TB-500: a seven-amino-acid fragment sold under the name of the protein thymosin beta-4

What is breaking down, and what would have to be built back

Take the two injuries that bring people here.

A tendon is rope. Type I collagen fibres, laid down in parallel, bundled, and anchored into bone. It goes wrong slowly: the fibres lose their parallel order, the cells inside stop laying down good collagen and start laying down disorganised patch material, and tiny vessels grow in where they do not belong while the tissue overall stays starved of blood. That last part is the whole problem. A tendon has very little blood supply, so the cells and oxygen needed to rebuild it arrive slowly, and the repair loses the race against the next load you put through it.

A disc is worse. The soft centre of the disc holds water under pressure; the tough outer ring contains it. With age and load the centre dries out, the ring cracks, the height drops, and the joints above and below start taking force they were not built for. A disc has almost no blood supply at all in its middle — nutrients seep in slowly through the bone above and below it. Nothing in the body repairs more slowly than the middle of a disc. What degenerating discs actually do and what a herniation actually does are covered separately.

So the chain that breaks tissue down looks like this: load exceeds repair, the cells switch from building rope to building patch, blood supply stays too thin to feed either, the patch scars instead of remodelling, and the tissue ends up weaker — so the same load now breaks it faster.

Anything that would build it back has to hit one of five links. This is the frame to hold every claim below against.

The vial and the studies contain different molecules

Two products sit side by side on the same shelf at the same 10 mg strength — one labelled "TB-500", one labelled "TB-500 / Thymosin Beta-4" — and nothing on either label tells you which molecule is inside. This is not a technicality.

Where the fragment came from is documented. In 2003 a group cut the protein apart looking for the smallest piece that still grew new blood vessels, and found it: those seven amino acids, and nothing shorter. Peptides missing the seven were dead. The seven on their own worked at about 50 nanomolar. Adding loose actin at 5 to 50 nanomolar switched the effect off completely, which proves the activity comes from grabbing actin rather than from docking into a receptor.

That is the honest origin story. The fragment is real, it does one real thing, and one thing is not the same as everything the parent protein does.

A regulator wrote the mismatch down in July 2026

The Pharmacy Compounding Advisory Committee briefing package is the only document from a government agency that examines TB-500 as a substance. It says in plain terms that the two are not interchangeable, and it names the marketing error.

U.S. Food and Drug Administration

Official disclosure

FDA Briefing Document, Pharmacy Compounding Advisory Committee — TB-500-Related Bulk Drug Substances (TB-500 free base and TB-500 acetate)

“TB-500 (free base) is understood today to be a seven amino acid synthetic fragment of thymosin beta-4 (β4) from amino acids 17 to 23 and an acetyl group at the N-terminal leucine amino acid (Ac-Leu-Lys-Lys-Thr-Glu-Thr-Gln-OH). … thymosin beta-4 and TB-500 are not the same substance. … no articles were found in which TB-500 was administered to humans. … Therefore, potential safety risks associated with the use of TB-500 in humans are unknown.”

Aug 4, 2026 · fda.gov

2026-08-04 08:16s22 · #181a7b2d744e

Three findings in that package change how everything else reads.

Nobody has been given it and measured. The agency searched the published literature through March 2025 and reports that "no articles were found in which TB-500 was administered to humans." The people who nominated it supplied no clinical data and no record of anyone taking it, by any route. The agency's safety verdict is the word "unknown" — not clean, not dangerous, unknown.

The cap on the front end is not a detail. The wound-healing papers vendors point at used the uncapped seven amino acids, LKKTETQ. FDA states that capping the front end "irreversibly alters their charge, hydrophobicity, and size," and that what the uncapped version does "cannot be directly extrapolated" to TB-500. Change the electrical charge and the water-repelling behaviour of a molecule whose entire job is sticking to another molecule, and you have changed the thing that matters.

The one time anybody tested TB-500 itself in a dish, it did nothing. In a scratch test — you scrape a line across a sheet of skin-building cells and time how fast they close it — TB-500 free base "appeared to be devoid of wound-healing properties." Its breakdown product, capped LKKTE, at 50 micrograms per millilitre did close the scratch by a small but real amount. That leaves open the possibility that TB-500 is a delivery form for a shorter piece. Nobody has tested that either way.

The same package records that the nomination itself carried a molecular formula that was wrong and a CAS number matching neither the free base nor the acetate salt. It also records where TB-500 actually came from commercially: a 2011 veterinary preparation sold to make racehorses and racing greyhounds perform better.

Seven amino acids are enough to pull one lever

The proposed mechanism runs in five steps. None of it is special injury machinery. It is the ordinary business of a cell changing shape, happening where tissue is being rebuilt.

1. Holding loose actin. The protein grabs free actin building blocks, one molecule to one molecule. A held block cannot join a growing strand, so the cell keeps a stock of ready parts instead of a stiff finished mesh.
2. Releasing it on command. Letting those blocks go where and when the cell wants lets it extend a strand in one direction and push. Holding is not blocking. It is stockpiling, and the stockpile is what makes movement in a chosen direction possible.
3. Crawling. A cell that can build a leading edge crawls. Skin cells crossed a filter two to three times faster with as little as 10 picograms of the protein present. The cells that line blood vessels do the same thing.
4. New blood vessels. Those seven amino acids are the piece that does this, and this is the only step where the fragment is the active part rather than a bystander. Vessel-lining cells crawl, sprout, and form tubes.
5. Inflammation settling down. The full 43-amino-acid protein reduces the number of inflammatory cells at a wound and turns down three inflammatory signals — TNF-alpha, interleukin-1β and interleukin-6.

