Where This Field Actually Stands
It is worth putting the summary first, because most writing on this subject does not.
Exosome biology is one of the more interesting developments in regenerative medicine of the last twenty years. It reframed a genuine puzzle — why do patients sometimes improve after stem cell infusions when the cells demonstrably do not stay? — and produced a mechanistic answer that has held up.
What has not happened is the clinical part. There is no approved exosome therapy in any major jurisdiction. The evidence assembled below, which is substantial and in places striking, comes almost entirely from mice, rats and cell culture. A promising preclinical literature is not a treatment, and the gap between those two things is where most of the confusion in this area lives.
That is the honest position, and it is a more interesting one than either enthusiasm or dismissal.
What Exosomes Are
Exosomes are a class of extracellular vesicle — small membrane-bound packages that cells release to communicate with one another.
A note on terminology that matters when reading this literature: the field has largely moved to the broader term extracellular vesicles (EVs), because separating true exosomes from other small vesicles by isolation method alone is not currently reliable. The 2023 international guidance recommends defining preparations by their source material, isolation method and measured markers rather than by a name that implies a biogenesis route that has not been demonstrated [1][2]. Much published work labelled "exosome" is more accurately "small EV preparation".
How they are formed
The exosome pathway is well characterised [3]:
- Endosome formation — the cell membrane folds inward, creating an internal compartment
- Cargo loading — specific molecules are packaged into small vesicles inside that compartment
- Multivesicular body — those internal vesicles accumulate
- Release — the multivesicular body fuses with the cell membrane, releasing its vesicles
- Uptake — the vesicles travel through body fluids and are taken up by recipient cells
Packaging is selective rather than random: cells load cargo according to their state and the signals they are responding to.
What MSC Exosomes Carry
MSC-derived vesicle preparations contain a broad range of bioactive molecules [6][7].
Proteins and growth factors detected in MSC exosome preparations include VEGF, HGF, TGF-β, PDGF, IGF-1, angiopoietin and SDF-1 — factors involved in angiogenesis, tissue repair, immunomodulation and cell migration.
MicroRNAs are the more distinctive cargo. Unlike proteins acting at the cell surface, miRNAs enter recipient cells and regulate gene expression directly. Systems-level analysis of the MSC exosome miRNA landscape has catalogued the species present and their predicted regulatory targets [8]; commonly reported ones include miR-21, miR-29, miR-125, miR-146a, miR-181, miR-199 and the let-7 family.
Lipids — phospholipids, cholesterol, sphingomyelin and ceramides — form the membrane and have signalling roles of their own.
⚠️ An important limit on all of this. Detecting a molecule in a preparation is not the same as showing it does something in a patient. Cargo catalogues are descriptive. The 2019 four-society position statement was explicit that the field suffers from "limited reproducibility of in vitro and in vivo functional assays" [5]— meaning that even the functional experiments behind these catalogues do not reliably replicate between laboratories.
Why This Changed How Stem Cell Therapy Is Understood
This is the genuinely important part of the story, and it is well evidenced.
The original assumption failed
Early stem cell research assumed therapeutic benefit came from engraftment and differentiation — cells travelling to damaged tissue, integrating, and becoming new tissue.
The human data did not support it. Von Bahr and colleagues examined autopsy material from 18 patients who had received HLA-mismatched MSCs, analysing 108 tissue samples by PCR [9]. They found:
- No ectopic tissue formation and no malignant tumours of donor origin on macroscopic or histological examination
- Donor DNA detectable in one or more tissues in 8 patients, at levels of 1/100 to under 1/1,000
- Detection negatively correlated with time — found after 9 of 13 infusions given within 50 days of sampling, but only 2 of 8 given earlier
- No correlation between engraftment and treatment response
Their conclusion: MSCs "appear to mediate their function through a 'hit and run' mechanism", and "the lack of sustained engraftment limits the long-term risks of MSC therapy" [9].
That is human evidence, and it cuts two ways — it undermines the engraftment story and it provides real reassurance about long-term safety.
