SIDE BY SIDE
Four Peptides, Four Different Kinds of Evidence
They are discussed as interchangeable repair compounds. They are not. They differ in mechanism, in the species their data comes from, in how mature that data is, and in how close any of it gets to a hard tissue — which is to say, not close.
The short version
These four peptides get grouped together because people use them for the same reason — something hurts and is not healing. But the four have very little in common underneath.
BPC-157 works mainly by encouraging new blood vessels to grow. TB-500 works on the internal scaffolding cells use to move. GHK-Cu carries copper and tells connective-tissue cells to build matrix. KLOW is simply all of those plus a fourth peptide, KPV, in one vial.
Their evidence differs even more than their mechanisms. BPC-157 has the biggest pile of animal studies. TB-500 has a strong piece of molecular structure work — but most of its healing evidence belongs to a bigger protein, not the fragment being sold. GHK-Cu has the best human trials of the four, and they are all about skin and hair. KLOW has no studies of its own at all.
One thing they share completely: not one of them has a controlled study measuring bone healing or cartilage repair, in any species.
At a glance
| BPC-157 | TB-500 | GHK-Cu | KLOW | |
|---|---|---|---|---|
| What it is | 15-amino-acid synthetic peptide from a gastric juice protein | 7-residue fragment (Ac-LKKTETQ) of thymosin beta-4 | Gly-His-Lys tripeptide chelated 1:1 to copper(II) | Four peptides co-dissolved in one vial |
| Primary mechanism | Angiogenesis via VEGFR2 up-regulation and internalisation, Akt-eNOS downstream [4] | G-actin sequestration, 1:1, dual-end capping — cell migration [12] | Matrix synthesis and MMP/TIMP rebalancing; broad transcriptional shift [14][16] | All of the above plus NF-κB suppression by KPV [18] |
| Best animal evidence | Transected Achilles tendon, Wistar rats [6]; gastric ulcer, Wistar rats [5] | Embolic stroke, Wistar rats, full-length protein [9] | Rodent and cell studies; no primary animal study in this set | Mouse colitis, for the KPV arm only [18] |
| Best human evidence | IV safety pilot, n=2, no efficacy endpoint [1] | Phase 1 IV safety, n=40, full-length protein, no efficacy endpoint [11] | 45-man 6-month hair-count trial of a combination topical [15]; topical wrinkle trials [13][16] | None — no study of the blend exists |
| Evidence maturity | Large preclinical body, three human pilots total [2] | One clean structure paper, borrowed efficacy literature [8][12] | Small but genuinely controlled human topical trials | Extrapolation only |
| Pharmacokinetics | Characterised in rats and dogs: half-life under 30 min [3] | Dose-proportional for the full-length protein [11]; fragment uncharacterised | No validated human PK for injectable or systemic use | Inherent mismatch across the four components |
| Regulatory | Not approved; FDA 503A category 2; WADA-prohibited at all times | Not approved; WADA-prohibited; prescription-only in some jurisdictions | Legal topical cosmetic ingredient; injectable use unapproved | Not approved; WADA implications via TB-500 |
| Bone or cartilage endpoint | None | None | None | None |
Where they actually differ
Mechanism class. BPC-157 is vascular, TB-500 is cytoskeletal, GHK-Cu is transcriptional and structural. Those are three genuinely different levels of biology, and the difference matters for the hard tissues in particular. A vascular mechanism has an obvious relationship to bone, which heals through a blood-supplied callus, and almost none to articular cartilage, which is avascular. A matrix-synthesis mechanism is the closest thing here to a cartilage-relevant story — and the studies that measured it measured skin [16].
Species profile. BPC-157's efficacy literature is rat, with pharmacokinetics in rat and dog [3][5][6]. TB-500's efficacy literature is rat plus reviews of multi-model work, largely on the parent protein [9][10]. GHK-Cu's efficacy literature is the only one in this group that is substantially human — and it is human skin [13][15][16][17]. KLOW's is mouse for one arm, human-cosmetic for another, and nothing for the combination [18].
Evidence maturity. Ranked by how close each comes to a controlled human efficacy result: GHK-Cu first, by a wide margin, on the strength of a 45-man randomised hair trial and placebo-controlled topical skin work [15][16]. TB-500 second, but only on a technicality — its 40-volunteer Phase 1 study measured safety, not efficacy, and tested the parent protein [11]. BPC-157 third: a large, coherent preclinical body and three human pilot studies total [2]. KLOW last, with nothing of its own.
Identity risk. Two of the four carry a definitional problem. TB-500 is routinely marketed on data generated with a molecule five times its size [8]. KLOW is routinely marketed on the sum of its parts, with a pharmacokinetic mismatch that makes the sum implausible as delivered. Neither problem shows up in a spec sheet.
Regulatory exposure. BPC-157 and TB-500 are both prohibited in sport at all times, and anti-doping laboratories detect TB-500 and its breakdown products. GHK-Cu is not currently on the Prohibited List as of the 2024–2025 lists, though the S0 catch-all can cover non-approved substances. KLOW inherits TB-500's prohibition wholesale.
How to read the four together
The honest synthesis of this comparison is unflattering to all four, and it is not a close call.
The compound with the most evidence relevant to hard tissue is BPC-157, and that evidence is a severed tendon in a rat [6] plus a mechanism that plausibly matters for bone and does not obviously matter for cartilage [4]. The compound with the best human evidence is GHK-Cu, and that evidence is about wrinkles and hair [15][16]. The compound with the best molecular science is TB-500, and that science describes an actin interaction, not a healing outcome [12]. The compound with the most confident marketing is KLOW, and it has the least data of the four.
Two reviews frame the whole category the same way, from different disciplines. A 2025 review of BPC-157 for musculoskeletal healing: only three human pilot studies, no rigorous large-scale trials, treat as investigational [2]. A 2026 sports-medicine review of unapproved peptide therapies: favourable animal-model tissue repair, scarce human safety data, potential for serious harm, largely outside regulatory oversight [8].
What none of the four supports is the claim that brought most readers here. No peptide on this site has been shown in a blinded human trial to restore cartilage, and adult articular cartilage's near-absent intrinsic repair capacity is exactly why that claim would need such a trial before anyone believed it. What the studies measured is what this site reports; the rest is a hypothesis wearing a citation.