Steps 3 and 4 were measured together in the original rat wound work. Applied to the skin or injected into the belly cavity, the full protein grew the skin surface back over the wound 42% faster at four days and 61% faster at seven, with more collagen and more vessels in the treated wounds than in the salt-water controls.

Step 4 is the step that matters for the tissue you are worried about. Tendon, ligament, the sheet of tissue around muscle and a spinal disc all share one problem: not enough blood. A mechanism aimed at growing vessels is aimed at the actual bottleneck. That is a good argument on paper. It is also where the argument stops, because nobody has grown a vessel into a human tendon with this and measured it.

Why it gets run with BPC-157, and what happened when somebody tested that

BPC-157 and TB-500 are sold as a pair because their proposed mechanisms sit at different points in the same repair sequence.

The theory is that one supplies the machinery for moving and the other supplies the signal telling cells where to move. It is coherent. It has been tested head to head exactly once, and the pairing failed.

Thirty-two male Sprague-Dawley rats, twelve weeks old, about 330 grams each, had the Achilles tendon cut through and repaired, then four weeks of injections into the belly cavity, in four groups of eight: BPC-157 at 10 micrograms per kilogram per day, TB-500 at 60 micrograms per kilogram per day, both together, or nothing.

TB-500 on its own was the arm that reached statistical significance. Maximum load before the tendon failed: p < 0.05. Total Bonar score, which grades how normal the tissue looks under a microscope: p = 0.016. Movin score, a second grading scale: p = 0.017. A collagen stain showed the most orderly type I collagen in the TB-500 group. A protein stain found no real difference between groups in how much type I collagen was being made.

And the combination "did not confer additional benefits compared to either agent alone." The synergy the pair is sold on did not appear in the one experiment that looked for it.

The scar-tissue claim comes off a piece this fragment cannot produce

A large share of the "breaks up scar tissue and adhesions" marketing traces back to Ac-SDKP, a four-amino-acid piece that calms inflammation and reduces scarring, released from the parent protein.

Getting it out takes two enzymes working in order. First an enzyme called meprin-alpha cuts the front section of the protein into pieces under 30 amino acids long. Then a second enzyme, prolyl oligopeptidase, snips Ac-SDKP off. Neither enzyme can do it alone.

Now put the two facts side by side. Ac-SDKP comes off the front end of the protein. TB-500 is positions 17 to 23 — the middle. The four-amino-acid piece cannot be made from it at any dose, by any enzyme, in any tissue. Every scar-reduction claim built on Ac-SDKP belongs to the 43-amino-acid protein and does not carry over to what is in the syringe.

There is a separate scar result for the full protein that does not depend on Ac-SDKP, and it is a good one. In rats, 100 micrograms of the protein delivered on days 2, 3 and 4 through implanted sponges produced wounds at day 14 with very few myofibroblasts — the contracting cells that pull a wound closed into a scar — and very few of their internal tension fibres. The collagen bundles were longer and thicker, and under polarised light they glowed the yellow-red of mature collagen where the untreated wounds glowed the green of immature patch material. Stopping cells from turning into myofibroblasts is the mechanism behind "repair instead of scar."

The 100 micrograms went into the wound on a sponge. It was the full protein. Neither of those matches a milligram injected into belly fat.

Sort the whole literature by tissue and the tendon row is one rat study

Somebody counted the entire field. A 2026 scoping review from an orthopaedics and rehabilitation group at the University of Utah screened 1,772 records and included 80 studies.

Their findings, in their own words: the evidence is "weighted toward mixed and in vitro designs, and most studies evaluated TB4 rather than TB-500"; "direct musculoskeletal tissue categories such as tendon, ligament, muscle, cartilage, and spine/intervertebral disc were comparatively sparse"; and "human evidence was concentrated in ocular/cornea and wound/skin/soft tissue settings, whereas direct TB-500 evidence was limited to a single included study."

One study out of eighty tested the molecule that is actually sold.

Read the "Molecule" column before you read the "Result" column. Ten of the eleven rows are the 43-amino-acid protein.

The best tendon result that exists is eight rats an arm, injected into the belly

That sentence deserves its own heading, because it is the single strongest piece of evidence in the world for the thing people buy TB-500 to do.

It ran four weeks. It was a rodent. The dose went into the abdominal cavity, not under the skin. The authors call their own study "exploratory" and stop at calling both compounds "candidate adjuncts to tendon repair, pending dose-optimization and longer-term studies." Nothing in it supports a milligram-per-week injection under the skin of a person, and the authors do not claim it does.

The delivery problem is out in the open in the tendon-engineering literature, which is itself a signal: the people who study getting this peptide into a tendon do not expect an injection into belly fat to do it. When the full protein was loaded into spun polymer yarns it released steadily over 28 days and pushed human stem cells taken from fat toward becoming tendon cells — they crawled more, multiplied more, and switched on tendon genes together.

The scaffold is the point. Somebody built a 28-day slow-release fibre to hold the peptide at the tendon, because holding it there is the hard part.

