Oxytocin and vasopressin are nine-residue peptides made in the same part of the brain, stored in the same gland, and different from each other at only two positions. One contracts the uterus and triggers milk ejection; the other tells the kidney to keep water. Both are among the most heavily studied peptides in biology, and oxytocin was the first peptide hormone to be made in a laboratory[1].
They are also the clearest example in this field of how far a reputation can run ahead of the evidence. Oxytocin is widely known as the “love hormone”, and that name rests on experiments in rodents whose authors explicitly warned against reading them across to people[2].
| Oxytocin | Cys-Tyr-Ile-Gln-Asn-Cys-Pro-Leu-Gly-NH2 |
|---|---|
| Vasopressin | Cys-Tyr-Phe-Gln-Asn-Cys-Pro-Arg-Gly-NH2 |
| Shared features | Nine residues; disulfide between Cys1 and Cys6; C-terminal glycinamide |
| Molecular weight | 1007.2 Da and 1084.2 Da; calculated |
| Made by | Magnocellular neurons of the paraventricular and supraoptic nuclei of the hypothalamus |
| Released from | Posterior pituitary, into the blood |
| Receptors | OT receptor; vasopressin V1a, V1b and V2 receptors; all G protein-coupled |
Two hormones, two differences
Every neurohypophysial hormone shares the same plan: a disulfide bridge between the cysteines at positions 1 and 6 closes a ring of six residues, and a tail of three more residues ends in a C-terminal glycinamide[3]. What separates the two families is the residue at position 8. The vasopressin family carries a basic amino acid there, arginine or lysine; the oxytocin family carries a neutral one[3].
The textbook version of the rule is simple: a basic residue at position 8 for vasopressin receptors, isoleucine at position 3 for oxytocin receptors[3]. Measured directly, the selectivity is much softer than that. The oxytocin receptor binds oxytocin with only about 10 times the affinity it has for vasopressin, and vasopressin still acts on it as a partial agonist: around 100 times more vasopressin is needed to produce the same response[3].
Position 8 contributes less to that binding than the ring does. Peptides sharing the same ring bind the oxytocin receptor equally well whether they carry arginine or leucine at position 8, while isoleucine at position 3, which sits inside the ring, contributes more to affinity than phenylalanine does[3]. So the two substitutions shift the preference rather than switching it. Pigs use lysine instead of arginine at position 8 and the hormone works[4], which shows that what matters at that position is a basic residue rather than arginine specifically.
The pairing is old. Almost every vertebrate has one peptide of each kind, and two evolutionary lines have been proposed: isotocin to mesotocin to oxytocin, tied to reproduction, and vasotocin to vasopressin, tied to water balance. Calculations from the genes put the common ancestral gene at more than 500 million years old, with the duplication occurring after the cyclostomes branched off[3]. Mesotocin, not oxytocin, is the oxytocin-like hormone in most land vertebrates from lungfishes to marsupials[3].
The first peptide hormone ever made
By the early 1950s, hydrolysis of purified oxytocin had accounted for its parts: one equivalent each of leucine, isoleucine, tyrosine, proline, glutamic acid, aspartic acid, glycine and cystine, three equivalents of ammonia, and a molecular weight near 1000. Degradation pointed to a cyclic disulfide[1].
In 1953 Vincent du Vigneaud and five colleagues at Cornell built the proposed structure from scratch. They assembled a fully protected chain, removed the benzyl protecting groups with sodium in liquid ammonia, and closed the disulfide by simply aerating a dilute aqueous solution at pH 6.5. From 360 mg of the protected peptide they recovered material with roughly 29 000 units of biological activity[1].
Proving the synthetic peptide was the hormone took six independent comparisons. It matched natural oxytocin in potency on repeated side-by-side assay; in specific rotation, −26.1° against −26.2°; in the melting point and crystal form of its flavianate derivative; in amino acid composition; in distribution coefficients in two solvent systems and in electrophoretic mobility at two pH values; and no differences were found in the infrared spectra[1]. The synthetic material contracted the isolated rat uterus, stimulated labour, and produced milk ejection within 20 to 30 seconds after about 1 µg given intravenously[1].
