The OG Peptide

I was having a conversation with a doctor, and peptides came up.

He was not rude about it. He was direct. Peptides are not researched, he said. It is bro science. People watch a YouTube video and start injecting themselves with something nobody has ever studied.

I did not argue with him. He is a doctor and I am not. But I did ask him about two things.

Insulin. And the GLP-1 drugs.

Both of those are peptides. One of them has been in human medicine for over a hundred years. So I went home and spent a while reading, because I wanted to know whether I had a point or whether I was just being difficult.

Turns out he is right and he is wrong, and which one depends entirely on what he meant by the word peptide.

START HERE

What a peptide actually is

Your body is built out of amino acids. Think of them like beads on a string. String a few beads together and you have a peptide. String a lot of them together and fold the whole thing up and you have a protein.

That is the entire difference. Length. The National Library of Medicine puts a short peptide at roughly 2 to 12 amino acids, a longer chain at about 13 or more, and it describes proteins as larger versions of peptides.

So peptide is not a product. It is not a brand, a supplement aisle, or a category of thing you buy. It is a shape. It is how long the string is.

Which is why "are peptides researched" is a broken question. It is like asking whether pills work. Which pill?

1921

The OG peptide

If you want to know how well studied a peptide can get, you start with insulin.

Insulin is a chain of 51 amino acids in two pieces, an A chain of 21 and a B chain of 30, clipped together by sulfur bonds. It is the hormone that tells your body what to do with sugar. Without it, your blood turns into syrup and you die.

In May of 1921, a surgeon named Frederick Banting and a medical student named Charles Best started working in a University of Toronto lab that belonged to a physiologist named J. J. R. Macleod. The FDA's own history exhibit points out that neither Banting nor Best was an experienced researcher. In December, a visiting biochemist named James Collip joined them and got the extract clean enough to put into a person.

On January 11, 1922, a 14 year old boy named Leonard Thompson got the first insulin injection ever given to treat diabetes.

The first shot did not work.

They went back, purified it further, and tried again. His medical records say he "became brighter, more active, looked better and said he felt stronger." He went home a few months later. He lived another 13 years and died in 1935 at 27 years old.

Before that shot, type 1 diabetes was a death sentence and the best available treatment was starving the patient.

Banting, Best and Collip held the US patents. They sold them to the University of Toronto for one dollar each.

Banting and Macleod took the 1923 Nobel Prize in Physiology or Medicine, and the citation is four words long: "for the discovery of insulin." They were unhappy that Best and Collip got left out of it, so they split the prize money with them.

1951 TO 1958

The first protein anybody ever read

Here is the part I did not know until I went looking.

For a long time nobody knew what was actually inside a protein. People knew proteins were built out of amino acids. Whether a given protein had one exact order, every single time, or whether it was some kind of jumble, was an open question.

A chemist named Frederick Sanger spent about fifteen years working out the exact order of insulin's 51 amino acids. He published the first chain in 1951, the second in 1953, and the sulfur bonds holding them together in 1955.

He won the 1958 Nobel Prize in Chemistry for it. The award ceremony speech says it flat out: this "was the first time one had succeeded in determining the structure of a substance belonging to this very important group."

Insulin was the first protein human beings ever read end to end. That result is the reason anybody could go looking for the order of anything else.

1982

The first drug ever made this way

It gets better.

Until the early eighties, insulin came out of cows and pigs. Animal supply, animal impurities, and a real fear of running short as more people needed it.

In 1978, scientists at City of Hope and Genentech worked out how to put the human insulin gene into common bacteria and let the bacteria make the stuff for us. Eli Lilly signed on to take it commercial.

On October 28, 1982, after five months of review, the FDA approved Humulin. In the FDA's own words, it was "the first biosynthetic human insulin product and the first approved medical product of any kind that derived from this technology."

Read that again. Not the first insulin made this way. The first approved medical product of any kind, for anything, made this way.

