
BPC-157: Body Protection Compound
BPC-157 is a 15-amino-acid peptide first isolated from human gastric juice. It is one of the most-studied and least-approved compounds in the recovery world: more than a hundred animal papers describe it healing tendon, gut, muscle, nerve, and bone, roughly three small human reports exist, and in July 2026 an FDA advisory committee split 8-to-6 over whether pharmacists should even be allowed to compound it. This page is the complete picture — what it is, the single mechanism the whole theory rests on, the full breadth of the animal record, the thin human evidence, how people actually dose and reconstitute it, the live regulatory fight, the cancer question everyone asks, and where it fits for a spinal disc. It assumes you know nothing going in.
Made in the stomach, stable in acid
The name is literal: Body Protection Compound. It is a chain of 15 amino acids taken from a protein sequence found in gastric juice, and its one genuinely unusual physical property is that it survives stomach acid — resistant to degradation in human gastric juice for more than 24 hours. That acid-stability is why oral dosing is even on the table for a peptide, and it points to the gut as its home turf. Almost the entire research record traces to one group in Croatia, led by Predrag Sikirić, running since 1993 and now past 150 papers. A journalist who spent four months tracing the peptide's path from that post-communist lab into the American supplement scene told the story in an AMA worth reading before you trust any vendor.
The single-lab origin cuts both ways. It is a coherent, decades-deep research program — and it is also most of the evidence coming from one invested group, which is exactly why independent human trials matter and why their near-total absence is the headline, not a footnote.
One mechanism the whole theory rests on: angiogenesis
Strip the marketing and BPC-157's proposed action is a short, testable chain. It raises signals that grow new blood vessels at an injury; new vessels bring oxygen and repair cells; alongside that it drives collagen production, fibroblast activity, and a specific push on cell migration.
The cell-migration step has a named molecular basis: in tendon cells BPC-157 dose-dependently increases phosphorylation of FAK and paxillin — the machinery a cell uses to crawl and adhere. That is a real mechanism, not a hand-wave.
It also modulates nitric oxide in both directions depending on context, paired with counteracting free-radical damage.
The animal record is broad — and consistent
This is the strong part of the file. A systematic review screened 544 papers and kept 36; 35 were preclinical, and across them the results converge: better functional, structural, and biomechanical outcomes in muscle, tendon, ligament, and bone.
Transected Achilles tendon and severed myotendinous junctions recovered better with the peptide.
In the gut — its native territory — it healed colocutaneous fistulas in rats, working through the nitric-oxide system, and it entered a human IBD trial program (PL14736).
Most relevant to this site's purpose: after experimental spinal-cord injury in rats, a single injection produced consistent motor recovery and resolved spasticity by day 15.
Consistency across dozens of models in different tissues is real evidence of a real biological effect. It is not evidence that the effect reaches a useful size in a human tendon, gut, or disc — a distinction the honest reviews keep making.
The human evidence is three studies
Said plainly: the controlled human record is a two-person intravenous safety study (up to 20 mg, no adverse effects), plus a small number of pilot reports.
McGill's science office counted it precisely — to date, only three publications have reported giving BPC-157 to humans.
There is one registered randomized, placebo-controlled Phase 2 trial, for acute hamstring strain, that could actually move the needle. It is not yet published.
How people actually take it
Because no human trial set a dose, every regimen in circulation is extrapolated from animals or copied from clinic practice — treat all of it as folklore with a citation, not medicine.
- Dose. Rodent work is effective across an enormous range, most commonly around 10 mcg/kg. Human practitioner protocols that grew out of that typically run ~200-500 mcg per day, often as a 4-8 week course, sometimes split into a loading and maintenance phase. No validated human dose, cycle, or taper exists.
- Route. Subcutaneous injection near the injury is the dominant musculoskeletal route, on the theory of higher local exposure; systemic subcutaneous dosing is used for gut, nerve, and organ targets. Oral capsules are used specifically for gut problems, exploiting the acid stability — but oral stability is not the same as systemic absorption, and systemic effects from oral dosing are unproven.
- Reconstitution. It ships as a lyophilized powder and is reconstituted with bacteriostatic water (the benzyl-alcohol preservative matters for a multi-dose vial). Standard handling: bring the vial to room temperature, run the diluent down the wall, swirl — never shake, which shears and aggregates the peptide — then store at 2-8 °C and use within about four weeks.