The authors stated the result carefully: "If the synthetic product truly represents oxytocin, which it does in so far as we have been able to ascertain, this would constitute the first synthesis of a polypeptide hormone."[1] It did, and du Vigneaud received the Nobel Prize in Chemistry in 1955[5]. The name oxytocin is older still, coined by Henry Dale from the Greek for swift birth[5].
A note for anyone opening the 1953 paper: its title reads "The Synthesis of an Octapeptide Amide with the Hormonal Activity of Oxytocin", while the text of the same paper calls the product a nonapeptide derivative[1]. The difference is in how the cystine is counted, as one residue or as the two cysteines it is built from. The hormone has nine residues.
How the body makes and releases them
For release into the bloodstream, both hormones are made by large neurons in two regions of the hypothalamus, the paraventricular and supraoptic nuclei, whose axons end in the posterior pituitary. The peptide is synthesised as part of a larger precursor, processed on its way down the axon, and stored in the nerve terminals until a nerve signal releases it into the blood[6].
Each precursor also carries a binding protein. The oxytocin gene sits on chromosome 20p13 and has three exons; the first encodes a signal sequence, the hormone itself, the Gly-Lys-Arg processing signal and the first nine residues of a neurophysin[3]. Neurophysin is a small protein of 93 to 95 residues, rich in disulfides, which binds the hormone one to one and is needed to pack and store it in the granules before release[3].
The vasopressin precursor yields three fragments at once: vasopressin, neurophysin II, and a C-terminal glycopeptide now called copeptin[4]. Copeptin has become useful in the clinic for a practical reason. Vasopressin itself is hard to measure in blood because it is small, present in low amounts, short-lived, and binds heavily to platelets; copeptin is released alongside it and is far more stable, so it is increasingly measured as a stand-in[4].
A second route matters for vasopressin. Smaller neurons of the paraventricular nucleus send their axons not to the posterior pituitary but to the median eminence, releasing vasopressin into the portal vessels that supply the anterior pituitary, where it drives the release of ACTH[6].
Once in the blood, oxytocin does not last. Its half-life there is 1 to 2 minutes, against 28 minutes in the cerebrospinal fluid, where it is present at 10 to 50 pM[3]. A hormone that disappears from the blood this quickly can only act in pulses, which is what the release pattern below describes.
What oxytocin does
The two classical actions are contraction of the uterus during labour and milk ejection during breastfeeding[3].
Milk ejection is the better-established of the two. Suckling generates sensory impulses that travel to the oxytocin neurons of the hypothalamus, which fire in a synchronised burst lasting 3 to 4 seconds and recurring every 5 to 15 minutes. Each burst releases a pulse of oxytocin into the blood, which contracts the myoepithelial cells around the milk-producing alveoli and ducts. In humans, milk begins to flow within 30 to 60 seconds of a baby starting to suck[3]. The reflex can be triggered before any touch, by the baby crying[3]. In mice engineered to lack oxytocin, milk ejection fails[3].
Labour is more complicated than the textbook version. The uterus becomes far more sensitive to oxytocin as birth approaches, not because more hormone is released but because the muscle grows more receptors: in humans, myometrial oxytocin receptor levels rise during early labour to around 200 times those in the non-pregnant state, then fall rapidly after delivery[3]. Yet mice that cannot make oxytocin at all give birth normally[3]. The knockout does not identify what starts labour instead; it shows that in mice oxytocin is not indispensable for it, and the authors of the review take this to mean that other uterotonins are also at work[3].
What vasopressin does
Vasopressin is named for the blood vessels, but its everyday job is in the kidney. Its older name, antidiuretic hormone, describes it better.
At the V2 receptor on the collecting duct, vasopressin acts through a chain that is now known step by step. The receptor activates the G protein Gs, which stimulates adenylate cyclase to make cAMP; cAMP activates protein kinase A; and the kinase phosphorylates the water channel aquaporin-2 at serine-256 in its cytoplasmic tail. Phosphorylated channels, held in intracellular vesicles, are then inserted into the apical membrane, and water can cross[7]. Phosphorylation happens within a minute of the hormone arriving[7].