The ligament study worked because the peptide was put in the wound

The rat knee-ligament study gets summarised as "TB-500 heals ligaments." What it actually did was place 1 microgram of the full-length protein, carried in 100 microlitres of fibrin glue, directly into a ligament that had just been cut through. Local. Held in place by a carrier. One microgram. A different molecule.

The paper opens by noting that as of 2013 nobody had ever published on this protein in ligament repair at all.

In muscle, your body already sends this signal after an injury

The muscle result is not a treatment result, and it is worth understanding why.

Injuring a muscle causes the muscle itself to make more thymosin beta-4 — in the fibres that are regenerating, and in the blood-cell-making cells that move into the damage. The protein then acts as a chemical attractant, pulling muscle precursor cells toward the injury. Both the ordinary protein and its oxidised form sped up closure and pulled cultured muscle cells across a gap. Precursor cells taken from adult muscle followed the oxidised form.

So the pathway is real in skeletal muscle, and your body switches it on by itself when you tear something. What has never been tested is whether adding more from outside improves a strain. No study has asked that question.

Every controlled result in a person is an eye or a skin wound

Here is the entire human record for the 43-amino-acid protein, sorted by tissue. None of it is muscle, tendon, ligament or disc.

Eyes — the only place the numbers turned positive. A nine-patient dry-eye trial reported eye discomfort down 35.1% and surface damage staining down 59.1% against the vehicle drops.

A Phase III trial in nerve-damaged cornea randomised ten patients against eight. Complete healing at four weeks was 60% versus 12.5%, p = 0.0656 — which misses the usual cutoff for calling a result real. At day 43 it was 50% versus 0%, p = 0.0359, which clears it.

Three completed Phase 3 dry-eye trials sit behind those: ARISE-1 with 317 patients, ARISE-2 with 601, and ARISE-3 with 700, the last finishing in October 2021. Total: 1,618 people. There is still no FDA approval.

Skin — completed, but safety was the question being asked. The pressure-ulcer trial enrolled 72 patients on gel at 0.01%, 0.02% and 0.1% by weight, once daily for up to 84 days. The main question was safety and tolerance. Healing was a secondary question.

A matching 72-patient trial in leg ulcers from poor vein drainage completed on the same design.

Two more were stopped early: one in the blistering skin disease epidermolysis bullosa, 30 patients, and one in corneal wounds in diabetic patients after eye surgery, 12 patients.

Heart — withdrawn twice, then restarted in another country. The animal finding that started the programme is specific: the protein forms a complex with two other proteins, PINCH and integrin-linked kinase, which switches on the survival signal Akt. After a coronary artery was tied off in mice, heart muscle cells survived better and the heart pumped better.

The Phase 1 safety study in healthy volunteers was designed around single rising intravenous doses of 42 mg, 140 mg, 420 mg and 1,260 mg. It never enrolled anybody. The reason on the record: "Study never initiated due to contract manufacturing issues."

The Phase 2 heart-attack trial behind it was also withdrawn with zero patients enrolled. The only heart programme still producing human data is a Chinese full-length recombinant product called NL005: a Phase 1a in 54 healthy volunteers that established the highest tolerated dose, measured how fast the body cleared it, and checked whether the immune system made antibodies against it, then a Phase 1b in 30 people and two completed Phase 2 trials in heart attack.

Look at the dose scale in that withdrawn Phase 1 design. Milligram quantities of the full protein, dripped into a vein, in a hospital, under a protocol, are not the same object as 2.5 mg of a seven-amino-acid fragment injected under the skin of a belly at home — even though both are written in milligrams.

One registry record names TB-500, and it says it is made up

Search ClinicalTrials.gov for TB-500 and you get exactly one interventional record where the intervention is TB-500 itself: NCT07487363, "TBRIDGE-CV," sponsor listed as Hudson Biotech, first posted March 2026.

Its summary reads, word for word: "This fictional study is an example of a ClinicalTrials.gov-style record."

It is a demonstration template. Its intervention description has an unclosed bracket. Its dose field says the levels "are not provided in this public example." Any page that cites NCT07487363 as proof TB-500 is in human trials is citing a placeholder — and that citation will keep appearing, because it is the only registry entry that looks like the thing people want to find.

Nothing has ever been injected into a person and measured

One row, stated plainly, because the tables above allow no other reading.

There is no human trial of TB-500. None finished, none running, none registered. Every controlled human result belongs to the 43-amino-acid protein, given as eye drops or a skin gel, for dry eye, a nerve-damaged cornea, or a long-standing skin ulcer. For tendon, ligament, muscle, fascia, cartilage or disc, the count of human trials is zero for both molecules.

Asked whether it is proven for a tendon: no. Not partly proven, not proven in a small way. The study has never been run.

That is one row on this page, not the page. Everything else here is what is actually known.

Mixing the vial, in numbers a syringe can read

A freeze-dried vial holds a fixed mass of powder. The strength is whatever the water you add makes it, and the volume you inject follows from that. Insulin syringes are marked in units, where 100 units is 1 millilitre, so one unit is 0.01 mL.

Take the common 10 mg vial and add 2 mL of bacteriostatic water:

• 10 mg ÷ 2 mL = 5 mg per mL
• 1 mL is 100 units, so 100 units holds 5 mg
• 1 unit = 0.05 mg = 50 micrograms

Add 5 mL instead and you get 2 mg/mL. Now 1 unit is 20 micrograms, and a 2.5 mg dose is 125 units — more than one syringe holds, so a single dose takes two injections. That is why 2 mL is the practical fill for a 10 mg vial: it puts every commonly used dose inside one syringe.

