Uroguanylin and Guanylin: The Gut Peptides Found by Chasing a Toxin

Peptide & protein science

Uroguanylin and its sibling guanylin are two of the very few human hormones that were found by reverse-engineering a poison. In 1990, researchers established that the heat-stable enterotoxin from E. coli — the molecule behind a great deal of travellers’ diarrhoea — works by binding an enzyme on the gut lining called guanylate cyclase C. That result implied something odd: the human intestine had built a receptor that a bacterium happened to have a key for. Somebody went looking for the lock’s real key, and found two.

Found by working backwards from a toxin

The chronology is unusually clean, which is what makes it a good story.

1990. Schulz and colleagues, writing in Cell, cloned the receptor and showed that “a heat-stable enterotoxin receptor responsible for acute diarrhea is a plasma membrane form of guanylyl cyclase.” The receptor is now known as GC-C, encoded by the gene GUCY2C.

1992. Currie and colleagues stated the problem plainly in their abstract: an endogenous ligand for this cyclase had not previously been found. Using a human colon cell line that spikes its cyclic GMP when the toxin is applied as a bioassay, they purified a 15-amino-acid peptide from rat jejunum and named it guanylin. It showed clear sequence homology with the bacterial enterotoxins, and it lost all activity when its disulfide bonds were reduced.

1993. Hamra and colleagues purified a second peptide from the urine and intestinal mucosa of opossums, named it uroguanylin for where they found it, and reported it was about ten-fold more potent than guanylin — though both were still less potent than the bacterial toxin. The closing line of that abstract planted a hypothesis that people are still testing: receptors in the kidney’s proximal tubules, and peptide in the urine, hinted that these molecules might regulate kidney function too.

Then the direction of travel reversed. Having learned the receptor from a toxin and the hormones from the toxin’s receptor, drug developers went back and built molecules deliberately. Two of them are now approved medicines.

What guanylin and uroguanylin actually are

Both are short, cysteine-rich peptides cut from the C-terminal end of larger precursors. Guanylin is the product of the GUCA2A gene; uroguanylin comes from GUCA2B. Mature guanylin is 15 amino acids long; mature human uroguanylin is 16, ending in an extra leucine that turns out to matter. Each carries two disulfide bonds, arranged so that the first cysteine pairs with the third and the second with the fourth.

That arrangement produces a genuinely unusual property. With that particular disulfide pattern, the loop in the middle of the peptide can sit either above or below the plane formed by the cross-links, giving two distinct three-dimensional forms — topoisomers — from one identical covalent structure. They are the same molecule by every chemical measure and different molecules by shape. Only one of them is an effective agonist at GC-C; whether the other is truly inert or merely weak is disputed in the literature.

How the receptor works

GC-C is not a G-protein-coupled receptor, which makes it an interesting contrast to most of the hormone receptors covered elsewhere on this site. It is a receptor that is an enzyme: the outside of the protein binds the ligand, and the inside is a guanylate cyclase that starts manufacturing cyclic GMP.

The rest of the chain is described in identical terms in both approved drugs’ FDA labels. Rising intracellular cyclic GMP opens the CFTR channel — the same channel that is defective in cystic fibrosis — which pushes chloride and bicarbonate into the gut lumen. Water follows osmotically. The stool softens and transit speeds up.

Run that circuit at normal intensity and you have a bowel movement. Run it flat out, with a bacterial toxin that binds harder and is not easily switched off, and you have travellers’ diarrhoea. It is the same mechanism at two different volumes.

There is a second, more speculative limb. Some of the cyclic GMP is exported out of the epithelial cell rather than staying inside it, and the proposal is that this extracellular cGMP quietens pain-sensing nerve endings in the gut wall. The elegance of the idea is that the mucosa expresses GC-C while the sensory neurons do not, so the signal has to be handed across. The supporting work is rodent electrophysiology, and the FDA labels are careful about it: both describe the effect as occurring “in animal models of visceral pain.” The pain benefit in adult trials is real; the explanation for it is provisional.

The two drugs, and the detail nobody should blur

Linaclotide (Linzess) was first approved in the United States in 2012. It is 14 amino acids with three disulfide bonds. Its FDA label says it is “structurally related to human guanylin and uroguanylin” — but the third disulfide bond is the toxin’s signature, not the human peptides’, so structurally it sits closer to E. coli STa than to either endogenous ligand. It is approved for irritable bowel syndrome with constipation and chronic idiopathic constipation in adults, and, through a series of later extensions, for functional constipation in children.

Plecanatide (Trulance) followed in 2017. It is 16 amino acids with two disulfide bonds and differs from human uroguanylin at exactly one position — an aspartate replaced by a glutamate. It is approved for adults only, for the same two indications.

Both carry a boxed warning about serious dehydration in young children, and this is the single most important thing to get right: the two warnings use different age cut-offs. Linaclotide is contraindicated under 2 years of age. Plecanatide is contraindicated under 6, and its label additionally says to avoid use from 6 to under 18. They are not interchangeable statements.

The reason those warnings exist is mechanistic rather than toxicological. GC-C expression in the intestine is age-dependent and highest in the youngest gut. In neonatal mice, a single clinically relevant dose of linaclotide caused deaths from dehydration within 24 hours. When the FDA looked for human data, a study measuring GC-C messenger RNA in gut samples from 99 children concluded there was insufficient information to assess the risk under 2 years of age. The warning marks the boundary of what is known, not a proven human harm.