The evidence that this particular residue is the switch is direct: when serine-256 is replaced by alanine, neither vasopressin nor forskolin can move aquaporin-2 to the membrane, while the unmodified channel relocates normally[7]. At least three of the four subunits in each channel need to be phosphorylated for it to be transported[7]. The idea that water channels sit in vesicles and are shuttled into the membrane on demand was proposed long before the channels themselves were identified[7].
Vasoconstriction is real but requires more hormone. In normally hydrated people, plasma vasopressin after an overnight fast runs about 2 to 4 pg/ml, rising to 20 to 30 pg/ml under water restriction[6]. The blood levels needed to raise systemic blood pressure are much higher than those needed for the antidiuretic effect[6]. The V1a receptor on vascular smooth muscle is what does the constricting: in mice lacking it, the pressor response to vasopressin disappears, and their resting blood pressure is lower than normal[6].
The third receptor, V1b, sits on the corticotrophs of the anterior pituitary and mediates the release of ACTH[6], which is how vasopressin takes part in the stress response.
Vasopressin is used in intensive care to support blood pressure, and the largest trial of it has a careful result. VASST randomised 778 patients with septic shock to low-dose vasopressin or to norepinephrine, each added to the vasopressors they were already receiving. Mortality at 28 days was 35.4% with vasopressin and 39.3% with norepinephrine, a difference that was not statistically significant (P = 0.26); at 90 days it was 43.9% against 49.6% (P = 0.11). Serious adverse events were equally common in both groups[8]. The trial also documented how low these patients’ own vasopressin was: baseline plasma concentrations had a median of 3.2 pmol/l, and the infusion raised them to roughly 70 to 100 pmol/l[8]. A subgroup with less severe shock did better on vasopressin, but the test for interaction was not significant, and the authors present that finding only as a hypothesis for future trials[8].
The “love hormone” question
Much of the influential evidence for a role in social bonding comes from one group of animals. Only 3 to 5% of mammals are socially monogamous, and prairie voles are. In a partner-preference test, oxytocin infused into the brain of a female prairie vole speeds the formation of a pair bond, vasopressin does the same in males, and receptor antagonists prevent it. Monogamous voles carry more oxytocin receptors in the nucleus accumbens and more vasopressin V1a receptors in the ventral pallidum than their non-monogamous relatives[2]. Transferring the V1a receptor gene into the ventral pallidum of a non-monogamous meadow vole strengthens its partner preference[2].
These are striking experiments. Their authors were also explicit about their limits, writing that "there are no hard data demonstrating common physiological mechanisms for pair-bond formation in voles and man"[2].
In people, the hormone is usually given as a nasal spray, and a 2016 methodological analysis found a statistical problem in the literature built on that method. Pooling the meta-analyses available at the time, it found an average effect size of d = 0.28 in healthy subjects, with a median sample size of 49. That combination gives those studies about 16% statistical power; for clinical trials, with a median of 26 participants, it gives 12%. Detecting effects of that size with the usual 80% power would need around 350 and 310 participants[9]. Of 33 intranasal oxytocin papers published in 2014, 29 reported at least one positive finding, yet only 17% of all the statistical tests in them were significant, and just three of the 33 mentioned any correction for multiple comparisons[9]. The authors' conclusion is that most published positive findings in this field are likely to be false positives, while noting that the underlying hypothesis is not the problem[9].
A large clinical trial fits that picture. Published in 2021, it gave intranasal oxytocin or placebo to 290 children and adolescents with autism spectrum disorder for 24 weeks, at a target dose of 48 IU a day; 277 of them contributed data to the main analysis. On the primary measure, a 13-item social-withdrawal scale on which a higher score means less social interaction, the least-squares mean change from baseline was −3.7 with oxytocin and −3.5 with placebo. Both groups improved; the difference between them was −0.2 points (95% CI −1.5 to 1.0; P = 0.61). The secondary measures showed no difference either, and adverse events were as common on placebo as on the drug[10]. By the time the trial reported, many children with autism were already being given intranasal oxytocin[10].
