Can exosomes cause cancer? What is shown and what is not

Why this page exists

The question reaches us more often than any other on this subject, usually from people who have read that exosomes are involved in how cancer spreads and have then been offered an exosome treatment for their skin or their hair. Both things they read were true. The inference between them is the part that needs care.

Two questions that are not the same question

Question one is about tumour biology. Cancer cells, like all cells, release extracellular vesicles, and those vesicles carry material from the cell that made them. The research asks what those particular vesicles do.

Question two is about treatment safety. It asks whether a preparation derived from a non tumour cell, given to a person who does not have a cancer, changes that person's chance of developing one.

The answer to question one does not answer question two, in the same way that the fact that a wasp's venom gland makes venom tells you nothing about what a bee's honey sac makes. The contents of a vesicle reflect the cell it came from, which is the whole point of the tumour work and the reason it cannot be read across.

What the tumour vesicle research actually showed

Two papers are cited more than any others when this subject comes up, and it is worth being precise about what each did.

In 2012 a group reported in Nature Medicine that vesicles from melanoma cells altered bone marrow progenitor cells in mice in a way that favoured metastasis, with the receptor MET implicated in the effect. In 2015 a group reported in Nature that integrins carried on tumour vesicles were associated with which organ a tumour spread to, again in animal models.

Both are careful papers about tumour biology. Neither administered vesicles from a healthy cell source to a healthy subject, and neither set out to measure the safety of any treatment.

There is a further detail that this publication considers part of reading the evidence properly. In 2018 an independent replication of the 2012 melanoma work, run under the Reproducibility Project in Cancer Biology, reproduced some elements and did not reproduce others: the direction of the metastatic effect was similar but the result did not reach statistical significance, and phosphorylated MET could not be reliably detected in the vesicles. That does not overturn the original. It does mean the finding is less settled than the way it is usually quoted, and we set out why that pattern is normal in translation failure is normal and in what peer review does.

Exosome treatments cause cancer.

Proposed mechanism

Vesicles delivered to skin or scalp carry signalling cargo that drives malignant transformation or promotes an existing occult tumour.

What has been shown

Vesicles released by tumour cells alter recipient cells and influence metastatic behaviour in cell culture and in animal models. No study has tested whether a preparation from a non tumour source, given for an aesthetic indication, changes cancer incidence in people. No such outcome has been reported to us in the published literature.

Highest level reached

Animal models, and only for tumour derived vesicles

Main confounders

Reading a finding about vesicles made by tumour cells as a finding about vesicles made by other cells. Cancers take years to appear, so a short follow up would miss an effect that existed. No registry records who has had these treatments, so nobody is counting.

GradeNOT ASSESSABLE

What would change thisLong term follow up of a defined cohort who received a characterised preparation, compared with a matched group who did not, with cancer incidence as a named endpoint. Nothing short of that moves this claim in either direction.

Exosome treatments have been shown to be safe with respect to cancer risk.

Proposed mechanism

Safety has been established by use, by trials, or by regulatory assessment.

What has been shown

We are not aware of any long term safety study of aesthetic exosome use with a cancer endpoint, and we are not aware of any exosome product holding a UK marketing authorisation for an aesthetic indication, which means no regulator has assessed it for that use. Widespread use without systematic follow up is not a safety finding.

Highest level reached

Not shown

Main confounders

Absence of reports being read as absence of events, in a setting where no mechanism exists to collect them.

GradeNOT SHOWN

What would change thisA prospective cohort with a defined follow up period and a named oncological endpoint, or a regulatory assessment of a specific product for a specific use.

Why a laboratory finding does not transfer to a vial

Three things separate the two settings, and all three cut in the same direction.

The source cell is different, and the cargo follows the source. The dose, the route and the tissue are different: an animal model of metastasis delivers vesicles systemically to an animal that already carries a tumour, which is not what happens when a preparation is applied to skin after microneedling. And the question being asked is different: those studies measured tumour behaviour, not tumour initiation in a healthy subject.

