Appetite circuits: hypothalamus, brainstem and reward

The weight loss from these drugs is almost entirely a story about eating less. So the real mechanism question is: where in the brain, and how?

Three regions, three jobs

The arcuate nucleus (hypothalamus). This sits at the base of the brain next to the median eminence, a region without a tight blood-brain barrier. Secher and colleagues showed that labelled liraglutide reaches the arcuate nucleus in mice, activates POMC/CART neurons (which suppress eating) and indirectly inhibits NPY/AgRP neurons (which drive it), and that weight loss depended on this region (PubMed 25202980). Sisley and colleagues showed that deleting the GLP-1 receptor from neurons abolished liraglutide's effect on food intake but not its effect on glucose, separating the two actions cleanly (PubMed 24762441).

The brainstem: area postrema and NTS. The area postrema is another barrier-free zone. Neighbouring it, the nucleus of the solitary tract (NTS) receives vagal signals from the gut. Fortin and colleagues found GLP-1 receptors on GABA neurons in the rat NTS that mediate reduced intake (PubMed 32132220). Huang and colleagues then showed in mice that these are dissociable circuits: NTS GLP-1 receptor neurons produce satiety without aversion, while area postrema neurons produce aversion and nausea-like behaviour (PubMed 38987598). That result is the clearest anatomical basis yet for the idea that the fullness and the sickness are different signals, covered on the nausea page.

Reward regions. In people, GLP-1 receptor activation changes how the brain responds to food. van Bloemendaal and colleagues gave exenatide to adults with obesity or type 2 diabetes during fMRI and found reduced responses to food pictures in the insula, amygdala, putamen and orbitofrontal cortex, alongside lower food intake (PubMed 25071023). Whether the drug reaches these regions directly or they respond to upstream hypothalamic and brainstem input is not known; Gabery's tracing study found semaglutide largely confined to barrier-free regions, with downstream activation spreading from there (PubMed 32213703).

What the human data say

Animal tracing tells you where the drug goes. Human studies tell you what changes. Blundell and colleagues ran a crossover study in 30 adults with obesity: after 12 weeks of semaglutide, ad libitum energy intake at a test lunch was about 24% lower than on placebo, hunger and prospective food consumption ratings fell, fullness and satiety rose, cravings decreased and preference for high-fat foods dropped (PubMed 28266779). The intake reduction is the mechanism of weight loss in a sentence.

Reduced resting energy expenditure was not the mechanism; it fell in line with body weight, as it does with any weight loss. The why weight returns page picks that up.

What is debated

Which region dominates in humans is unknown. Kanoski, Hayes and Skibicka's review makes the case that GLP-1 receptor populations across the hindbrain, hypothalamus and mesolimbic system each contribute and that lesioning any single one only partly blunts the effect (PubMed 27030669). Gabery's phrase, distributed neural pathways, is the honest summary. Claims that the drugs "work on the hypothalamus" or "reset the set point" compress a distributed, partly-mapped system into a slogan. Whether GLP-1 receptor agonists also reduce alcohol or other reward-driven consumption in people is under active study and is not established.

Where to go next

• The intake effect over 68 weeks translates into the STEP 1 weight curve, digested at FormBlends Research.
• Practical eating strategies during treatment are at Weight Loss on GLP-1s.
• Click the brain nodes in the mechanism map for the graded evidence.