Part of the difficulty is physical. How much of a nasal dose reaches the brain, and how quickly, is still uncertain[10], and the same point has been raised as a reason why effects in humans are small to begin with[9].
In short
- Oxytocin and vasopressin are nine residues long, share a Cys1–Cys6 disulfide and a C-terminal glycinamide, and differ only at positions 3 and 8. Those substitutions shift receptor preference without making either hormone exclusive: the oxytocin receptor binds oxytocin only about 10 times more tightly than vasopressin, which acts on it as a partial agonist.
- Oxytocin was the first peptide hormone synthesised, by du Vigneaud and colleagues in 1953, and matched the natural hormone on six independent tests.
- Both are made in the hypothalamus with a neurophysin carrier and released from the posterior pituitary. Oxytocin's half-life in blood is 1 to 2 minutes. Vasopressin's precursor also yields copeptin, which is measured in its place because it is far more stable.
- Oxytocin is required for milk ejection in mice; for labour it is not, although human uterine oxytocin receptors rise around 200-fold by early labour.
- Vasopressin's main job is water retention through V2 receptors, cAMP and the phosphorylation of aquaporin-2 at serine-256. Raising blood pressure needs much higher levels than the antidiuretic effect does.
- The “love hormone” evidence is strong in voles and weak in humans: the vole researchers denied any demonstrated common mechanism with people, a 2016 analysis put the statistical power of the intranasal studies it examined at 12–16%, and a 290-participant autism trial found no effect.
References
- du Vigneaud V, Ressler C, Swan JM, Roberts CW, Katsoyannis PG, Gordon S (1953). The synthesis of an octapeptide amide with the hormonal activity of oxytocin. Journal of the American Chemical Society 75(19):4879–4880. DOI: 10.1021/ja01115a553
- Young LJ, Wang Z (2004). The neurobiology of pair bonding. Nature Neuroscience 7(10):1048–1054. DOI: 10.1038/nn1327
- Gimpl G, Fahrenholz F (2001). The oxytocin receptor system: structure, function, and regulation. Physiological Reviews 81(2):629–683. DOI: 10.1152/physrev.2001.81.2.629
- Antoni FA (2017). Vasopressin as a stress hormone. In: Stress: Neuroendocrinology and Neurobiology, chapter 9, pages 97–108. Elsevier. DOI: 10.1016/B978-0-12-802175-0.00009-7
- Viero C, Shibuya I, Kitamura N, Verkhratsky A, Fujihara H, Katoh A, et al. (2010). Oxytocin: crossing the bridge between basic science and pharmacotherapy. CNS Neuroscience & Therapeutics 16(5):e138–e156. DOI: 10.1111/j.1755-5949.2010.00185.x
- Koshimizu T, Nakamura K, Egashira N, Hiroyama M, Nonoguchi H, Tanoue A (2012). Vasopressin V1a and V1b receptors: from molecules to physiological systems. Physiological Reviews 92(4):1813–1864. DOI: 10.1152/physrev.00035.2011
- Nielsen S, Frøkiær J, Marples D, Kwon TH, Agre P, Knepper MA (2002). Aquaporins in the kidney: from molecules to medicine. Physiological Reviews 82(1):205–244. DOI: 10.1152/physrev.00024.2001
- Russell JA, Walley KR, Singer J, Gordon AC, Hébert PC, Cooper DJ, et al. (2008). Vasopressin versus norepinephrine infusion in patients with septic shock. New England Journal of Medicine 358(9):877–887. DOI: 10.1056/NEJMoa067373
- Walum H, Waldman ID, Young LJ (2016). Statistical and methodological considerations for the interpretation of intranasal oxytocin studies. Biological Psychiatry 79(3):251–257. DOI: 10.1016/j.biopsych.2015.06.016
- Sikich L, Kolevzon A, King BH, McDougle CJ, Sanders KB, Kim SJ, et al. (2021). Intranasal oxytocin in children and adolescents with autism spectrum disorder. New England Journal of Medicine 385(16):1462–1473. DOI: 10.1056/NEJMoa2103583