Appetite and Its Control

Appetite control is best understood as a negative-feedback homeostatic system, structurally analogous to thermoregulation but built around the variable of energy stores rather than temperature. The controlled variable is, on a long timescale, adipose tissue mass; on a short timescale, it is the moment-to-moment nutrient state of the gut and blood. The hypothalamic arcuate nucleus (ARC), sitting adjacent to the median eminence where the blood-brain barrier is relatively permeable, functions as the primary sensor and integrator. It contains two functionally antagonistic neuron populations occupying the same anatomical space but driving opposite behaviors. The first population co-expresses neuropeptide Y (NPY) and agouti-related peptide (AgRP). These neurons are orexigenic — their activation drives hunger, food-seeking, and reduced energy expenditure. The second population co-expresses pro-opiomelanocortin (POMC), the precursor of α-melanocyte-stimulating hormone (α-MSH), and cocaine- and amphetamine-regulated transcript (CART). These neurons are anorexigenic — their activation suppresses appetite and increases energy expenditure. Both populations project to a shared downstream target, the paraventricular nucleus (PVN) of the hypothalamus and the lateral hypothalamic area (LHA), where their outputs are summed and relayed onward to brainstem feeding-control centers. Critically, both neuron populations express receptors for circulating metabolic hormones, making the ARC a site where humoral signals are converted into neural output. Leptin and insulin are "tonic" or long-term signals whose plasma concentrations track adipose tissue mass over days to weeks; they suppress NPY/AgRP neurons and stimulate POMC/CART neurons, so more fat mass produces less hunger drive, on average. Ghrelin, cholecystokinin (CCK), peptide YY (PYY), and glucagon-like peptide-1 (GLP-1) are "phasic" meal-related signals that rise and fall across the timescale of a single meal, giving the system the fine temporal resolution needed to start and stop individual episodes of eating. The overall behavioral output — hunger, satiation (the process of stopping a meal), and satiety (the post-meal suppression of hunger) — emerges from the algebraic sum of these inputs onto the ARC and its downstream targets.

Step 1 — Long-term adiposity signaling: leptin. Leptin is a 167-amino-acid protein hormone secreted by adipocytes in direct proportion to fat cell size and total triglyceride content: more fat, more leptin. Leptin circulates in plasma at concentrations roughly proportional to body fat percentage (typical fasting levels range from about 5–15 ng/mL in lean humans to well over 30 ng/mL in obese individuals). It crosses the blood-brain barrier via a saturable transport mechanism and binds the long-form leptin receptor (Ob-Rb), a class I cytokine receptor, on ARC neurons. Binding activates JAK2/STAT3 signaling: JAK2 phosphorylates STAT3, which dimerizes, translocates to the nucleus, and alters transcription. In POMC/CART neurons, leptin increases POMC expression and neuronal firing. In NPY/AgRP neurons, leptin inhibits firing and reduces NPY/AgRP transcription. The net behavioral effect of leptin is appetite suppression and increased energy expenditure, making it a signal that reports "fat stores are adequate" to the brain.

Step 2 — The melanocortin output pathway. POMC is cleaved by prohormone convertases into several peptides, the physiologically central one here being α-MSH. α-MSH is released onto second-order neurons in the PVN and LHA, where it binds melanocortin-4 receptor (MC4R), a Gs-coupled receptor that raises cAMP and suppresses feeding. AgRP, released by the opposing NPY/AgRP neurons, is an inverse agonist/competitive antagonist at MC4R — it physically blocks α-MSH's action, tipping the balance toward hunger. NPY itself acts on Y1 and Y5 receptors on PVN neurons to independently stimulate feeding. This arrangement — one ligand (α-MSH) and its endogenous antagonist (AgRP) competing at the same receptor (MC4R) — is the single most important convergence point in mammalian appetite control; mutations in MC4R are the most common monogenic cause of severe human obesity, accounting for roughly 2–5% of cases with severe early-onset obesity.

