Can Exosomes Cause Cancer? A Sourced Evidence Checklist

Exosomes cannot be assigned a single cancer risk as a class. Risk assessment depends on the starting cells, manufacturing process, cargo, route, exposure and evidence quality. This checklist separates findings that show a biological signal from evidence capable of supporting a product-specific safety conclusion, using named scientific and regulatory standards where available.

What this checklist can establish

This is a product-assessment checklist for a reader reviewing claims or supplier evidence about whether extracellular vesicles, including products described as exosomes, can cause cancer. It does not decide whether any person should receive a product. For that question, exosomestherapy.co.uk addresses the patient decision separately.

There is no single internationally adopted test standard that can answer the question for every vesicle preparation. “Exosome” does not describe a uniform ingredient, a fixed dose, or a fixed biological effect. A meaningful conclusion must connect a defined product to an appropriate test system and a stated exposure. It must also distinguish tumour initiation, promotion of existing tumour cells, altered tumour spread, and a laboratory finding that has not been shown to occur in a living organism.

1. [ ] Write down the exact conclusion being claimed. Check whether the document claims absence of genotoxicity, absence of tumour formation, lack of effect on tumour-cell behaviour, or broad product safety. These are different propositions requiring different evidence. A short cell-culture experiment cannot by itself establish long-term tumour safety.
2. [ ] Identify whether the claim concerns the tested batch or an entire product line. A study of one batch is evidence about that batch unless comparability data show that other batches are materially equivalent. This matters especially where donor material, cell culture conditions, purification and storage can alter the preparation.
3. [ ] Record what the evidence does not test. The International Society for Extracellular Vesicles, in its MISEV2018 guidance, stresses that extracellular-vesicle studies should report separation and characterisation methods and discuss limitations. Missing information is not evidence of no risk.

A defensible review records the boundaries of each experiment before considering its result. The checklist below is designed to make those boundaries visible rather than to turn incomplete evidence into a reassuring label.

1. Confirm that the tested material is identified

The first task is to establish whether the material in the experiment can be connected to the material being assessed. The International Society for Extracellular Vesicles advises in MISEV2018 that authors use operationally defined terms where biogenesis cannot be directly demonstrated, and report relevant characterisation. This is important because a preparation called “exosomes” may contain a mixture of extracellular vesicles, non-vesicular particles, soluble proteins, lipoproteins or process-related residues.

1. [ ] Check the starting material and cell source. Record species, tissue or cell type, whether cells are primary or immortalised, and whether they have been genetically modified. A transformed or tumour-derived source raises different questions from a non-transformed source. The source is an essential part of the identity, not background detail.
2. [ ] Check how the preparation was separated. MISEV2018 asks for reporting of separation methods and their rationale. Note whether the method enriches small vesicles, removes soluble medium components, or leaves likely co-isolates. A result cannot automatically be assigned to vesicles if the comparator and purity controls are absent.
3. [ ] Check characterisation by complementary approaches. MISEV2018 sets out categories of markers and recommends quantifying the source material and recovered preparation. Look for evidence addressing particle-associated components, contaminants and particle properties, rather than one marker alone.
4. [ ] Check batch identity and storage history. The certificate or study record should identify the batch, manufacture date or equivalent traceable identifier, storage conditions and any freeze-thaw history. Without this, a reader cannot know whether the tested material represents the offered material.

Do not infer that a particle count proves identity, purity or biological equivalence. It is one measurement of a heterogeneous preparation. The question is whether the material used in a cancer-related assay is described well enough for the finding to be attributed to that material.

2. Check whether the source and process address plausible hazards

Cancer-related risk can arise through more than one mechanism. Cargo may affect cell proliferation, inflammatory signalling, angiogenesis, migration or survival. Process impurities may also contribute to an observed effect. The presence of a molecule associated with cancer in another setting is not proof that a given preparation causes cancer, but it can identify a question that needs targeted testing.

1. [ ] Look for a documented source-cell qualification process. Ask whether the record explains how the cell bank or donor-derived starting material was selected, maintained and monitored. For cell-based manufacturing generally, the source and its passage history can affect product attributes. A label alone does not show that the source was suitable for a safety conclusion.
2. [ ] Check whether culture conditions are disclosed. Record media components, supplements, conditioning duration, cell density and harvest conditions where available. These conditions can alter secreted material and can introduce non-vesicular constituents. A study should not leave the reader unable to distinguish culture-medium effects from vesicle-associated effects.
3. [ ] Check for a process control. A useful control may be conditioned medium processed without cells, or an appropriately matched fraction depleted of the particle preparation. The appropriate design depends on the question. Its purpose is to test whether the reported activity follows the preparation rather than the processing workflow or medium.
4. [ ] Separate cargo description from functional proof. RNA or protein profiling can generate hypotheses, but does not demonstrate delivery, persistence or biological effect at the exposure tested. MISEV2018 treats functional studies and characterisation as related but distinct evidence streams.

A supplier document that lists selected microRNAs or proteins without explaining the source, method, normalisation and batch variation does not establish a cancer-related hazard or its absence. It is descriptive information. The reader should seek a bridge between composition, exposure and an observed biological outcome.

3. Check the experimental design before accepting a cancer signal

A cancer-associated signal is only as interpretable as the experiment that produced it. Cell proliferation, migration and invasion assays are frequently used as exploratory tools. They can show that a tested preparation changed a measured endpoint under specific laboratory conditions. They do not, without further work, show tumour causation in humans or even in a whole animal.

