BPC-157 (body protection compound 157; sequence GEPPPGKPADDAGLV; ~1,419 Da) is a synthetic pentadecapeptide derived from a gastric-juice protein fraction first characterised by Sikiric and colleagues in Zagreb in the early 1990s. It is not an MHRA-authorised medicine. It is not a controlled substance under the Misuse of Drugs Act 1971. It is, in the United Kingdom, an unlicensed research chemical whose legal status turns on how it is presented: laboratory material is one thing; a product offered with claims of tissue repair, recovery or disease treatment is a medicinal product under the Human Medicines Regulations 2012.
That distinction matters more in 2026 than it did a year ago. Two independent orthopaedic reviews published in 2025, a biopharmaceutical appraisal in May 2026, a first registered randomised musculoskeletal trial with no results, and an MHRA investigation into UK peptide clinics have made the evidence hierarchy unusually clear. This explainer sets out what that literature actually shows — and what it does not.
1. What the molecule is
BPC-157 is a linear 15-residue peptide with an N-terminal polyproline cluster (positions 3–5) that is widely cited as the structural basis of its unusual stability in gastric juice. No high-affinity receptor has been identified after three decades of work. Proposed signalling routes in preclinical models include VEGFR2–PI3K–Akt–eNOS angiogenesis, Egr-1 transcriptional priming, nitric-oxide system modulation, and FAK–paxillin engagement in tendon fibroblasts. These are overlapping, partly redundant hypotheses, not a defined receptor-occupancy model.
The absence of a characterised binding site is not a curiosity. It means standard PK/PD modelling — the framework used to justify first-in-human dose selection for almost every licensed peptide medicine — cannot yet be applied.
2. The preclinical column is real, and concentrated
Vasireddi and colleagues, writing in HSS Journal in 2025, searched PubMed, Cochrane and Embase for English-language BPC-157 papers from 1993 to June 2024. They screened 544 records and included 36 studies. Thirty-five were preclinical. One was clinical. In rodent models of Achilles transection, ligament injury, muscle laceration and fracture, the peptide has been associated with faster structural repair, improved biomechanical parameters and enhanced neovascularisation of poorly vascularised connective tissue.
That ratio — hundreds of laboratory papers to a single qualifying human study in an orthopaedic systematic review — is the single most important fact about the field. McGuire and colleagues, in a December 2025 narrative review in Current Reviews in Musculoskeletal Medicine, reached the same structural conclusion from a slightly different cut of the literature: robust regenerative and cytoprotective signals in animals; human data “extremely limited”.
Two methodological caveats sit on top of that column. First, Mateescu and colleagues note that more than 80% of published BPC-157 studies, by author-affiliation analysis, originate from a single research group at the University of Zagreb. Internal consistency is not independent replication. Second, most preclinical experiments use a single dose level, which precludes construction of a dose–response curve, a minimum effective concentration, or a therapeutic window — all of which are regulatory prerequisites for clinical dose selection.
Preclinical activity is a reason to run a proper trial. It is not a substitute for one.
3. The human dataset: fewer than 30 subjects, no randomised trial
Across the entire peer-reviewed literature, three small uncontrolled human studies account for the published clinical exposure:
- A retrospective chart review of intra-articular administration in patients with chronic knee pain, identified by Vasireddi et al. as the sole clinical paper meeting their orthopaedic inclusion criteria. Uncontrolled, unblinded, no standardised pain instrument, no placebo arm. Intra-articular injection trials in knee pain routinely show placebo response rates of 40–50%, which makes an uncontrolled observation in this indication particularly hard to interpret.
- A 12-patient interstitial-cystitis pilot using intravesical administration (Lee, Walker and Ayadi, 2024). Again uncontrolled; interstitial cystitis has high symptom variability and a substantial placebo response.
- A two-subject intravenous pharmacokinetic and safety pilot (Lee and Burgess, 2025) in a 58-year-old man and a 68-year-old woman. Serial plasma sampling showed a half-life under 30 minutes, with concentrations returning toward baseline within 24 hours. No adverse events were reported. A sample of two provides no meaningful characterisation of inter-individual variability and cannot detect uncommon harms.
Mateescu et al., writing in Pharmaceutics in May 2026, place this combined enrolment — fewer than 30 subjects, none in a randomised controlled design — at Oxford CEBM Level IV–V, below the threshold required to support a clinical practice recommendation under any major evidence-based-medicine framework. McGuire et al. are equally direct: until well-designed clinical trials are conducted, BPC-157 “should be considered investigational, and its use approached with caution.”
A Phase I trial registered in 2016 (NCT02637284), designed to enrol 42 healthy volunteers for safety and pharmacokinetics, was terminated without publication of results. That missing dataset is the vacuum into which uncontrolled observations and grey-market use have expanded.
4. Pharmacokinetics: a short plasma half-life and a long pharmacodynamic tail
The only formal ADME characterisation is He et al. (2022) in Sprague-Dawley rats and beagle dogs. After a single intravenous dose, elimination half-life was 15.2 minutes in rats and 5.27 minutes in dogs. Intramuscular bioavailability was 14–19% in rats and 45–51% in dogs — a species gap that makes informal allometric scaling to humans unreliable. Metabolism was rapid proteolytic hydrolysis to small fragments, with urinary and biliary excretion.