Step 3 — Insulin as a second adiposity signal. Insulin, secreted by pancreatic β-cells in proportion to plasma glucose and, chronically, in rough proportion to adiposity, also crosses into the CNS and acts on insulin receptors in the ARC. Like leptin, insulin activates POMC/CART neurons and inhibits NPY/AgRP neurons via PI3-kinase signaling, reinforcing the same net direction: adequate stores suppress hunger, low stores (or low CNS insulin/leptin action, as in Regulation of Gastrointestinal Function-adjacent starvation states) permit it.

Step 4 — Short-term signals from the gut: ghrelin. Ghrelin is a 28-amino-acid peptide secreted mainly by X/A-like (ghrelin) cells in the gastric fundus. Unusually among hormones discussed here, ghrelin stimulates appetite: plasma levels rise before an anticipated meal (often peaking just before habitual mealtimes, reflecting learned timing as well as fasting state) and fall sharply within an hour of eating. Ghrelin crosses into the ARC and directly activates NPY/AgRP neurons via the growth hormone secretagogue receptor (GHSR), while indirectly inhibiting POMC/CART neurons through local GABAergic connections. Ghrelin is essentially the physiological mirror image of leptin — it is the principal "start eating" signal, whereas leptin, insulin, PYY, CCK, and GLP-1 are "stop/reduce eating" signals.

Step 5 — Meal termination signals: CCK, PYY, GLP-1. As food enters the duodenum, CCK is released from I cells in response to fats and proteins; it acts locally on vagal afferent CCK-A receptors, slows gastric emptying, and — via the vagus nerve to the nucleus of the solitary tract (NTS) in the brainstem — produces the sensation of fullness and terminates the meal. CCK's action is short-lived (minutes) and mainly governs meal size rather than the interval to the next meal. PYY (specifically the PYY₃₋₃₆ fragment) and GLP-1 are secreted by L cells in the distal small intestine and colon in response to the arrival of nutrients (particularly fat and carbohydrate), rising over 30–60 minutes and remaining elevated for hours. PYY₃₋₃₆ acts on Y2 receptors, which are inhibitory autoreceptors on NPY/AgRP neurons, thereby damping the hunger signal. GLP-1 acts both peripherally (slowing gastric emptying, potentiating glucose-stimulated insulin secretion — the "incretin effect") and centrally, engaging GLP-1 receptors in the ARC, PVN, and NTS to promote satiety.

Step 6 — Brainstem integration. All of the ascending vagal and circulating hormonal signals converge on the NTS in the caudal brainstem, which also receives direct mechanical distension signals from gastric stretch receptors and taste/chemosensory input. The NTS is reciprocally connected to the hypothalamic ARC/PVN circuitry, forming a hypothalamic-brainstem loop: the hypothalamus sets the overall "threshold" or drive for hunger over long timescales, while the NTS executes the moment-to-moment decision to terminate an individual meal once cumulative satiety signals cross that threshold. This two-tier architecture explains why an animal with normal circulating leptin can still eat a normal-sized meal (the NTS still works to stop it) but why chronically deficient leptin signaling (as in genetic leptin deficiency) causes both larger individual meals and higher meal frequency (the hypothalamic threshold is pathologically low).

The parabiosis experiments (Hervey, 1959; Coleman, 1969–1973). Long before any hormone was cloned, physiologists suspected a circulating "satiety factor." G. R. Hervey surgically joined pairs of rats in parabiosis (shared blood circulation) after making one rat obese via hypothalamic lesion. The lesioned rat overate and grew fat, while its surgically joined partner — receiving whatever the fat rat's blood contained — drastically reduced its food intake and lost weight, implying the obese rat's blood carried a satiety signal that its own (lesioned) brain could no longer read, but its partner's intact brain could. Douglas Coleman extended this logic decisively using two mouse mutants discovered at the Jackson Laboratory: ob/ob mice, obese due to a mutation later shown to be in the gene encoding leptin itself, and db/db mice, obese due to a mutation in the leptin receptor. Coleman parabiosed ob/ob mice with wild-type mice and found the ob/ob mice lost weight — they could respond to a circulating factor they weren't making. He parabiosed db/db mice with wild-type mice and found the wild-type partner stopped eating and lost weight, while the db/db mouse was unaffected — because db/db mice overproduce the missing factor (compensating for their broken receptor) and flood their parabiotic partner with it, while their own defective receptor makes them insensitive. Coleman correctly inferred, purely from these crosses, that ob/ob mice lack a satiety hormone and db/db mice lack its receptor — years before either molecule was identified.