Diarrhoea is the most common adverse reaction to both drugs, which is exactly what you would expect from a mechanism that works by moving water into the bowel. In the adult IBS-C trials of linaclotide at 290 micrograms, diarrhoea was reported in 20% of patients versus 3% on placebo. In the adult IBS-C trials of plecanatide at 3 mg, it was 4.3% versus 1%. Those figures come from different trials with different populations, doses and adverse-event definitions, and there has never been a head-to-head study, so they cannot be read against each other.

One more consequence of the design is worth spelling out. Both drugs are essentially non-systemic: plasma concentrations sit below the limit of quantitation, and the labels state outright that area under the curve, peak concentration and half-life cannot be calculated. In a plecanatide food-effect study using three times the recommended dose, the drug was detectable in exactly one of 24 subjects. That is why no drug-drug interaction studies were needed, and why the safety story is not organ toxicity but too much of the intended effect in the intended place.

The pH claim, examined

Plecanatide is frequently described as pH-sensitive — active in the acidic microclimate of the upper small intestine and quieter further down — and this is offered as a reason it might behave more gently than linaclotide. It is worth separating the layers of that claim.

The underlying biology is real, and it is about the natural peptides. A 1997 study found that at a mucosal acidity comparable to the gut surface, uroguanylin’s receptor-binding affinity rose roughly ten-fold, while guanylin’s fell about a hundred-fold — at pH 5.0, guanylin was essentially inactive. The acid sensitivity traces to uroguanylin’s acidic N-terminal residues. That is solid primary science.

The extension to plecanatide is thinner. The paper most often cited for it is a computational molecular-dynamics modelling study from 2017, not an experimental or clinical one. And the words “pH,” “acidic” and “pH-dependent” appear nowhere in the FDA-approved Trulance label, including its mechanism-of-action section.

And it has never been tested in people against the alternative. No head-to-head randomised trial of plecanatide versus linaclotide exists, and indirect network meta-analysis has not found significant differences between the individual drugs. The pH story is mechanistically plausible and unproven as a clinical differentiator. Treat it as the former.

A second life as a cancer target

GC-C has an unusual property for a tumour antigen: in healthy people it is confined to the intestinal lining and is largely hidden from the immune system, yet it is overexpressed in gastrointestinal adenocarcinomas. That combination has driven a small vaccine programme. A phase 1 trial of an adenoviral vaccine against GUCY2C in ten patients with resected early-stage colon cancer reported no adverse events above grade 1, though pre-existing antibodies to the viral vector blunted responses; a redesigned chimeric vector has since moved into a phase 2a dose-finding study. This is early work, and none of it has produced an approved product.

Does the gut tell the kidney about salt?

Here is the hypothesis that has followed uroguanylin since 1993. A salt load taken by mouth produces more sodium excretion than the same amount of salt given intravenously. Something in the gut appears to be warning the kidney that salt is on the way, before blood sodium has moved. Uroguanylin is the leading candidate for that messenger — an intestinal-renal axis.

The load-bearing evidence is a 2003 study in the Journal of Clinical Investigation: mice lacking uroguanylin have higher blood pressure and an impaired ability to excrete a salt load delivered into the gut, mostly because their kidneys hold on to sodium when they should not. There is also an elegant supporting detail — uroguanylin appears to reach the kidney largely as the unprocessed propeptide and to be converted to its active form within the renal tubules themselves.

But the honest summary is that this remains a mouse-genetics hypothesis. There is no human knockout, no randomised human natriuresis trial, and no approved uroguanylin-based natriuretic drug. Human data are sparse and have not cleanly reproduced the rodent findings. And there is a structural reason the question stays open: both approved GC-C drugs were deliberately designed never to leave the gut, so neither of them can test the endocrine hypothesis in people at all.

Frequently asked questions

If these drugs work the same way as the bacteria that cause travellers’ diarrhoea, how can they be safe?

They use the same receptor, but not the same intensity. The bacterial toxin binds harder and is not readily switched off, which is why it produces a purge rather than a bowel movement. The approved drugs are dosed to nudge the circuit, and diarrhoea is the most common adverse reaction precisely because the mechanism is shared. It is a difference of degree, deliberately engineered.

Why do these drugs carry a boxed warning for young children if they barely get absorbed?

Because absorption is not the risk. GC-C expression is age-dependent and highest in the youngest intestine, so the same luminal dose produces a much larger fluid shift. Neonatal mice died of dehydration within a day of a single dose, and the human data the FDA reviewed were not sufficient to rule the risk out. Note again that the two drugs use different cut-offs — under 2 for linaclotide, under 6 for plecanatide.

Is uroguanylin the same thing as a natriuretic peptide?

Not in the family sense. The classical natriuretic peptides — ANP, BNP and CNP — are a structurally distinct group that act on their own receptors, and uroguanylin is not one of them. It has been described as having natriuretic actions, which is a claim about physiology rather than family membership, and even that claim rests largely on mouse genetics.

Could someone use guanylin or uroguanylin as research peptides the way other peptides are used?

The pharmacology makes this awkward. Both peptides act on receptors facing the inside of the gut, and the entire therapeutic logic of the approved analogs is that they stay in the lumen and are never absorbed. A molecule delivered into the bloodstream is not reaching the compartment where GC-C signalling normally happens. That mismatch is worth understanding before drawing conclusions from any experiment involving them.

References

Informational only — not medical advice. Nothing here is a treatment recommendation or a dose for any person; prescribing information for linaclotide and plecanatide is linked above and should be read in full by anyone with a clinical question. Consult a qualified healthcare professional about your own care. Intended for readers 21 and over.