Some entries in this library exist because a compound is confusing. This one exists because a name is confusing, and the name is now searched tens of thousands of times a month.
There is no GLP-3 in a human being. There is one in a shark. Neither fact is what the search is looking for.
Where the human list actually stops
The hormones in this family all come from a single precursor protein called proglucagon, encoded by the gene GCG. The precursor is 180 residues long, and enzymes cut it into different products in different tissues. Its annotated products are:
| Peptide | Residues of the precursor | Length |
|---|---|---|
| Glicentin | 21–89 | 69 aa |
| Glicentin-related polypeptide | 21–50 | 30 aa |
| Oxyntomodulin | 53–89 | 37 aa |
| Glucagon | 53–81 | 29 aa |
| Glucagon-like peptide 1 | 92–128 | 37 aa |
| Glucagon-like peptide 2 | 146–178 | 33 aa |
That is the list. It ends at GLP-2, at residue 178 of a 180-residue protein — there is no room left and no third glucagon-like peptide annotated. A molecule cannot be missing from its own precursor.
But GLP-3 is a real peptide, somewhere else
This is the part that makes the entry worth writing rather than just correcting.
Chondrichthyans — sharks, rays and chimaeras — carry a third proglucagon-derived peptide in addition to GLP-1 and GLP-2, and it is named glucagon-like peptide 3. It is not a naming mistake and it is not speculative; it is a recognised feature of that lineage's proglucagon gene.
The one study asking what it does was published in General and Comparative Endocrinology in May 2024 and worked in the Pacific spiny dogfish, Squalus suckleyi. The researchers perfused GLP-3 at 10 nM for fifteen minutes and looked for metabolic effects. What they found was mostly an absence: no change in the rate of glucose or oleic acid acquisition in the spiral valve, and no measurable effect on intermediary metabolism in that tissue. The one thing that did move was ketone handling — flux through the perfused liver switched from net production to net consumption.
So the real GLP-3, in the only animal that has one and in the only experiment that has tested it, did not shift glucose. Anybody hoping the name implied a stronger metabolic drug is reading the literature backwards.
Counting the literature, and why the count is misleading
Eleven PubMed records contain the string "GLP-3". That sounds like a small but genuine body of work. It is not, because most of those eleven are not about this at all:
- Glypican-3 abbreviated GLP-3, in a liver-cancer diagnostic paper
- "GLP 3-month", meaning a Good Laboratory Practice study of three months' duration, in a rat toxicology paper
- Polysaccharide fraction labels — GLP-3 as the name of the third fraction off a column — in three separate natural-product chemistry papers
- A roundworm gene: glp-3 in Caenorhabditis elegans, reported in Genetics in 1997 as required for mitosis and meiosis in the germ line, and entirely unrelated to glucagon
Two of the eleven concern the fish peptide. That is the whole real literature.
For scale on the same day, GLP-2 — a peptide that does exist in humans, and which almost nobody searches for — returned 1,613 records. The thing with a name people search has eleven papers, most of them about something else; the thing with no name recognition has a century of work behind it.
What people mean when they type it
Almost always: a drug that acts at more than one receptor. GLP-1 drugs became household names, then came dual agonists, then triple agonists, and "GLP-3" is the sensible-sounding guess for the next rung — the same hormone, one number stronger.
The drugs are real and the guess is wrong. Retatrutide engages three receptors — GIP, GLP-1 and glucagon — in one molecule. The three counts receptors, not peptides. Tirzepatide engages two. And a genuinely new arrival in this class is not a peptide at all: orforglipron is a small molecule that switches on the same GLP-1 receptor from a completely different chemical starting point.
The registry check settles it. On 2026-09-16 ClinicalTrials.gov held zero studies for GLP-3 or glucagon-like peptide-3, and 34 for retatrutide. The same query, run the same minute. A drug in development appears in that registry under its code number years before anyone gives it a commercial name, so an empty result for a name that is already being searched heavily means the name arrived before the molecule — and in this case without one.
If something is being sold under this name
Then the label is the problem, not the starting point for research. There is no human molecule called GLP-3, so there is no published sequence, no molecular weight and no specification that a certificate of analysis could be checked against. A buyer cannot verify what is in the vial because there is no agreed answer to what should be in it.
That is a different situation from the rest of this library, where a compound is at least defined even when the evidence for it is thin. Here the name is the whole exercise, and it does not survive being read carefully. Any decision about taking anything belongs with a clinician.