The paracrine answer
If the cells do not stay, the effect must be in what they release. Caplan — who named mesenchymal stem cells in the first place — argued the point forcefully enough to propose renaming them, on the grounds that their therapeutic action is signalling rather than stem-ness [10].
Lai and colleagues then isolated the active fraction. Working with human embryonic-stem-cell-derived MSC conditioned medium, they showed the cardioprotective activity resided in 50–100 nm particles, and that these particles reduced myocardial ischaemia/reperfusion injury in mice [11].
This is a real and consequential scientific advance. It reframed cell therapy from a transplantation problem into a signalling problem, and it opened the possibility that the signal could be delivered without the cell. That possibility is what the rest of this page is about — and it remains, at present, a possibility.
What the Preclinical Evidence Shows
The findings below are consistent, mechanistically coherent and, in several cases, striking. Every one of them is from animals or cell culture. The species is stated in each case, and again in the reference list.
Osteoarthritis — mice
MSC-derived exosomes and microvesicles protected cartilage and bone from degradation in a murine model of osteoarthritis, reducing cartilage degradation and inflammatory markers [14].
Bone — rats and in vitro
Bone-marrow MSC extracellular vesicles regulated osteoblast activity and differentiation in vitro and promoted bone regeneration in rats [15].
Tendon — rats
Extracellular vesicles from bone-marrow MSCs modulated inflammation and enhanced tendon healing in a rat model [16].
Wound healing — rats
Human umbilical-cord MSC exosomes accelerated cutaneous wound healing in a rat skin burn model, acting through Wnt4 signalling [17].
Stroke — rats
Systemically administered MSC exosomes promoted neurovascular remodelling and functional recovery after middle cerebral artery occlusion in rats [18].
⚠️ This study is also the usual basis for the claim that exosomes cross the blood-brain barrier. What it shows is that systemically given exosomes had effects in the brain after an experimental stroke — a condition in which the barrier is itself disrupted by the injury. Whether exosomes cross an intact human blood-brain barrier in therapeutically meaningful quantities is not established by it.
Heart — mice
MSC exosomes increased ATP levels, reduced oxidative stress, activated PI3K/Akt signalling and limited adverse remodelling after myocardial ischaemia/reperfusion in mice [19].
Inflammation — rats and in vitro
LPS-preconditioned MSCs released exosomes carrying let-7b that shifted macrophage polarisation toward resolution of chronic inflammation, in rats and in cell culture [13].
The Anti-Ageing Work
This deserves its own section because it is the most-cited and most-overstated finding in the area.
Lei and colleagues showed that neonatal umbilical cord is a source of MSC-derived extracellular vesicles (UC-EVs) that "contain abundant anti-aging signals" [12]. Specifically:
- UC-EVs rejuvenated senescing adult bone-marrow MSCs in culture, which then showed alleviated ageing phenotypes, increased self-renewal capacity and increased telomere length
- The mechanism was, at least partly, transfer of PCNA (proliferating cell nuclear antigen) into the recipient cells
- Rejuvenated cells showed enhanced regenerative capacity in bone formation, wound healing and angiogenesis
- Intravenously injected UC-EVs reduced bone and kidney degeneration in aged mice
The authors' own conclusion: UC-EVs "are of high translational value in anti-aging intervention" [12].
What that sentence says and does not say. "High translational value" is a statement about promise. The telomere finding is an in vitro measurement in cultured cells. The organ findings are in aged mice. Neither has been shown in a person, and no trial has tested whether giving a human being exosomes changes any measure of ageing.