The mechanics, in order. Wipe the rubber stopper with alcohol and let it dry. Draw up the water. Angle the needle so the water runs down the inside glass wall instead of jetting straight onto the powder. Swirl until it goes clear. Do not shake. Shaking tears peptide molecules apart and makes them clump together, and clumping is the exact mechanism behind the immune-reaction concern further down this page.

Every dose ever given in a study, and where yours would sit

No dose-finding study of TB-500 exists. Not in any species, at any dose. The loading-then-maintenance schedule that circulates online has no source under it: no trial produced it, no blood-level study calibrated it, no regulatory filing contains it. The nomination sent to FDA specified a formulation — 3 mg/mL, "Subcutaneous and/or Intramuscular Injection" — and specified no dose at all.

What does exist is every dose that has actually been given in a study, and the list is short enough to print in full.

Read the route column and count. Two doses in the entire record went under the skin: one of them was a horse, and the other does not exist. Every human dose ever given was put on a surface — a cornea or a wound bed — and none of them was TB-500.

The one animal musculoskeletal study that came out positive on placement put 1 microgram directly into the wound in a glue carrier. The schedules people run use milligrams injected into belly fat and expect them to reach a tendon in the shoulder. Those are not the same treatment scaled up. They are different treatments.

Three route facts are settled. There is no oral form that works, because your gut digests it. Under the skin and into the muscle are the routes the formulation was submitted for. Injecting into or right beside the injured structure — the route in the one positive animal result — is not the route almost anybody uses.

Ten milligrams under the skin of a horse peaked at 0.08 nanograms per millilitre

The only measurement of TB-500 in the blood of a living animal, in any species, comes from thoroughbred geldings given 10 mg of the free base under the skin.

Blood concentration peaked at 0.05 to 0.08 nanograms per millilitre, somewhere between 60 and 120 minutes after the injection. Between 6 and 10 hours it could no longer be measured at all. It broke down by losing amino acids one at a time off the tail end: Ac-LKKTET, then LKKTE, then LKKT, then LKK, then LK.

Two things follow from that, and both change how you would take it.

Dosing every three or four days has nothing behind it. Schedules that space injections out are usually defended by saying TB-500 has a long half-life. A compound that cannot be detected in blood 6 to 10 hours after a 10 mg dose does not have a long half-life. Whatever else might justify spacing doses out, measured persistence in the blood is not it — and no half-life for TB-500, in any species, has ever been published as a number.

Sub-nanogram blood levels off a 10 mg dose is a very small exposure. For scale: the concentrations at which this peptide and its breakdown products do anything measurable in a dish are in the tens of micrograms per millilitre. That is four to five orders of magnitude higher than what a 10 mg injection produced in a horse.

The 28-day rule comes from the water, not from the peptide

• Freeze-dried powder: keep at 2–8 °C, or freeze it for long storage. Keep it out of light.
• Mixed with bacteriostatic water, which contains 0.9% benzyl alcohol as a preservative: 2–8 °C, commonly cited as good for about 28 days. That 28 days is the preservative's specification. It is not a stability figure anybody measured on this peptide.
• Mixed with sterile water, no preservative: one use, discard after a single draw.
• Do not freeze and thaw it once it is in solution.

There are no pharmacopeial stability data for either form of TB-500. FDA's own finding is that TB-500 free base "is not physically and chemically well characterized," and that tests for contaminants, clumping, microbial contamination and bacterial endotoxin are missing from the public data and from vendor Certificates of Analysis.

How long each measured effect actually took to appear

Every timeline anybody can quote belongs to a study, and they cluster tightly.

The musculoskeletal readouts all sit at four weeks in rodents. Rat tendon repairs faster than human tendon, and the human version of that four-week endpoint has never been measured. Any statement about how long you should expect to wait is a rat's timeline stretched across a species gap, for a molecule the rat got into its abdominal cavity.

There is a second problem with judging this by how you feel. Tendon problems get better on their own over months. That is the ordinary course, with no treatment at all. If you improve on that timescale you cannot separate the compound from time, from having backed off the load, and from everything else that changed in the same window.

The cancer objection has real numbers behind it

The vessel-growing and cell-crawling machinery that would plausibly help a healing tendon is the same machinery a tumour uses to grow and spread. That is not a hypothetical. It has been measured in living animals.

Forcing a melanoma cell line to overproduce thymosin beta-4 gave a mean 2.3-fold increase in how far the cells crawled (95% CI 1.9–2.7, p<.001), a mean 4.4-fold increase in blood vessels inside solid tumours (95% CI 3.3–5.5, p<.001), and roughly four times as many tumour deposits in the lungs — 46.7 against 10.9. The authors concluded that thymosin beta-4 may drive tumour spread by switching on cell crawling and new vessel growth.

The paper shares an author with the wound-healing and vessel-growth work. The helpful finding and the harmful finding come from the same laboratory and the same mechanism. That is not a scandal. It is what an honest mechanism looks like when you follow it in both directions.

Three qualifications, all of them real.