Cloning of leptin (Zhang et al., 1994). Jeffrey Friedman's laboratory at Rockefeller University positionally cloned the mouse ob gene using linkage mapping, identifying it as a novel gene expressed specifically in adipose tissue, encoding a secreted 167-amino-acid protein. Administering recombinant leptin to ob/ob mice rapidly and dramatically reduced food intake and body weight; leptin had no comparable effect in db/db mice, consistent with Coleman's decades-earlier receptor-deficiency hypothesis. This work — published in Nature, 1994 — converted appetite regulation from a physiological black box into a defined molecular hormone-receptor axis and won Friedman and Coleman (posthumously honored alongside him) the Lasker Award in 2010.

Ghrelin's discovery and human infusion studies (Kojima et al., 1999; Wren et al., 2001; Cummings et al., 2002). Ghrelin was isolated from rat stomach as the endogenous ligand for the previously "orphan" growth hormone secretagogue receptor. Subsequent human studies by David Cummings and colleagues measured plasma ghrelin every few minutes across a 24-hour period and found a clear pre-meal rise and post-meal fall that tracked subjective hunger ratings almost exactly, strongly implicating ghrelin as a physiological meal-initiation signal rather than a mere correlate. Intravenous ghrelin infusion in humans (Wren et al.) increased food intake at a subsequent buffet meal by roughly 28% compared with saline, directly demonstrating a causal, not just correlative, orexigenic effect.

MC4R knockout and human mutation studies (Huszar et al., 1997; Farooqi et al., 2000s). Mice with targeted deletion of the Mc4r gene develop marked obesity, hyperphagia, and increased linear growth, phenocopying much of the agouti-yellow obesity syndrome and confirming MC4R as a necessary node for melanocortin-mediated satiety. In humans, Sadaf Farooqi and Stephen O'Rahilly's group screened cohorts of severely obese children and identified heterozygous and homozygous loss-of-function MC4R mutations as the most common monogenic cause of severe early-onset human obesity, with a clear gene-dose effect on food intake measured in structured test-meal paradigms.

Example 1 — Interpreting a genetic obesity pedigree. A clinician sees a 4-year-old with severe obesity (BMI far above the 99th percentile), intense hyperphagia, tall stature for age, and red hair pigmentation. Blood leptin is markedly elevated (roughly 40 ng/mL versus an age-matched norm of ~5 ng/mL). Reasoning: elevated leptin in the presence of obesity generally rules out primary leptin deficiency (those patients have near-undetectable leptin, since they lack fat-derived leptin production... actually the opposite: leptin-deficient patients have low leptin because the mutant leptin protein is nonfunctional or unsecreted, and their obesity itself does not raise a hormone they cannot make correctly). Here, elevated leptin despite obesity implies leptin resistance, i.e., a defect downstream of leptin — consistent with a signaling-pathway lesion. The combination of red hair (a clue pointing to disrupted α-MSH action, since α-MSH also signals through MC1R in melanocytes) and hyperphagia with tall stature points toward a POMC deficiency (loss of the α-MSH precursor: red hair from unopposed non-signaling at MC1R, hyperphagia from absent α-MSH at MC4R, and comparatively preserved growth because ACTH deficiency requires cortisol replacement but GH axis is intact). This contrasts with isolated MC4R mutation, which causes hyperphagia and tall stature but normal pigmentation, since MC1R signaling in skin is unaffected. The differential is resolved by sequencing POMC versus MC4R and by measuring ACTH/cortisol (low in POMC deficiency, since POMC is also the ACTH precursor in the pituitary).