Two frailty trials that are frequently misattributed
Two randomised controlled trials in aging frailty are often cited in discussions of exosome anti-ageing effects:
- Tompkins and colleagues (2017) infused allogeneic bone-marrow-derived MSCs intravenously at 100 or 200 million cells in 30 patients, with no therapy-related serious adverse events at one month, and physical-performance improvement preferentially in the 100-million group [24]
- Zhu and colleagues (2024) infused human umbilical-cord MSCs intravenously in a randomised, double-blind, placebo-controlled phase I/II trial, reporting improvement in the SF-36 physical component score at 6 months (p = 0.042) and in EQ-VAS (p = 0.002 at 6 months), with significant improvement on the Timed Up and Go test [25]
Both trials administered cells, not exosomes. Neither measured exosomes, exosome cargo, or any exosome-mediated mechanism. They are evidence about MSC therapy. They cannot be used to support exosome therapy, and the inference that the cells worked because their exosomes worked — however plausible mechanistically — has not been tested in either study.
Exosomes Compared With Cells
Where cells retain clear advantages: they adapt their secretome to the environment they land in, they continue producing factors over time rather than delivering a fixed payload, and — decisively for a patient making a decision — there is far more human trial data on cells than on exosomes.
Where the cell-free approach is genuinely attractive: no viability to preserve, potentially simpler storage and logistics, no possibility of cell-derived proliferation, and a product that is in principle easier to characterise and standardise. Note the words potentially and in principle. The next section is about why they are still required.
What the Field Cannot Yet Do
This section is the one most often left out, and it is the reason there is no approved product.
In 2019, members of four scientific societies — SOCRATES, ISEV, ISCT and ISBT — published a joint position on defining MSC-derived small extracellular vesicles for therapeutic use [5]. Their statement of the problem:
"...discrepancies and controversies about the biology, functions, and potency of MSC-sEVs have arisen due to several factors: the diversity of MSCs and their preparation; various methods of sEV production and separation; a lack of standardized quality assurance assays; and limited reproducibility of in vitro and in vivo functional assays."
Six years later the same problems are named as the live constraints. Lundy and colleagues, reviewing the field's translational position, identify "an incomplete mechanistic understanding, a lack of standardized measurements, and inconsistencies in regulatory classification" [22], and argue the field must move "from descriptive studies into a reproducible and quantitative measurement science."
The practical consequences, in plain terms:
- There is no agreed way to state a dose. Particle count, protein content and marker-based quantification give different answers, and the same source cells processed by different isolation methods yield materially different products
- There is no validated potency assay. Nobody can currently certify that a given preparation will do anything
- Batch-to-batch variability is real, even from a single MSC source
- Isolation method changes the product — ultracentrifugation, size-exclusion chromatography, immunoaffinity capture and tangential flow filtration do not produce equivalent material
The operative international guidance is now MISEV2023 [1], with a dedicated companion addressing relevance to cell and gene therapies [2], superseding the 2018 [2]and 2014 [4]versions.
Regulatory Status
There is no approved exosome product for regenerative medicine indications — not in the United States, not in the European Union.
That is not a formality awaiting paperwork. A systematic analysis of EV regulation across the US, EU and major Asian jurisdictions describes a "fragmented and rapidly evolving regulatory landscape, with significant disparities", and identifies "critical hurdles in quality control, safety evaluation, and efficacy demonstration" [21]. Products of this kind, being substantially manipulated biological material, fall to be regulated as drugs or biologics and require the full trial and licensing pathway.
The path is harder than for cells, not easier. The often-repeated idea that exosomes offer a lighter regulatory route has it backwards: cells are the more established category, with approved products and settled frameworks, while EV classification is explicitly listed among the inconsistencies holding translation back [22].
⚠️ On products offered outside that pathway. The same regulatory review specifically identifies "the emerging direct-to-consumer market for EVs and secretome preparations" as a concern, and frames the purpose of regulatory harmonisation as "to safeguard patient welfare" [21]. That an exosome product is available somewhere is not evidence that it has been evaluated. It is often evidence of the opposite.
What this means for a reader: anyone offering exosome therapy today is offering something unapproved. That may be appropriate in a properly conducted clinical trial with ethical oversight and informed consent. Outside that setting, the honest description is experimental, and the burden is on the provider to say so.