It was a peptide.

1990

And then a lizard

GLP-1 is a peptide too. It is 36 or 37 amino acids, made mostly by cells in your intestine. After you eat, it tells your pancreas to let insulin out, slows your stomach down, drops your blood sugar, and turns your appetite off.

Great hormone. Terrible drug. Your body clears it in about two minutes.

In 1990, an endocrinologist named John Eng at the Veterans Administration center in the Bronx was going through lizard venom. Earlier NIH work had shown that venom from certain lizards made the pancreas swell up. Eng also knew the Gila monster can go long stretches without eating and still hold its blood sugar steady.

So he assayed the venom and found a peptide he named exendin-4. It looks like human GLP-1 and it acts like human GLP-1. The difference is that it stays active for hours instead of two minutes.

They built a synthetic version, called it exenatide, sold it as Byetta, and the FDA approved it in 2005.

A drug from a lizard. I read that three times to make sure I had it right.

THE RECEIPTS

How much study is "studied"

The GLP-1 drugs did not sneak through the back door. Take one trial out of many. SELECT, published in the New England Journal of Medicine in 2023, put 17,604 people on either semaglutide or a placebo and followed them for an average of about 40 months. A major heart event hit 6.5 percent of the drug group and 8.0 percent of the placebo group.

Seventeen thousand people. One trial. One drug.

Then there is the pile of published work. I searched PubMed, which is the National Library of Medicine's database, on September 23, 2026. Insulin comes back with 209,369 records. Semaglutide comes back with 5,716.

That is what "researched" looks like when a peptide actually gets researched.

THE OTHER SIDE

Where the doctor was right

Now the part where he was dead right, and I am not going to soften it.

BPC-157 is one of the peptides people actually buy online. Same PubMed search, same day: 232 records. Almost all of that is animals, or cells in a dish.

So what has actually been done in people? The FDA wrote up its own answer to that this year, and it is short. The briefing document it prepared on BPC-157 lists five human studies. Total. Ever.

Two of them were real randomized, placebo controlled trials, both run in the early 2000s. One in 32 healthy people, one in 53 people with ulcerative colitis. Both of them only ever came out as conference abstracts, which is the science version of a slide at a convention. Neither was ever published as a full paper that anybody could pull apart. The other three are a chart review of 17 knee patients, a 12 person study with no placebo group, and a safety study with TWO people in it.

So it is not zero. It is close to it, and the real problem is that nothing controlled ever got published in full.

There is one proper trial going right now. It is registered as NCT07437547: randomized, double blind, placebo controlled, phase 2, 120 people with hamstring tears, run by a company called Hudson Biotech. It started in February 2026 and it is still recruiting. It is not due to report until 2027.

One trial. One injury. Another year before anybody knows anything.

Now the FDA side of it, and this is the piece that moved recently.

BPC-157 used to sit in a bucket the FDA calls Category 2, which is where the agency puts compounding ingredients it thinks may carry real safety risks. It is not in that bucket anymore. Whoever nominated it pulled the nomination, so today it sits further down the same FDA page under a heading that reads "Bulk drug substances nominated but withdrawn." The page is "Certain Bulk Drug Substances for Use in Compounding that May Present Significant Safety Risks," and the FDA marks it current as of April 22, 2026.

What the FDA says about it did not change. The sentence is still sitting right there, word for word. The agency "has identified no, or only limited, safety-related information for the proposed routes of administration. Therefore, the agency lacks sufficient information to know whether the drug would cause harm when administered to humans."

Read that last line slow. It does not say dangerous. It does not say fine. It says nobody knows.

Then July happened, and you are going to hear about this one, so here it is with the caveats attached. On July 23, 2026, an FDA advisory committee called the Pharmacy Compounding Advisory Committee voted on whether BPC-157 belongs on something called the 503A Bulks List. That list is what makes it legal for a compounding pharmacy to mix a substance up for a patient. The reported vote was 8 in favor, 6 against, 1 abstention.