Example 2 — Quantifying a bariatric-surgery satiety effect. A 40-year-old undergoes Roux-en-Y gastric bypass. Preoperatively, fasting plasma ghrelin is 700 pg/mL and rises to 900 pg/mL before meals; postoperative fasting ghrelin falls to roughly 250 pg/mL and shows a blunted pre-meal rise. Meanwhile, postprandial GLP-1 and PYY, measured 30 minutes after a standard 400-kcal test meal, rise several-fold higher than preoperatively (a typical finding is GLP-1 increasing from a peak of ~15 pmol/L to ~45 pmol/L). Reasoning: bypass surgery removes or bypasses the ghrelin-rich gastric fundus and delivers nutrients rapidly to the distal small intestine, where L cells are far more numerous — producing an exaggerated, early "distal gut" nutrient signal. The physiological prediction, later confirmed clinically, is that patients experience reduced hunger and earlier, stronger satiation independent of caloric restriction per se, explaining why bypass produces more durable weight loss than purely restrictive dieting: the surgery permanently changes hormonal input to the ARC/NTS circuit rather than merely limiting mechanical capacity. This is also the physiological rationale that motivated GLP-1 receptor agonist drug development — mimicking the endogenous post-bypass GLP-1 rise pharmacologically.

Douglas Coleman spent most of his career at the Jackson Laboratory in Bar Harbor, Maine, working with inbred mouse strains rather than in a prestigious university department, and he never held a large, richly funded laboratory. Working largely without the tools of modern molecular biology, he relied on classical physiological surgery — parabiosis — and rigorous logical inference from cross-circulation experiments to deduce the existence of a satiety hormone and its receptor years before either could be sequenced. His 1969–1973 papers proposing that ob/ob mice lack a blood-borne satiety factor and db/db mice lack its receptor were, at the time, essentially unprovable predictions; they were vindicated only two decades later when Jeffrey Friedman's group cloned leptin and its receptor and showed the molecular defects matched Coleman's model almost exactly.

Obesity pharmacology. GLP-1 receptor agonists (exenatide, liraglutide, semaglutide) and dual GLP-1/GIP agonists (tirzepatide) directly exploit the gut-brain satiety axis described here, producing substantial weight loss (commonly 10–20% of body weight in clinical trials) by mimicking the natural post-meal GLP-1 signal at central and peripheral receptors. Setmelanotide, an MC4R agonist, is now used specifically for patients with genetic defects upstream of MC4R (such as POMC or leptin receptor deficiency) where the melanocortin pathway itself remains intact. Bariatric surgery (Roux-en-Y bypass, sleeve gastrectomy) achieves durable weight loss substantially through hormonal remodeling — reduced ghrelin, exaggerated GLP-1/PYY responses — rather than mechanical restriction alone, reshaping surgical practice and patient counseling. Cachexia and anorexia of chronic disease (cancer, heart failure, chronic kidney disease) involve pathological elevation of inflammatory cytokines that hijack the same hypothalamic circuitry, suppressing appetite inappropriately; understanding the ARC/NTS axis has driven trials of ghrelin agonists (anamorelin) to counteract this. Pediatric and genetic obesity clinics now routinely sequence LEP, LEPR, POMC, and MC4R in severe early-onset obesity, since a molecular diagnosis changes both prognosis and available targeted therapy. Finally, understanding these circuits informs public health nutrition policy, since highly palatable, energy-dense processed foods can produce supraphysiological hedonic signals that override the homeostatic ARC/NTS system, a mechanism now studied under the heading of "food reward" circuitry involving mesolimbic dopamine pathways working alongside — and sometimes overriding — the homeostatic system described here.

Ask the tutor: How do the hypothalamic appetite circuits described here communicate with peripheral organs like liver and muscle to coordinate whole-body energy balance?