• The experiment used a virus to force a melanoma line to overproduce the protein. It did not give a seven-amino-acid fragment at the doses a person injects. It establishes that the risk is mechanically plausible and open. It does not establish a proven human hazard.
• The literature runs both ways and depends on the cancer type. In multiple myeloma, this protein behaves as a tumour suppressor, and lower levels of it predict a worse outcome.

• No cancer-causation study of TB-500, in either form, has ever been run. FDA states that outright.

The defensible position is a specific exclusion, not a ban on the compound. Active cancer, or a significant cancer history, is a reason not to use this — because the risk is unmeasured in the direction that would hurt you and the benefit is unproven. That is a bad trade regardless of how the mechanism eventually resolves.

The wider literature supports that framing rather than either extreme. The protein is reported to help liver tumours in children spread, and to drive colon cancer cells to crawl through a specific signalling route, while in multiple myeloma losing it predicts a worse course. The effect depends on the tumour, which is exactly why no blanket reassurance is available in either direction.

What is in the vial has never been tested for the things that could hurt you

Separate from cancer, and specific to injecting a substance nobody has characterised:

The zero in that last row is not reassurance, and reading it as reassurance is the mistake this page most wants you to avoid. There is no legal human market for TB-500, so there is no channel through which a report could be filed. FDA's food and cosmetics system held two cases mentioning "blended TB-500 and BPC-157," with no safety assessment attached to either. No surveillance produces no data, and no data looks identical to safety if you do not check which one you are looking at.

Nobody has published an independent purity or endotoxin audit of the consumer TB-500 market. Searching for one returns vendor pages claiming 99% purity and nothing that tested the claim. FDA arrived at the same place from the paperwork side: the nomination package supplied a molecular formula that was wrong and a CAS number matching neither form of the substance. That is what an uncharacterised material looks like on paper.

The only safety information with any weight behind it belongs to the full-length protein, and even that is thin at whole-body doses. The one programme that has produced human immune-response and blood-clearance data for the protein given systemically is the Chinese recombinant product, whose Phase 1a in 54 healthy volunteers measured the highest tolerated dose alongside peak concentration, time to peak, total exposure, half-life, clearance rate and antibody formation.

Those numbers describe a 43-amino-acid recombinant protein dripped into a vein under a hospital protocol, and they are not published in the searchable literature. They do not describe the fragment, the route, or the setting in which TB-500 is actually used.

Banned in sport at all times, and the wording names tendons

The 2026 Prohibited List, in force from 1 January 2026, lists under section S2.3 — Growth Factors and Growth Factor Modulators — "Thymosin-ß4 and its derivatives e.g. TB-500."

World Anti-Doping Agency

Official disclosure

World Anti-Doping Code International Standard — Prohibited List 2026, section S2.3

“Thymosin-ß4 and its derivatives e.g. TB-500 … and other growth factors or growth factor modulators affecting muscle, tendon or ligament protein synthesis/degradation, vascularisation, energy utilization, regenerative capacity or fibre type switching. … PROHIBITED AT ALL TIMES (IN- AND OUT-OF-COMPETITION) … All prohibited substances in this class are non-Specified Substances.”

Aug 4, 2026 · wada-ama.org

2026-08-04 08:16s23 · #b92aa31e3f15

Four details decide what that means for you.

• Prohibited at all times, in competition and out of it. Using it in the off-season is still a violation.
• It is a Non-Specified Substance — the harsher sanction tier, with no starting presumption that the penalty should be reduced.
• The catch-all clause covers "other growth factors or growth factor modulators affecting muscle, tendon or ligament protein synthesis/degradation, vascularisation, energy utilization, regenerative capacity or fibre type switching." The exact reason you would take it is the exact behaviour the clause prohibits, written out in the rule.
• "Other substances with similar chemical structure or similar biological effect(s)" are prohibited too, which sweeps in unnamed fragments and lookalikes.

Enforcement is real, and it has hit the pairing specifically. The Canadian Centre for Ethics in Sport imposed a four-year ban on an athlete for BPC-157 and TB-500. The United States Department of Defense has adopted WADA categories S0 through S5, which puts service members under the same restriction.

Sanctions do not require a positive test. They can be imposed on an admission, on possession, or on a purchase record. If you are in a college programme, a national federation, an Olympic pathway, most professional leagues, or the military, that exposure exists whether or not anybody ever tests you.

The July 2026 vote started a process, it did not change the status

On 23–24 July 2026 the Pharmacy Compounding Advisory Committee voted 8 to 6, with one abstention, in favour of letting BPC-157, KPV and TB-500 be compounded — against FDA's own written recommendation that "a balancing of the criteria weighs against" listing them.

Pharmaceutical Executive

Official disclosure

FDA Panel Votes to Loosen Restrictions for Four Peptides

“voted 8-6 in favor of allowing Bpc-157, Kpv and Tb-500 to be compounded, with one abstention … The votes do not make the peptides FDA-approved drugs … FDA is not bound by the panel's recommendations … The advisory committee's recommendation still triggers a formal rulemaking cycle that realistically runs eight to twelve months”

Aug 4, 2026 · pharmexec.com

2026-08-04 08:16s24 · #9d6badacf186

What that vote does and does not do:

• It does not make TB-500 an approved drug.
• FDA is not bound by the recommendation.
• It starts a formal rulemaking cycle that realistically runs eight to twelve months.
• As of this writing TB-500 is not on the 503A Bulks List and is not legally compoundable. No outsourcing facility has reported compounding it, and no pharmacy was found doing so.