Three things that vote is not.

It is not binding. The FDA's own page for the meeting says advisory committees "make non-binding recommendations to the FDA, which generally follows the recommendations but is not legally bound to do so."

It is not approval. Nothing got approved that day. Landing on that list means a pharmacy is allowed to make the stuff. It says nothing at all about whether it works.

It is not evidence. A show of hands is not a result. The pile of data on BPC-157 was exactly the same size the day after that vote as it was the day before.

And nothing has happened since. I checked on September 23, 2026. The list itself is a federal regulation, 21 CFR 216.23, and it still holds the same six substances it has held since 2019. BPC-157 is not one of them. The Federal Register, which is where the government has to print it when it actually does something, carries exactly one document that mentions BPC-157, and that is the April 2026 notice announcing the meeting. No rule. No proposal. Nothing.

Here is the part I keep chewing on, though. The FDA's own reviewers walked into that meeting against it. Their briefing document says there is "a lack of evidence to support the effectiveness of BPC-157 (free base) and BPC-157 acetate as a treatment for UC," and that "the data available do not support the use of BPC-157" for it. The committee voted for it anyway, by two votes, with one person refusing to pick a side.

I do not read that as a scandal. I read it as what a room full of qualified people looks like when there is no study on the table. Fifteen experts, same file in front of all of them, and they land almost dead even. If the evidence were actually there, that vote is not 8 to 6. It is lopsided, one way or the other.

They were not arguing about the data. They were arguing because there is barely any.

That is the doctor's whole point and he is correct about it. Most of what is being sold and shipped and injected right now has not been through anything close to what insulin and semaglutide went through.

THE OTHER OTHER SIDE

Where he was wrong

He was wrong about the word.

Peptides are not unresearched. A peptide is a shape, and some of the molecules in that shape are the best documented drugs in the building. Insulin has a hundred years and 209,369 papers on it. It won two Nobel Prizes. It was the first protein ever sequenced and the first recombinant drug ever approved.

What he actually meant is true and it matters: the specific compounds people are ordering off the internet have thin evidence behind them. That is a real warning and people should hear it.

But "peptides are bro science" is not that sentence. It is a bigger, sloppier sentence, and when a doctor says it, a patient hears it as case closed. Then that patient goes and buys the thing anyway, and now they are doing it without telling anybody.

THE GAP

This is the actual problem

On one end you have insulin and the GLP-1s. A hundred years of work. Hundreds of thousands of papers. Trials with tens of thousands of people in them.

On the other end you have a list of compounds with a couple hundred papers each, most of that in animals or in a dish, a handful of small human studies and nothing controlled ever published in full, and a very large number of people using them anyway.

Both of those are peptides.

And for most of that second list the honest answer is not "it works" and it is not "it is junk." The honest answer is that nobody ran the study.

Nobody is going to run it by accident, either. Studies cost money.

WHAT I AM DOING

Peptide Science Alliance

At the end of the day, somebody has to pay for the research.

So I am starting something called Peptide Science Alliance.

The goal is an organization that raises money, sends it to qualified researchers, labs and universities to run real studies on peptides where the evidence is thin, and then publishes what comes back. Either way. If a properly run study shows a peptide does nothing, that gets published. If safety problems turn up, that gets published. A negative result is still an answer, and somebody needed it.

Here is the catch, and I would rather say it out loud than have you find it later. It is early. It is not a nonprofit yet, and I am not taking any money. I do not sell peptides, I do not make them, I will not point you at a vendor, and nothing on this page is medical advice or a dosing recommendation. If you are considering any of this, that conversation belongs with a licensed clinician.

All I am building right now is an interest list. If you want to know when it turns into something real, put your name in at robinheaston.com/peptide-science-alliance. It takes about ten seconds and it costs you nothing.

That doctor and I want the same thing. We want somebody to actually check.

He was right about the gap.

Let's go close it.