1. [ ] Check that the comparator is meaningful. The experiment should include a vehicle or untreated control and, where relevant, a process-matched control. If an effect is claimed to be specific to a preparation, controls must allow the reader to separate particle-associated activity from media, handling and assay background.
2. [ ] Check exposure reporting. Record the amount applied, how it was measured, the number of administrations, duration and final concentration. MISEV2018 recommends reporting dose in a way appropriate to the preparation, potentially using more than one metric. A particle number without volume, protein, source-cell input or treatment schedule may not be enough to compare experiments.
3. [ ] Check biological replication and independent batches. Technical repeats within one sample do not answer whether separately manufactured batches behave similarly. The study should state what constitutes an independent biological replicate and whether the result was repeated across batches.
4. [ ] Check whether the endpoint is blinded or objectively measured where feasible. The ARRIVE 2.0 guidelines from the NC3Rs identify study design, sample size, randomisation and blinding among information needed to judge animal-study reliability. These principles are especially relevant when assessing images, lesions or other outcomes open to observer judgement.

A positive finding can be important, but its scope should be stated narrowly: the named preparation changed the named endpoint in the named model. A negative result has the same boundary. Neither result is a universal statement about all extracellular vesicles.

4. Match the evidence to the cancer question being asked

Use the decision rule below to avoid treating unlike evidence as interchangeable. It separates directness from reassurance. A study can be competently performed and still be indirect for a particular safety question.

1. [ ] Identify whether the study tests initiation, promotion or progression. Tumour formation in a model is not the same endpoint as increased growth of implanted tumour cells. Keep the terms distinct in notes and in any certificate review.
2. [ ] Check route and distribution relevance. The route used in a study affects which tissues are exposed. A result from direct cell exposure does not establish the exposure pattern after another route. A study should report route, schedule and, where relevant, evidence of distribution or persistence.
3. [ ] Check duration against the claim. Short assays may be appropriate for acute cellular responses. They are not a substitute for evidence addressing delayed outcomes when a document makes a long-term safety claim.

The Organisation for Economic Co-operation and Development publishes test guidelines for specific toxicological endpoints, including genotoxicity assays. Those guidelines can support method selection, but they do not remove the need to justify whether an assay is suitable for a complex vesicle preparation.

5. Review genotoxicity and in vivo work without overstating either

Genotoxicity is one route by which substances may contribute to cancer risk, but it is not the only route. The International Council for Harmonisation guideline S2(R1) concerns the assessment of genotoxic potential for pharmaceuticals. It provides a recognised framework for choosing and interpreting a battery of tests in drug development. It should not be cited as proof that one isolated negative assay clears a vesicle preparation of every cancer-related concern.

1. [ ] Check whether a negative genotoxicity statement identifies the actual assay. The report should name the method, test material, concentration or exposure, controls, acceptance criteria and result. A bare statement such as “non-genotoxic” is not independently assessable.
2. [ ] Check metabolic competence and interference. Some assay systems use metabolic activation. Complex biological preparations may also interfere with read-outs. The report should explain how assay validity, cytotoxicity and interference were considered.
3. [ ] Check in vivo work against ARRIVE 2.0 reporting items. Look for species, strain, sex, group allocation, sample-size rationale, exclusion criteria, outcome measures, adverse events and statistical methods. The NC3Rs ARRIVE 2.0 framework helps a reader identify omissions that limit interpretation.
4. [ ] Check pathology and follow-up reporting where tumour findings are claimed. A claim about tumour formation or absence of tumours needs the observation period, tissues examined, pathology method and any unexpected findings. “No tumour observed” is not interpretable without those details.

Evidence should be read in both directions. A concerning signal deserves follow-up proportionate to its plausibility and reproducibility. A negative study narrows uncertainty only for the tested material, model, endpoint and exposure. It does not convert a biologically variable class of products into a single known-risk category.

6. Make a bounded conclusion and record the unresolved questions

The final step is to produce a conclusion that a later reviewer can audit. Avoid binary labels such as “proven safe” or “causes cancer” unless the available evidence truly supports that unusually broad statement. Most supplier records support a narrower conclusion, and a useful assessment makes the remaining uncertainty explicit.

1. [ ] Use this decision rule. If identity, batch traceability, relevant controls and exposure reporting are missing, classify the cancer-related claim as not assessable from the supplied record. If these are present but evidence is limited to exploratory cell work, classify it as mechanistic or hypothesis-generating. If a defined preparation has relevant, well-reported nonclinical evidence, classify it as product-specific nonclinical evidence, not a human safety conclusion.
2. [ ] List every evidence gap. Typical gaps include no matched process control, no independent batch replication, no relevant distribution data, no long-term observation, incomplete pathology reporting, or uncertainty about the material actually tested.
3. [ ] Check the date and jurisdiction of any regulatory statement. Product classification and regulatory expectations can change by jurisdiction and depend on intended use, presentation and route. A cancer-safety experiment does not itself determine a product’s legal classification.
4. [ ] Keep the conclusion separate from marketing language. State the observed result, the model, the material and the limitation. This preserves the distinction between evidence of a measured effect and a claim about real-world risk.

Limits of this checklist: it does not provide treatment advice, determine legal classification, assess infection control, or replace a product-specific toxicology programme. It does not apply a finding from one extracellular-vesicle preparation to another. It is intended for readers evaluating documentary evidence, not for diagnosing cancer risk in an individual or making a clinical decision.