None of that makes a preparation safe. It means the tumour literature is the wrong instrument for measuring it, and reaching for it produces a confident answer to a question it was never asked.

What would have to be true for a real risk

A responsible version of the worry is not the one usually stated, and it is worth writing down because it is answerable by a clinic in a way the vague version is not.

• An undefined cell source. If nobody can say which cell made the material, nobody can say what the cargo is. That is the first question, and it is covered in why the cell source matters.
• An immortalised or transformed cell line. Lines that have been made to divide indefinitely differ from primary cells, which is a reason to ask what was used rather than a demonstration that anything went wrong.
• Weak characterisation. A preparation that has not been characterised is a preparation whose contents are asserted rather than measured, which is the subject of what a real specification sheet shows.
• No batch release testing. The documented harms in this area to date have been infections traced to contaminated material, which is a manufacturing failure rather than a biological one.
• No follow up. Nobody is recording who had what, so no signal could be detected even if one existed.

Those are the conditions that would make a risk possible to have and impossible to notice. They are also, usefully, the things a person can ask about before agreeing to anything.

What regulators have and have not said

We are not aware of any regulator stating that exosome products cause cancer. What regulators have addressed is a different set of problems: approval status, manufacturing, and what may be claimed.

In the United Kingdom the position depends on what a product is made of and how it is presented, and we set it out at length in the UK position on exosome products, with the borderline test explained in borderline products and the MHRA. In the United States the regulator has issued a public safety notification on exosome products, and the US public health agency has published an account of bacterial infections associated with stem cell and exosome products, in which the problems found on inspection were manufacturing process failures and inadequate donor screening.

That is the shape of the documented harm so far, and it is worth stating plainly because it is both real and different from the fear people arrive with.

Why nobody can give you a number

A risk figure requires three things: a defined population who had the exposure, a comparable population who did not, and enough follow up time for the outcome to appear. Aesthetic exosome use has none of them. Treatments are given outside trials, no register records them, and cancers of the kind people worry about take years.

This is why we grade the claim not assessable rather than false. Not assessable is a statement about the evidence, not a reassurance and not a warning, and the difference between those three things is the whole discipline of reading the evidence.

What a careful reader does with this

Treat the tumour vesicle literature as what it is: good science about how cancers use vesicles, and no evidence at all about the safety of a cosmetic preparation. Treat an assurance that exosome treatments are proven safe as unsupported, because no study establishing that exists. Treat a claim that they cause cancer as equally unsupported.

Then ask the questions that can actually be answered: what cell, what characterisation, what release testing, what the supplier will put in writing. Those are set out in the standards checklist, and they are the part of this subject where a person still has some control.

Questions readers ask

Do exosomes cause cancer?

No study has shown that an exosome preparation given to a person caused a cancer, and no study has shown that it cannot. The research that links vesicles to cancer concerns vesicles released by tumour cells themselves, studied in cell culture and in mice. That is a finding about tumour biology, not a measurement of the risk of a treatment.

Are tumour derived vesicles the same as the ones in an aesthetic product?

No. The source cell is the defining difference, and preparations sold for aesthetic use are not made from tumour cells. What the tumour work does establish is that the contents of a vesicle depend on the cell that made it, which is why the cell source and the characterisation of a preparation matter more than the word on the label.

Has any regulator said exosome treatments cause cancer?

Not to our knowledge. The regulatory statements we can find concern approval status, manufacturing and marketing rather than cancer. The documented harms reported to date in this area have been infections traced to contaminated product, not malignancy.

Why can nobody give a number for the risk?

Because a number would require long term follow up of people who had the treatment, compared with people who did not, and that study has not been done for aesthetic exosome use. Absence of a reported harm in a setting where nobody is systematically looking is weak evidence in either direction.

What should a person ask a clinic about this?

What cell the material came from, whether the cell line is primary or immortalised, what characterisation and release testing each batch receives, and what the clinic will say in writing about long term data. A supplier who cannot answer the first question cannot answer this one either.