"The FDA just approved it" is wrong twice: there is no approval, and the rulemaking has not finished.

Thirty-one people wrote down what happened to them, and the split is not what the marketing says

Everything above this line is evidence. What follows is not, and it is counted anyway, because it is the only record of people using this for the reason you are reading about it.

Thirty-one first-person accounts, posted publicly on X between January and August 2026. Each one was located by search and then re-checked independently by pulling the post back and comparing the full text and the timestamp. Three candidates failed that check and are not counted. Nobody in this set was randomised, blinded, or controlled. Nobody tested the powder before injecting it, so the actual substance is unknown in every case. People who improved are far more likely to post than people who did not, so the set is selected before you even read it.

The count: 17 said it helped. 8 said nothing happened. 6 had a reaction or got worse. Denominator 31. Two of those 31 were treating a gut complaint rather than a musculoskeletal one. Drop those two and the split is 17 helped, 7 nothing, 5 worse, out of 29.

There is one number that matters more than the split, and it is at the bottom of this section.

Before the accounts themselves, the shape of them. The seventeen that helped cluster in shoulders and elbows run for four to eight weeks; the eight that reported nothing and the six that reported a reaction or a worsening are grouped under their own headings below, so you can read one group without reading all thirty-one. If you came here about a specific injury, the section titled injury by injury, what the record actually reaches is the one to read instead of these accounts: it sorts every tissue by what evidence exists in it, and for four of the seven rows the honest entry is none. No body-region tally is given for the thirty-one because the set is too small to divide without producing percentages that look like measurements, which is the failure this page spends its first section warning about.

Every account below is quoted in the person's own words, with the handle, the date and a link to the original post. All of it is anecdotal. None of it is evidence.

Seventeen said it helped

The strongest reports come from shoulders and elbows, run for four to eight weeks.

"Ran BPC 500mcg twice a day and 6mg of TB500 every 5th day for 6-8 weeks. Healed my torn rotator cuff" — @PayHereSukka, X, 30 July 2026. Anecdotal.

"Slight tear of my rotator cuff. Surgery was a 5 month rehab timeline, instead I took bpc-157 and tb-500. Was back to full rotation in my shoulder and benching good weight by a month and a half!" — @GrantWest14, X, 10 January 2026. Anecdotal.

"I was dealing with a nagging golfers elbow for over a year. Would be tender the to the touch after hitting balls. Within a week of taking shots the pain started to disappear. Months later its compmetely healed." — @ryalljohnny1989, X, 3 August 2026. Anecdotal.

"I took it alongside TB-500 and it fixed my tennis elbow and was back to lifting also within two week." — @Das_Bear_, X, 29 June 2026. Anecdotal.

Three of the seventeen were treating exactly the disc problem that may have brought you here, and all three describe substantial relief inside a month.

"I've been doing subcutaneous BPC TB+500 injections on both sides of the L5-S1 and a month in, the pain is dramatically less... the pain has been horrible for 6 years but this stuff really works. I don't wake up in awful pain anymore!" — @MithMuel, X, 30 June 2026. Anecdotal.

"Blew up my back in 2017. L4/L5/S1 severe disc herniation. Doc diagnosed me with degenerative disc disease, told me to quit lifting... Found the Wolverine stack and within a month I was pain free." — @jasonsvoboda, X, 26 July 2026. Anecdotal.

"My lowest disc in the lower back is dried out, almost no fluid left, classic herniated disc... In June I ran BPC-157 and TB-500 for about 6 weeks... After 3-4 weeks it already felt better. Now after the cycle, zero pain in biceps and shoulder. Disc still needs physio and rest, but already much better." — @MuroCrypto, X, 31 July 2026. Anecdotal.

Read that last one closely, because it is the most careful account in the whole set. He separates his shoulder result from his disc result, says the disc "still needs physio and rest," and adds "Don't know if it works the same for everyone, I'm no expert." That is what an honest anecdote looks like.

Some report the effect arriving far faster than any tissue can rebuild — hours to days. That is a timescale for inflammation and pain signalling, not for collagen.

"Wednesday I woke up in severe pain in my L Leg. It was so bad I couldn't bear any weight... Peptides started in the afternoon. Leg 75% better in the evening... Saturday: Woke up 100% better." — @L0V3lsKeY, X, 2 August 2026. Anecdotal.

"BPC-157 + TB-500 vastly accelerated my recovery timeline, eliminated my knee pain two days after i started, and brought the swelling in my jaw down within the week" — @howdymary, X, 9 March 2026. Anecdotal.

Two days is not tissue repair. Whatever changed there, it was not new collagen.

The remainder are knees, ankles, quads, and general joint pain.

"4 weeks 1mg bpc and then 4 weeks 2mg in knee. No acl or meniscus...knee has never felt better since before the injury years ago...also did tb500 twice a week total 4mg" — @caezcrypto1, X, 30 July 2026. Anecdotal.

"I'm 95% sure that BPC-157 and TB-500 helped me heal from several injuries including my left knee and right ankle. I really thought I was going to need surgery for both." — @JonahLupton, X, 26 July 2026. Anecdotal.

"I tore both quad tendons in separate years, the first recovery had taken almost a year. After the second surgery... I added BPC-157 and TB-500. By month five, I was sprinting and jumping again" — @pepfessions, X, 3 August 2026. Anecdotal.