The two-subject human infusion study is directionally consistent with that preclinical clearance. What it does not explain is the reported persistence of biological effects in animal injury models days to weeks after a single exposure. Candidate mechanisms — transcriptional priming via Egr-1, the slow cell biology of collagen deposition and neovascularisation, possible local tissue retention — remain unranked experimentally. No published study has related plasma concentration to a tissue-level effect. Without that relationship, dosing regimens used outside trials are informal extrapolations, not pharmacokinetically justified protocols.
No pharmaceutical-grade formulation has been standardised. No BCS classification, Caco-2 permeability, plasma-protein binding, or ICH-compliant forced-degradation package has been published. Grey-market and compounding preparations therefore cannot be assumed to deliver a known dose of a characterised substance.
5. NCT07437547: a trial worth watching, not a result
Hudson Biotech has registered a randomised, double-blind, placebo-controlled Phase 2 trial (NCT07437547) of pentadecapeptide BPC-157 in acute MRI-confirmed grade II hamstring strain. Target enrolment is 120 adults aged 18–45 at Peking University Shenzhen Hospital. Both arms receive a standardised rehabilitation programme. Primary endpoints include time to return to unrestricted sport and change in MRI-assessed injury volume at day 14, read by blinded central radiology. The study is listed as recruiting; estimated primary completion is February 2027. No results have been posted.
This is the first randomised musculoskeletal evaluation with an imaging endpoint. Phase 2 trials are signal-detection studies. They are not powered to confirm clinical benefit at the level required for a marketing authorisation. A positive signal would support a larger programme; a negative result would clarify whether the rodent literature translates. Until results are published, the trial changes the prospect of evidence, not the evidence itself.
6. UK law: presentation, not the research-use label, decides the class
BPC-157 holds no marketing authorisation from the MHRA. It is not named in the Misuse of Drugs Act 1971. Possession of a research chemical is not, by itself, an offence. Supply and presentation are another matter.
Under the Human Medicines Regulations 2012, a medicinal product includes any substance presented as having properties of preventing or treating disease — and “disease” includes injury, ailment or adverse condition — or administered with a view to restoring, correcting or modifying a physiological function by pharmacological action. Lynda Scammell, head of borderline products at the MHRA, stated in April 2026 that the agency determines classification case by case, taking account of effect, intended use, and all available evidence, and that it “disregard[s] claims that products are for ‘research purposes’ if it is clear that such claims are being used as an attempt to avoid medicines regulations.”
On 4 April 2026 The Guardian reported that the MHRA was investigating UK clinics offering unlicensed peptides, including BPC-157, MOTS-c, Cortexin and thymosin alpha-1, where websites and consultations made recovery, tissue-repair or metabolic claims. An MHRA spokesperson said that if clinics offering peptide injections make medicinal claims, the products will be considered medicines under the 2012 Regulations, and that the agency will take action against clinics identified as breaching those requirements. After the newspaper approached one clinic, the medicinal claims were removed from its site.
This is the same legal line Peptide Data applies across the knowledge base: research-use-only is an absolute framing, not a marketing workaround. Licensed peptide medicines — semaglutide (Wegovy, Ozempic, Rybelsus), tirzepatide (Mounjaro), liraglutide (Saxenda) — sit in a different legal category because they hold marketing authorisations. BPC-157 does not.
Athletes face a further constraint. BPC-157 has been treated as an unapproved substance under the World Anti-Doping Agency framework; anyone subject to a WADA-aligned testing programme should treat it as prohibited unless a current list entry says otherwise. NCT07437547 itself excludes participants in formal anti-doping programmes for that reason.
US compounding policy (FDA 503A Category 2 history and the July 2026 Pharmacy Compounding Advisory Committee review) is a different jurisdiction and does not determine UK classification. It is noted here only as a parallel illustration of regulators asking the same question: is there enough human evidence to treat this as a medicine?
7. What this does and does not establish
Established, with caveats. A large, internally consistent preclinical literature reports cytoprotective and regenerative signals in rodent gastrointestinal, tendon, ligament and muscle-injury models, with plausible angiogenic and nitric-oxide mechanisms. Gastric stability is a distinctive physicochemical property among linear peptides of this size. Short plasma half-life is confirmed in two species and directionally in two humans.
Not established. Human efficacy for any musculoskeletal, gastrointestinal or systemic indication. A safety database capable of detecting uncommon harms. A characterised pharmaceutical formulation. A pharmacokinetically justified human dose. Independent replication of the Zagreb preclinical programme at scale. Any MHRA-authorised indication.
The honest research position in 2026 is therefore narrow. BPC-157 is a molecule with biological plausibility that has earned a controlled trial. It has not earned a clinical claim. Clinics and vendors that convert rodent histology into recovery advertising are making a medicinal presentation of an unlicensed peptide. That is a regulatory problem under UK law, not an evidence disagreement.
Related Peptide Data profiles: BPC-157, TB-500, thymosin beta-4, MOTS-c. TB-500 / thymosin beta-4 is the compound most often stacked with BPC-157 in grey-market marketing; its human musculoskeletal dataset is, if anything, thinner.
Research-use framing only. This article describes published literature and UK regulatory status. It is not medical advice, not a protocol, and not a recommendation for human administration.
This article is AI-researched and editorially reviewed. It is provided for research and educational purposes only and is not medical advice. Research peptides are not licensed for human consumption in the UK.