What Is Known About Safety
The preclinical safety record is reassuring, and there are sound reasons to expect a favourable profile: vesicles do not replicate, cannot differentiate, express little surface MHC, and are cleared relatively quickly.
The human safety record for exosome therapy is not established, because the trials have not been done. Reviews of the area describe clinical trials as in progress and note that exosomes "may provide considerable advantages... potentially reducing undesirable side effects" [20]— prospective and conditional language, not a finding.
The nearest thing to reassurance from human data is indirect and concerns cells rather than vesicles: the autopsy study above found no ectopic tissue and no donor-origin tumours in 18 patients [9].
Areas that remain genuinely open: effects of repeated dosing over time; behaviour in immunocompromised patients; what a dose even is; and — because preparations are not standardised — whether safety findings for one product transfer to another at all.
Frequently Asked Questions
What is the difference between exosomes and stem cells?
Stem cells are living cells that self-renew and differentiate. Exosomes are vesicles that cells release to communicate. MSC exosomes carry a large part of the signalling that appears to drive MSC therapy — but they are a fixed payload, where a cell responds to its surroundings and keeps producing.
Are exosomes approved?
No. There is no approved exosome product for regenerative medicine indications in the United States or the European Union, and the global regulatory picture is fragmented and unsettled [21].
Is there human evidence that exosome therapy works?
Very little. The therapeutic literature summarised on this page is preclinical — mice, rats and cell culture. Human trials exist and are in progress [20], but the evidence base is not yet comparable to that for cell therapy, and it is far from what would support routine clinical use.
Are exosomes safe?
The preclinical safety profile is good, and there are mechanistic reasons to expect fewer risks than with living cells. But safety in humans is established by human trials, and those have not been completed. Anyone describing exosome safety as proven is going beyond the evidence.
Why is this field advancing so slowly if the science is good?
Because the science being good is not sufficient. Without an agreed way to measure a dose, a validated potency assay, or reproducible functional testing between laboratories [5][22], regulators cannot evaluate a product and trials cannot be compared. Those are the actual bottlenecks, and they are the subject of active international work [1][2][22].
What should I ask anyone offering exosome therapy?
- Is this an approved product, or is it unapproved?
- Am I being enrolled in a registered clinical trial with ethical approval? If not, what is the basis for the treatment?
- What is the source material, the isolation method, and how is the dose defined?
- What human evidence exists for this specific product in my condition?
Where This Is Heading
Active directions include engineered vesicles loaded with specific cargo, surface modifications intended to confer targeting, and vesicles from defined cell subpopulations. All are research, not practice.
The field's own recent commentary is worth reading plainly. One 2025 analysis maps EV research explicitly onto the Gartner hype cycle — "from initial enthusiasm to growing concerns about reproducibility, standardization and clinical translation" — naming heterogeneity, methodological inconsistency and publication bias as risks that "risk stalling progress" [23]. The proposed remedy is not less ambition but more rigour: standardised reporting, transparent methods, and the MISEV and EV-TRACK infrastructure that supports both [23].
That is what a serious emerging field looks like from the inside. It is not the language of a treatment ready for patients, and it should not be translated into that language.
The Bottom Line
Exosomes are a real biological phenomenon and a genuine scientific advance. They explain something that needed explaining — why infused stem cells help without staying — and the mechanism has held up across many models [9][10][11].
The preclinical results are consistent and in places remarkable: cartilage protection, bone and tendon healing, wound closure, neurological recovery, cardioprotection, and the transfer of measurable rejuvenating signals into aged cells [12][14][15][16][17][18][19]. Those results are in mice, rats and dishes.
What does not yet exist is a way to make a consistent product, a way to measure its dose or potency, a body of human trial evidence, or an approved therapy in any major jurisdiction [5][21][22]. The field says so about itself [22][23].
Promising preclinical science with no approved products and minimal human data is an honest description, and a genuinely interesting one. It is not a reason to dismiss exosomes. It is a reason not to buy them yet.