"After a week of BPC-157 (8 injections, 500mcg subcutaneous abdominal), persistent ankle tendon pain from a snowboarding injury a year ago is barely noticeable... Yesterday I added TB-500 (1 injection, 2mg) as well" — @jordaaash, X, 30 March 2026. Anecdotal.

"Have done a BLAST OF TB 500 and BPC -157 and KPV! Have serious shoulder tear! Blasted all three for 2 weeks!... Took my shoulder from barely joking a cupping hand to back in gym!" — @tj_underwood, X, 31 July 2026. Anecdotal.

"My 70 yr old dad tore rotator cuff and I had him do 1mg (BPC) injected daily in 2 doses for 4 months with 2-3x/weekly TB500 @ 2.5mg/dose. Within a few weeks, he was lifting his arm totally above his head again and all pain was gone." — @_REDwave_2024, X, 20 May 2026. Anecdotal, and second-hand — the son is reporting his father's result.

"Promise you peptides are No Joke, I used the wolverine stack (BPC-157 and TB-500) to alleviate lifelong shoulder and joint pains" — @BenWolf1425771, X, 3 August 2026. Anecdotal, and the same post links to a supplier, so treat it as promotional.

Two of these seventeen carry their own warning on their face. One reports the benefit disappearing about six weeks after stopping — which is the opposite of what tissue repair does. Repaired tendon does not un-repair when you stop injecting.

"All of the soreness, aches, and pains that disappeared while on BPC/TB500 returned after about 6 weeks for me. I start back on Monday" — @TheCaddieYardFL, X, 1 August 2026. Anecdotal.

The other is @jordaaash above. He added TB-500 the day before he posted, after a week on BPC-157 alone, and the improvement he describes had already happened before the TB-500 went in. It is counted as a positive because the person counted it that way. It attributes nothing to TB-500.

Eight said nothing happened

These are not softer versions of the positives. They are people who ran it for weeks at ordinary doses and got no change at all.

"Did 3mg of BCP & 5mg of TB500 daily for two weeks into my elbow. It didn't do jack shit. This is my 3rd time trying peptides for various joint/tendon stuff and my 3rd time not noticing any improvement at all." — @prax1s_mm, X, 30 July 2026. Anecdotal.

"Bpc and tb-500 did nothing for my elbow arthritis sadly. Used it for 6 weeks" — @replacementBux, X, 12 April 2026. Anecdotal.

"Age 62... I am very limited on what I can do at this point due to the severity of pain. I have been doing BPC 157 +TB 500 for 7 weeks, no improvement yet." — @ToddMil03223111, X, 9 July 2026. Anecdotal.

"I ran a heavy dose of klow into a heavy dose of bp-157 and tb500 for 2 months or more for my shoulder, did nothing, facing a shoulder replacement" — @moon25martin, X, 30 May 2026. Anecdotal.

That last one is the harshest data point in the set: two months at heavy doses, no effect, still headed for a shoulder replacement.

Somebody treating the same lower-back problem as the three positive disc reports got nothing at all.

"I'm trying it to see if it helps at all with inflammation for my lower back. L4/5 bulge and L5/S1 herniation. Bpc and Tb did nothing I could notice." — @HangryDBowman, X, 16 July 2026. Anecdotal.

Three people saying a disc improved and one saying nothing changed, all treating an L5/S1 problem, none of them scanned before or after — that is a clean picture of exactly what uncontrolled reports can and cannot tell you.

Two more report the effect as too small to matter rather than absent.

"I personally used bpc 157/ tb 500 combo for my shoulder issue. Barely moved the needle. 3/10." — @travfourr, X, 15 April 2026. Anecdotal.

"I had some pain there and leading up to my bicep area. Didn't get much relief from bpc157/tb500." — @MittelstaedtTim, X, 11 June 2026. Anecdotal.

One took an oral form for a gut complaint and got nothing, which is unsurprising given the peptide does not survive digestion.

"Just finished a bottle of oral liposomal BPC157+TB500 and it seems it did nothing :/ (hoping it would help heal my gut)" — @Gronnet, X, 26 April 2026. Anecdotal.

Six had a reaction or got worse

These are the reports vendor pages do not carry, and the most informative account in the whole set is the first one.

"TB-500 made my shoulder left worse. I have bursitis and stopped TB and it's better. I was getting pain in left thumb and arm as well. I was pinning SubQ. I've used 3 different vendors who are legit. So I ruled out bad product. No issues with BPC-157." — @biotides, X, 7 July 2026. Anecdotal.

He had bursitis, the shoulder got worse on TB-500, it improved when he stopped, he had used three different suppliers so a bad batch is unlikely, and he had no problem with BPC-157. That is about as close as an uncontrolled self-report ever gets to isolating one variable.

"i ran bpc 157 for few weeks felt decent then add tb 500 hand and skeletal felt worse i discontinued it it was the blend version 20mg" — @cryptotray79, X, 14 June 2026. Anecdotal.

"I ran TB-500 for a shoulder injury and got a head rush that lasted maybe 20 minutes. Injection site was red for a day." — @UndergroundBio, X, 3 August 2026. Anecdotal.

Three describe immune-type reactions — hives, itching. That is the exact category FDA flagged in advance, on the grounds that clumped peptide and manufacturing contaminants provoke it.

