In the TV adaptation of John le Carré’s cold war novel, Smiley’s People, a foreign agent is interrogated by intelligence officers about his first meeting with Karla, his austere boss and manipulator: “On the hill, is a small dacha, a few lights, no luxuries. Who is this strange baron, I ask myself, who can live in a hut and look at the whole world without wishing to possess it?”
But that 1980s TV series was way back when, before our time. Nowadays, asceticism can be bottled in the form of molecular lookalikes of hormones that our body already creates. One such lookalike is the mimic to the hormone Glucagon-Like Peptide-1 (GLP-1), commercially known as the weight loss drug Ozempic.
The mechanism of action for these hormones-re-engineered-as-drugs is under active research. Extrapolating from the little we know is tempting: there is no shortage of polite fictions on how our appetites are muzzled. ‘The gut natively speaks to the brain through a satiety hormone,’ one story goes. ‘In people who carry excess weight that signal has become noisy, and the drug emphatically repeats that same message more clearly.’ It is a satisfying account, and it has the virtue of being easy to picture. The imagery is compelling because it suggests that some long-lost set point from our youth can be reclaimed.
A good starting point for our mechanistic trail is the mucosal lining of the intestine, which houses a scattered population of ‘L cells’. In a sense, these cells act as taste buds: food arrives in the gut and L cells sense it directly, sampling the sugars flowing past their apical tips. These cells are electrically excitable: feed them glucose and they depolarize, fire action potential signals, releasing native GLP-1 that spreads to nearby tissues and spills into the circulation. Then, this gut-produced hormone triggers the pancreas to release insulin and, critically for our story, tells the brain to curb appetite.
It’s a tidy story. But the reality that research is uncovering suggests otherwise: the gut’s native version of GLP-1 lasts about two minutes and is thought to communicate with the brain primarily via one kind of neuron about that one meal that triggered its release. Conversely, the drug version of GLP-1 is engineered to linger in the body for about a week, and it triggers neural reactions in brain regions that the native hormone can barely perturb.
But let’s continue along the gut: what kind of neuron reads GLP-1, which is to say: what neuron has molecular receptors sensing that hormone? The obvious candidate would be one that runs along the vagus nerve, whose sensory fibers thread through the wall of the intestine and report upward to the brainstem and onwards to centres that control appetite: a reasonable guess, held for years. Presumably, these vagal fibers would dip their endings into the gut’s mucosal lining, where the nutrients are, and detect the hormones released by the L cells. Except, when researchers went looking for the vagal sensors that actually carry the GLP-1 receptors, they found them in a different neighborhood. These neurons don’t fan out into the gut lining where nutrients are absorbed and GLP-1 is released. Instead, GLP-1-sensing neurons sit in the muscle layers of the gut. There, they sense stretch in the muscle that wraps the stomach and intestines. These GLP-1-sensitive neurons are not nutrient detectors but mechanical ones.
Why would the neuronal endings of a stretch-gauge evolve to take into account GLP-1 levels? Answer: GLP-1 adjusts the ‘gain’ of such gauges so that the sensation of fullness relayed to the brain reflects both the quantity (i.e., the expansion of the gut wall) and quality (i.e., the nutrient density) of the food being digested. And this ‘gain’ genuinely matters for behaviour: experimentally cutting the vagus or blocking the GLP-1 receptor on this neuronal pathway in rodents annuls the appetite-suppressing effect of native GLP-1.
But the above experiments revealed much more: disable the native GLP-1 gut pathway and the rhythm of eating is also deranged. The normal daily pattern of when hunger rises and falls gets scrambled in rodents, because physiological GLP-1 secretion helps keep feeding on a schedule. So native GLP-1 is a consequential circuit: it sets the size of the meal consumed, coordinates the insulin release that a meal demands, and keeps feeding to a regular beat. However, experimental interference of the gut pathway did not significantly impact long-term body weight because these GLP-1 secretions are so short-lasting. Native GLP-1 signalling is a within-meal controller, not a weight thermostat. The evolution of GLP-1 signalling in the gut did not include ‘weight regulation’ on its job description.
Which raises the question: if the body’s own gut-brain GLP-1 circuit only regulates the meal at hand, how are the drug-versions of GLP-1 engineered to strip away weight and keep it off, compared to the native version?
The answer is that while the native GLP-1 hormone has a shut-down switch, the drug form does not. GLP-1 produced by L cell is chewed up by a neighbourhood enzyme, DPP-4, within a couple of minutes of release, while the drug version of GLP-1 has this fragility removed. The drug is chemically modified to last longer by resisting DPP-4, lingering in the body for days rather than minutes. Furthermore, the medication is administered at doses far above anything the intestine would ever produce. Because of this higher sustained dosage, the drug stimulates the permeable regions of the blood-brain barrier more effectively.
What are these permeable regions? Most blood vessels supplying the brain are walled off behind a selective barrier, but there are a few deliberate ‘windows’: small structures whose blood vessels are left porous precisely so that the brain can sample what is circulating. Two of them matter. The area postrema sits in the brainstem monitoring the blood for signs of poisoning, and can trigger nausea and vomiting when it finds them. The median eminence sits at the base of the hypothalamus, next door to the arcuate nucleus, that hub of appetite control. Thanks to its higher blood levels, the drug version of GLP-1 exerts a long-lasting stimulation of these brain areas whereas the native hormone only appears in circulation for perhaps a minute or two after a meal.
What happens next sets apart the drug from the native hormone: the neurons at those brain windows respond to sustained GLP-1 drug in the blood and project their signal across a distributed network of appetite, satiety and reward circuits that the molecule itself never physically reaches. The gut-brain axis, in other words, is not the theatre of operations for weight loss: the main action is a small cluster of neurons in permeable brainstem regions being held, week after week, in a state that the body’s own hormone could sustain only fleetingly. Engineered GLP-1 may not, in the end, need the gut’s vagal relay for appetite suppression.
Held for long enough, drug effects begin to spill past appetite, affecting wanting in general, not merely hunger. In a small trial in people with alcohol use disorder, these drugs reduced craving and drinking; among those who happened to smoke, cigarettes per day fell too. The evidence is early but the direction is clear: engineered GLP-1 acts on the machinery of craving in the broadest sense, of which food is only the most visible instance.
Which returns us to that dacha on the hill. Karla, ‘who could look at the whole world without wishing to possess it’, was a chain-smoker. Today a physician might offer him a weekly dose of a GLP-1 drug. But perhaps something would be lost if he took it. Le Carré knew what he was doing when he gave this ascetic a vice: a man who wants nothing at all is not human. Appetite is not simply a fault to be corrected.
===