The FDA just fought over it — in public
This is the live, moving part of the story. In 2023 the FDA placed BPC-157 in Category 2 of the 503A bulk-substances list — the "significant safety risk or insufficient data" bucket that bars pharmacy compounding.
Then in 2026 it was removed from Category 2, and on July 24, 2026 the FDA's own Pharmacy Compounding Advisory Committee voted 8-6 (one abstention) to recommend adding it to the 503A list — over the objection of FDA staff, who cited a "lack of evidence to support the effectiveness of BPC-157."
That vote is a non-binding recommendation and still faces a long rulemaking process. But it captures the exact tension of this compound better than any review: real enthusiasm, thin evidence, and regulators split down the middle.
The sourcing problem is not a footnote
Nearly all consumer BPC-157 is grey-market "research chemical," labeled "not for human use" — which is a legal shield, not a safety standard. Independent testing of grey-market peptides is grim.
30% contained incorrect amino acid sequences and two-thirds of samples fell below the 95% purity threshold.
And a certificate of analysis reading "99% by HPLC" does not clear it — HPLC confirms the target peptide is present but says nothing about what makes up the rest, which can include pyrogens or the wrong sequence entirely. If someone uses this compound, third-party HPLC and mass-spec per batch is the floor, and vendor-only certificates are worthless.
Does it cause cancer? The question everyone asks
The fear is logical: a compound that grows blood vessels could, in theory, feed a tumor, which needs blood vessels to grow. The actual evidence points the other way. In the lab, BPC-157 inhibited growth and VEGF signaling in human melanoma cells, reduced melanoma lung metastases in mice, and countered tumor cachexia while extending survival.
The framing that fits the data is that it modulates a balance toward physiological, self-limiting angiogenesis rather than driving the pathological kind. No published rodent study shows it accelerating tumor growth. That said: there is zero human oncology safety data, and caution with active malignancy remains the sane default given the mechanism.
What people report — including the ones it hurt
The reputation was built on anecdote, and it is broad and cross-community. A climber called a stubborn elbow "almost 100% healed"; a finger-pulley tear rebuilt in a week; a sprinter with wrecked knees went back to landing flips; an EDS patient cleared lingering injuries faster than usual.
But a tool with no human safety file is one whose bad reactions are also uncatalogued — and they exist. One user hit extreme dizziness, migraine, and sensory overload within two hours of a low dose.
These are uncontrolled, unblinded, self-selected reports. They explain why people keep trying it. They are not evidence it works, and the bad ones are a real part of the picture.
Timelines, and one honest trap
In animal soft-tissue models, better biomechanics show up around 2-3 weeks and strength keeps climbing past a month. Human timelines are anecdotal: reduced pain and inflammation often in the first 1-4 weeks, structural change claimed over 2-3 months. The trap worth naming: early pain relief is an anti-inflammatory effect and can arrive before any actual collagen repair — feeling better is not proof the tissue is fixed.
Where it fits for the spine
BPC-157 is the tissue-repair base layer of a disc-recovery stack. Its rat spinal-cord recovery data and its angiogenic, connective-tissue-healing mechanism are exactly aimed at the disc's core weakness — it is nearly avascular, and healing there is blood-supply-limited. It pairs with TB-500 (a second repair signal), while ARA-290 handles the neuroinflammatory nerve pain and retatrutide handles mechanical load. The ceiling, stated flatly: no one has run BPC-157 against a human disc. The case is mechanistic alignment plus animal data, and it belongs in the framework laid out in the disc-stack, herniated-disc, and degenerative-disc-disease pages, not in a prescription.
Proven, unproven, unknown
Proven: BPC-157 heals across many rat tissue models (tendon, gut, nerve, bone), is acid-stable, and caused no acute harm in a two-person IV study; in the lab it opposes, rather than drives, tumor angiogenesis. Unproven: that any of it works in people — the human file is three small studies. Unknown: real dosing, long-term safety, drug interactions, and whether the disc rationale holds beyond mechanism. The regulators are split; the sourcing is a minefield; the mechanism is genuinely interesting. All three are true at once.
Not medical advice. BPC-157 is not an approved drug; consumer product is grey-market and unverified. Nothing here is a dosing or treatment recommendation.
Key evidence
136 more ranked claims
Low-confidence / auto-generated 4
Model review212 contributions · 8 modelsExpand the recursive review layer
/api/articles/bpc-157/contributionsAsk this article · 8 suggested prompts
Text the build (+14245134626) or WhatsApp — slug|question creates a question node. Paste evidence with ingest slug|q:NODE_ID|your paste.