"Not my proudest moment, but after BPC-157 and TB-500 from the gray market, I got hives all over my body." — @stateofnomind, X, 10 March 2026. Anecdotal.

"I took tb500 and bpc 157 mix for 3 days at 0.750 7.5 units on a 3mm needle i had some Hives show up on each side of love handles not the injection sites... Ive done research my self and stopped for now" — @GetterBankWest, X, 2 March 2026. Anecdotal.

"nothing moved the needle more re gut barrier symptoms than TB500. Until the development of an MCAS-like picture, it began to flair immune Sx: itching, hives from heat, etc." — @AmbrosesDrink, X, 4 August 2026. Anecdotal.

Two of the thirty-one used TB-500 without BPC-157, and both are in the last group

This is the number that matters. Twenty-nine of the thirty-one accounts ran TB-500 stacked with BPC-157, and several stacked a third and a fourth compound on top of that. Not one of those twenty-nine can attribute anything to TB-500. Every positive report in this set is a report about a combination.

The two accounts that isolate TB-500 are @biotides, whose bursitis shoulder got worse and recovered when he stopped, and @UndergroundBio, who got a head rush and a red injection site. Both sit in the "worse" column.

That is not evidence that TB-500 is harmful. Two reports settle nothing. It is a precise statement of how thin the isolated human experience with this compound actually is: out of 31 people who wrote down what happened, 2 took it on its own, and neither described a benefit.

What these 31 reports can and cannot settle

They can settle that people are running this at milligram doses for weeks, mostly stacked, mostly for shoulders, elbows, knees and lower backs, and that a substantial fraction of them believe it worked.

They cannot settle whether it works. Nobody was blinded. Nobody imaged the injury before and after. Almost everybody changed something else in the same window — rest, physiotherapy, load reduction, a second compound, surgery. The conditions being treated improve on their own over the same timescale. And negative reports are systematically harder to find than positive ones, which means 17 out of 31 overstates the true positive rate by an unknown amount.

Hold this section against the evidence sections above and the shape is clear. The self-reported record is large, warm, and confounded. The controlled record for this exact molecule, in this exact tissue, is a single study in eight rats.

Injury by injury, what the record actually reaches

Two of those rows carry an extra warning. The frozen-shoulder and disc arguments both rest on reducing scar, and the strongest scar-reduction mechanism in this literature runs through Ac-SDKP — the piece the sold fragment cannot make.

Every human study ever registered on this molecule

Here is the whole clinical programme for the parent protein, with what actually happened to each trial. TB-500 appears in it once, as the fictional example.

Four things fall out of that table.

No musculoskeletal indication has ever been registered. Not tendon, not ligament, not muscle, not fascia, not disc. Not by any sponsor, in twenty years, for either molecule. That absence is not a hole in the search. It is the state of the field.

Both heart trials from the original sponsor were withdrawn before a single patient enrolled, and the programme then moved to a different sponsor, a different country, and a recombinant full-length product.

The published "Phase III" cornea result comes from a terminated trial. SEER-1 stopped at 18 patients enrolled, and the ten-versus-eight comparison you read in the literature is that terminated trial's data. Its successor, SEER-2, is still recruiting toward 70.

Three completed Phase 3 dry-eye trials covering 1,618 patients have produced no approval. ARISE-1 finished in 2016, ARISE-2 in 2018, ARISE-3 in October 2021. That is the largest and most mature dataset this molecule has, in the indication where it performs best, and it has not turned into a medicine anybody can buy.

The skin programme closed the same way — two trials completed with safety as the main question, two stopped early.

What is solid, what is not, and the gap underneath both

Solid. Thymosin beta-4 holds loose actin and controls how a cell builds its internal skeleton. Its seven-amino-acid actin-grabbing piece is the part that grows new blood vessels, working at around 50 nanomolar. The full protein closes rat skin wounds 42 to 61% faster, stops repair cells turning into the scar-pulling kind, works through integrin-linked kinase and Akt in injured heart muscle, pulls muscle precursor cells toward damage, and improves dry eye and nerve-damaged cornea in small controlled human trials.

Not solid. That any of that heals a human tendon, ligament, muscle, disc or fascia. That the fragment sold as TB-500 does what the full protein does. That milligram-per-week injections under the skin reach an injured structure at a concentration where the mechanism runs. That the compound is safe in a person over any length of time, because that study has never been run.

The gap underneath both. The evidence and the product are not the same substance. A regulator has written that down. The capped form has never been shown to do what the uncapped form does. And the one time anybody tested TB-500 itself in a dish, it did nothing. Every transfer of evidence from the 43-amino-acid protein to the seven-amino-acid fragment is an assumption, and it is the assumption the whole market stands on.

The practical consequence is narrow and specific. Tendon pain from overuse, an impingement, and a partial tear are three different problems with three different treatment plans, and no peptide tells them apart. Getting the diagnosis right changes what you should do this week far more than anything in this vial does. The scoping review that counted this entire literature lands in the same place, in its own words: the evidence "remains unevenly distributed and largely preclinical, with limited human evidence directly relevant to musculoskeletal applications."

Not medical advice. TB-500 is sold for research use only, is not an approved drug, is not currently legal to compound, and is prohibited at all times in WADA-tested sport. Nothing here is a dosing or treatment recommendation.

The sibling objects for this page, each one inspectable on its own terms: