TB-500: A Seven-Residue Fragment Carrying a Larger Protein's Reputation

The short version

TB-500 is a short synthetic peptide, seven amino acids long. Those seven residues are a slice cut out of a much bigger natural protein called thymosin beta-4 — specifically the piece that grips actin, the structural fibre cells use to change shape and move.

Here is the problem that runs through everything else on this page. When people cite research to explain what TB-500 does, they are usually citing research on the whole protein, which is about five and a half times larger. Whether the seven-residue piece reproduces what the whole protein does has not been established in a controlled human trial. Nobody has shown that it does. Nobody has shown that it does not.

What there is: cell-biology work explaining how the parent protein binds actin ^[12], rat injury studies ^[9], a review of the parent protein's repair activity ^[10], and one safety trial in 40 healthy volunteers — again with the full-length protein ^[11].

On bone and cartilage specifically: nothing. No study cited here measured either.

What it is

TB-500 is a synthetic, N-terminally acetylated heptapeptide, sequence Ac-Leu-Lys-Lys-Thr-Glu-Thr-Gln-OH. It corresponds to residues 17–23 of the endogenous 43-amino-acid protein thymosin beta-4 (Tβ4, gene TMSB4X). That LKKTETQ stretch is the conserved actin-binding region shared across the beta-thymosins.

The naming is where the confusion begins. In commerce and in the analytical and anti-doping literature, "TB-500" means the heptapeptide, roughly 889 Da. In the efficacy literature, the molecule under test is overwhelmingly full-length recombinant or synthetic Tβ4, roughly 4963 Da. Those are different molecules with a shared motif, and this file flags which one was used every time a finding appears.

Regulatory position: TB-500 has no approved therapeutic indication anywhere and is not approved by the FDA for human use. The FDA has reviewed peptide bulk drug substances including thymosin β4-related peptides for compounding and flagged safety concerns. It is prohibited by the World Anti-Doping Agency under prohibited peptide and growth-factor categories, banned in and out of competition for the relevant classes, and anti-doping laboratories have built LC-MS methods that detect it in both equine and human matrices. It is classified as a prescription medicine in some jurisdictions, including Australia and New Zealand, and it has been encountered as a designer drug in racehorses. Suppliers sell it for laboratory research use only.

How it works

The mechanism is cytoskeletal rather than vascular. Full-length thymosin beta-4 is the major intracellular sequestering agent for monomeric, globular actin: X-ray crystallography of a gelsolin-domain-1–Tβ4 hybrid bound to actin, solved at 2 Å, established that Tβ4 forms a 1:1 complex with G-actin and holds the monomer by capping both of its ends, preventing polymerisation, with the WH2 motif underlying the interaction ^[12]. Functionally, that buffers a pool of unpolymerised actin and regulates how quickly a cell can rebuild its cytoskeleton — which is what a cell must do in order to migrate into a wound.

A consolidating review describes the downstream picture for the parent protein: actin binding drives cell mobilisation and migration and stem-cell activity; myofibroblast numbers fall, which reduces scar formation; platelets and macrophages release Tβ4 at injury sites where it limits apoptosis, inflammation and microbial growth; and it promotes angiogenesis. That combination is the stated rationale for clinical development in dermal wounds, corneal injury, and cardiac and central-nervous-system repair ^[10].

The mechanism targets listed for this compound are monomeric actin, vascular endothelial cells, keratinocytes and corneal and dermal epithelium, hair-follicle bulge stem cells, and cardiomyocytes and epicardial progenitor cells. Three of those five are attributed specifically to the full-length protein. None of them is a bone or cartilage cell type.

Whether the isolated seven-mer reproduces any of this at the exposures used in peptide research is not established in controlled human trials.

What the research shows

Each result below is labelled with the species and, critically, with which molecule was tested.

Structural biochemistry, in vitro. A 2 Å crystal structure established 1:1 G-actin sequestration by dual-end capping and the WH2 structural basis for it ^[12]. This is the strongest, cleanest result in the whole TB-500 literature — and it is a description of a molecular interaction, not a healing outcome.

Embolic stroke, male Wistar rats, full-length Tβ4. Intraperitoneal thymosin beta-4 given at 2, 12 or 18 mg/kg starting 24 hours after stroke and then every three days for four further doses improved neurological function at 2 and 12 mg/kg, significantly from day 14 through day 56 (p<0.05). The 18 mg/kg arm produced no significant benefit, and a modelled optimal dose of around 3.75 mg/kg was proposed ^[9]. The dose-response is non-monotonic: more was not better. That directly undercuts the community reasoning behind front-loading protocols, which have no basis in controlled human trials.

Healthy human volunteers, full-length Tβ4, safety. A randomised, placebo-controlled Phase 1 study gave intravenous synthetic thymosin beta-4 to 40 healthy volunteers in four cohorts of ten — a single dose then daily for 14 days at 42, 140, 420 or 1260 mg. It was well tolerated, with only infrequent mild-to-moderate adverse events, no dose-limiting toxicities and no serious adverse events; pharmacokinetics were dose-proportional with half-life increasing with dose ^[11]. This is a safety study of the parent protein. It is the strongest human dataset in this file, and it measured tolerability, not repair.

