration is negligible. Airway irritation, antimicrobial activity and local anti-inflammatory effects belong in that category.

• Cannabidiol with Δ9-THC: documented functional interaction in humans, dose- and ratio-dependent.
• Pharmacokinetic interaction between constituents: the best-evidenced entourage mechanism.
• Aroma-mediated expectancy: real, measurable, psychological, not pharmacodynamic.
• High local concentration at skin or airway: receptor-independent terpene pharmacology where the concentration argument does not apply.

Sources: Boggs DL 2018 · Santiago M 2019 · Russo EB 2011

β-caryophyllene is the exception that proves the rule animal

One terpene does engage a cannabinoid receptor, and noticing which one is instructive. β-caryophyllene is a selective CB2 agonist — Gertsch and colleagues demonstrated binding, functional activation, an anti-inflammatory effect in mice, and loss of that effect in CB2-knockout animals. It appears in the Santiago negative result too, where it did not modulate Δ9-THC at CB1 or CB2, which is entirely consistent: it is an agonist in its own right at CB2, not a modulator of another ligand at either receptor. So the one clear terpene-to-cannabinoid-receptor link in the literature is a direct agonism, orally bioavailable, at a non-psychoactive receptor, and it is not an entourage effect in the sense the term is usually used. That is the shape of a real finding in this area: specific, mechanistically defined, receptor-attributed by knockout, and narrower than the slogan.

Sources: Gertsch J 2008 · Santiago M 2019