Is there a GLP-4?

The story so far reads like a version history. First a single agonist — semaglutide, one receptor. Then a dual — tirzepatide, two. Then a triple — retatrutide, three receptors from one molecule. One, two, three. So the next leap is obvious, isn’t it? A bigger number. A GLP-4.

Except there is no GLP-4. There is no GLP-3 either. And the reason that sounds surprising is the reason this whole essay exists: the intuition that progress means a higher number is exactly the intuition the field is leaving behind.

Start with what GLP-1 actually is. The “1” does not mean “first generation.” It is an accident of anatomy. The body makes a large precursor protein called proglucagon, then cuts it into pieces — and depending on which cell does the cutting, you get glucagon, glucagon-like peptide-1, and glucagon-like peptide-2, plus a few others. GLP-1 and GLP-2 are simply the first and second GLP-shaped fragments researchers named as they mapped the precursor. The number is a catalog index, not a performance tier.

And GLP-2 — the one number that does exist beyond GLP-1 — proves the point, because it isn’t a stronger metabolic drug at all. GLP-2 grows intestinal lining. Its analog, teduglutide, is used for short bowel syndrome, a gut-repair problem with nothing to do with appetite or blood sugar. If “GLP-2” meant “GLP-1 but more,” it would act on metabolism harder. Instead it does something else entirely. The sequence of numbers was never a ladder of power; it was a list of what fell out of one prohormone.

So if not a bigger number, what? Look at what is actually in development and a different shape appears. The frontier isn’t one axis pushed further — it’s several new axes opening at once, some of them pulling in directions the “more agonism” story would call backwards.

Four moves, none of which is “the same thing, stronger.” Take them one at a time.

The triple agonist recruited GIP and glucagon alongside GLP-1. The next recruit isn’t another incretin at all — it’s amylin, a hormone co-secreted with insulin that curbs appetite through a separate brainstem circuit. Because it works through a different door than the incretins, its effect stacks rather than overlaps. Cagrilintide (a long-acting amylin analog), the cagrilintide-plus-semaglutide combination, and amycretin (a single molecule that is both a GLP-1 and an amylin agonist) are all investigational bets on exactly that: not a fourth GLP, but a whole new signaling axis added to the mix.

Here is the move that breaks the intuition cleanly. Tirzepatide is a GIP agonist — it turns the GIP receptor on. Yet one of the most closely watched next-generation molecules, maridebart cafraglutide, pairs GLP-1 agonism with GIP receptor antagonism — it turns the same receptor off. Two drugs, opposite actions at the very same target, and both drive weight loss in trials.

That should be impossible under a “more activation is better” model, and it is the strongest evidence that the model is wrong. The GIP system can be tuned productively from either direction depending on what the rest of the molecule is doing. When the field can’t even agree on the sign of a receptor’s contribution, “just crank it to 4” stops meaning anything.

A different frontier ignores receptors entirely and attacks the delivery problem. Peptides are, chemically, food — swallow one and your gut digests it before it works, which is why this class has lived on the needle. Getting the same biology into a tablet is its own kind of breakthrough, and it’s arriving two ways: oral formulations of existing peptides, and non-peptide small molecules like orforglipron that hit the GLP-1 receptor but survive the stomach because they were never peptides to begin with. A once-daily pill — or a once-monthly injection, another active direction — changes who will actually use these drugs far more than a marginal bump in receptor count ever could.

The last move redefines the goal. The first era chased a single number on the scale. But weight lost is a mix of fat and lean tissue, and stripping muscle alongside fat is not a win. So the frontier turns to the quality of the loss: pairing an incretin with myostatin inhibition to burn fat while sparing — even building — muscle. That is a myostatin-pathway drug bolted onto a metabolic one: a completely different mechanism recruited to make the same weight loss better, not larger. No amount of GLP-agonism, at any number, does that.

None of this is an accident of naming — there are hard reasons the field stopped chasing a bigger single number:

• Receptors adapt. flood a signaling system and it down-regulates; past a point, more agonism buys less and less real effect
• Tolerability is the true ceiling. the nausea that shadows this class limits how hard any one axis can be pushed — so you add new axes at moderate strength instead of maxing one out
• More weight isn't strictly better. once loss is dramatic, its composition — fat versus muscle — matters more than its size, and that's a quality problem, not a potency one

Push a single lever hard enough and you meet diminishing returns, a nausea wall, and a quality problem all at once. Adding a lever sidesteps all three. That is why the map fans out instead of marching up.

The honest near-term answer: amylin combinations and oral delivery reaching the clinic, the GIP agonism-versus-antagonism question getting settled by data rather than theory, and metabolic drugs increasingly judged by the muscle they keep, not just the fat they shed. Every one of these is investigational — promising, not proven — which is why the ones already in the catalog wear honest evidence badges rather than settled ones.

But the shape of the future is already clear, and it isn’t a number. The instrument that learned to play chords isn’t about to play a higher note — it’s growing new strings, learning to bend them the other way, and finding out it can be played without the needle. That’s a richer frontier than GLP-4 ever promised. The best part of this field is that the interesting direction turned out to be sideways.