Mechanistic review, multi-model. Actin binding, migration, anti-scarring, anti-inflammatory and angiogenic actions consolidated across models, with clinical trials in dermal wounds, corneal injury and heart and CNS repair as the stated rationale ^[10].

The class-level verdict. A 2026 Sports Medicine narrative review listing TB-500/thymosin β4 and BPC-157 among unapproved peptides concluded that many such peptides show favourable tissue-repair outcomes in animal models but that rigorous human safety data are scarce, with potential for serious harm, and that these compounds operate largely outside regulatory oversight ^[8].

And the results that cut the other way. In dystrophin-deficient mice, chronic thymosin beta-4 increased the number of regenerating muscle fibres but did not improve muscle strength, cardiac function or fibrosis. In a porcine study, systemic Tβ4 failed to attenuate myocardial ischaemia-reperfusion injury. Both are worth stating plainly: more visible regeneration did not equal better function, and one large-animal model returned a null result. Commercial momentum also stalled — an injectable Tβ4 acute-stroke programme was withdrawn — so a presumed clinical pipeline overstates where the evidence actually is.

Reported effects, cautions and safety

The reports in this section are anecdotal, not clinical evidence. They come from research-use communities and self-tracked write-ups, without controls, blinding or verified product identity, and no doses are reproduced here.

The dominant benefit report, by a wide margin, is faster recovery from tendon, ligament and muscle injuries — nagging soft-tissue problems described as returning to activity sooner than expected, with timelines varying substantially between individuals. Frequently reported: looser, less achy joints with better range of motion after a few weeks, and a general sense of improved flexibility and physical resilience noticed around three to four weeks in. Occasionally reported: a vaguer feeling of reduced inflammation or calmed-down soreness, and faster healing of cuts, surgical sites or skin irritation. Rarely, users report more hair growth over four to eight weeks, usually while also using other peptides or light therapy, which makes any individual result impossible to attribute.

On the adverse side, injection-site redness, swelling or aching is very commonly reported and typically described as gone within a day or two. Temporary tiredness or lethargy is frequently reported, especially in the first few doses, and is the most consistent systemic effect mentioned. Occasionally reported: a brief head rush, light-headedness or mild headache shortly after administration, and a short-lived flu-like feeling in the first day or two. Rarely reported: nausea, a heightened awareness of an existing injury during the first week or two, and short-lived mood changes. There is no clinical evidence tying the fragment to mood effects.

Cautions from the literature. Human safety for the heptapeptide is essentially unstudied — there are no completed controlled human trials of TB-500 for any use, and the 2026 review's conclusion applies squarely ^[8]. The parent protein is overexpressed in several cancers, including pancreatic and colorectal, and is implicated in metastasis and tumour angiogenesis; the same pro-migratory, pro-angiogenic properties that might aid repair could in principle support tumour progression, which is why current or past cancer is the most frequently flagged reason for caution. It is prohibited in sport and detectable, so a tested athlete risks eligibility regardless of any claimed benefit. The mdx-mouse result is a standing warning that felt or visible regeneration need not mean functional repair. The fragment-versus-protein gap means claimed effects could be weaker or simply different from the parent protein's. Research-grade material is not made to medicine-grade standards, and identity, purity and exact sequence vary between suppliers — which also makes every anecdote above harder to interpret. Because the parent protein influences vessel formation and cell migration and is released by platelets at injury sites, clotting disorders and imminent surgery are a precautionary, mechanism-based concern, unstudied for the fragment in humans. And since the compound acts on cell movement and new vessel growth — processes central to development — pregnancy, breastfeeding and the still-growing form a group where the conservative position is avoidance.

Where it fits in bone, cartilage and hard-tissue repair

TB-500's presence in hard-tissue conversations is inherited rather than earned. The parent protein has a real regenerative literature; the fragment sold under the name has a crystal structure, a set of extrapolations, and no efficacy trial of its own.

Read against the hard tissues, the mechanism is a cytoskeletal one — cells become more able to migrate. That is upstream of nearly every repair process, which is exactly what makes it so easy to attach to any tissue a marketer likes. But the tissues actually measured in the cited work are brain after stroke ^[9], cornea, skin, heart and skeletal muscle ^[10], and the safety endpoints of healthy volunteers ^[11]. No fracture model. No chondral defect. No osteoarthritis endpoint. No bone density, no cartilage thickness, no joint-space measure — in any species.

The mdx-mouse finding deserves the last word here, because it is the sharpest available correction to the extrapolation habit: the parent protein increased regenerating fibres without improving strength, cardiac function or fibrosis. Visible tissue activity is not the same as restored function — and in the hard tissues, restored function is the only claim that would matter.