The science
Multilineage-differentiating Stress-Enduring cells: an endogenous, pluripotent-like subpopulation identified in adult connective tissue. Not engineered, not synthetic — already present, and rare.
Origin
In plain terms
Muse cells are a rare, naturally occurring type of stem cell found inside ordinary bone marrow, blood and connective tissue — everyone already has some. Researchers first described them in 2010. What makes them interesting: in lab and animal studies they appear to travel toward injured tissue on their own and turn into whatever cell type is needed there, without triggering immune rejection — even from an unmatched donor.
Muse cells were first described in 2010, in Proceedings of the National Academy of Sciences (Kuroda Y, Kitada M, Wakao S, et al., 2010;107:8639–43). Isolation and expansion protocols followed in Nature Protocols in 2013, and the population has since been reported by independent groups across several mammalian species.
They are identified as SSEA-3-positive cells in bone marrow, peripheral blood and the connective tissue of various organs, with a reported diameter of roughly 13–15 µm.
Why the marker travels across species
SSEA-3 is a surface glycolipid rather than a protein encoded by a gene sequence, so the same antibody detects the population in different mammals without redesign — confirmed in 2026 work on canine and feline tissue.
Signature
The table below uses the shorthand researchers use for these markers. In short: Muse cells are identified by one defining marker (SSEA-3), and alongside it they carry genes tied to flexibility — the ability to turn into different cell types — plus markers shared with ordinary mesenchymal cells, and an immune profile that appears to help them avoid rejection.
| Feature | Reported |
|---|---|
| Defining marker | SSEA-3 (Stage-Specific Embryonic Antigen-3) |
| Pluripotency genes | Oct3/4, Sox2, Nanog, TERT — upregulated in suspension culture, where the cells aggregate into ES-cell-like M-clusters |
| Differentiation | Endodermal (SOX17, AFP), mesodermal (GATA2, DESMIN, SMA) and ectodermal (NESTIN, NF) marker-positive lineages |
| Mesenchymal markers | CD105, CD90, CD73 |
| Immune profile | HLA-G positive, IDO positive, HLA-DR negative |
| Homing receptor | S1PR2 — sphingosine-1-phosphate receptor 2 |
| Stress tolerance | Tolerates genotoxic stimuli better than mesenchymal stromal cells (Alessio et al., Oncotarget 2018;9:19328) — the property the name derives from |
Mechanism
These three properties are the actual scientific case for Muse cells — each shown in lab or animal studies, not yet confirmed the same way in people.
Homing
Damaged tissue releases a distress signal, a molecule called S1P. Muse cells appear to sense that signal and move toward it after being infused into the bloodstream. Shown specifically for heart tissue by Yamada et al., Circulation Research 2018;122:1069–83.
Differentiation
At the injury site, the cells absorb debris from dying cells — and that appears to act as a local signal telling them what to turn into. Shown for nerve-cell formation by Uchida et al., Stroke 2017;48:428–35.
Immune profile
Across the published human trials, patients received donor cells with no matching and no anti-rejection drugs, and tolerated it well — which would make this kind of therapy much simpler to deliver at scale, if the finding holds up in larger studies.
Evidence
In plain terms: there have been six published human studies of one specific manufactured Muse cell product, adding up to about sixty-seven patients total. That's a small, early body of evidence — enough to start learning about safety, not enough to prove the treatment outperforms existing options. No study has reached the large, late-stage size (called Phase 3) that regulators require before approving a therapy.
All published human efficacy and safety data comes from trials of a single allogeneic bone-marrow-derived product. Six trials have been published, enrolling roughly sixty-seven patients in total. No indication has reached Phase 3.
| Indication | n | Reported result |
|---|---|---|
| Ischaemic stroke | 35 | Randomised, double-blind, placebo-controlled — the only controlled Muse cell trial. Its primary endpoint was safety, not efficacy. 40% of treated patients reached mRS ≤2 at 12 weeks against 10% on placebo, but that 30-point gap was not statistically significant (95% CI −7.8 to 65.2). At one year the between-group difference remained around 30 points. |
| Cervical spinal cord injury | 10 | Phase 1, open-label, single IV dose. Improvements in motor scores, activities of daily living and quality of life. Authors note a control arm is needed to establish efficacy. |
| Neonatal HIE | 9 | Dose-escalation. Safe and tolerable; normal developmental scores in 67% of participants. |
| Amyotrophic lateral sclerosis | 5 | Open Phase 2, six monthly infusions. Safe across 12 months; slower ALSFRS-R decline in 3 of 5, reported as non-significant. |
| Dystrophic epidermolysis bullosa | 5 | Phase 2, open-label, single-arm. The sponsor announced the efficacy endpoint was met; the peer-reviewed report of the same trial gives a week-4 ulcer reduction whose confidence interval spans zero, with ulcer size back to baseline by week 12. |
| Acute myocardial infarction | 3 | First-in-human exploratory. Left ventricular ejection fraction improved from approximately 41% to 52% at three months, without arrhythmias. Safety was the primary endpoint; LVEF was secondary. |
Read this alongside the table
Study sizes range from just three to thirty-five patients, and the stroke study is the only one with a placebo comparison group — every other trial compared patients only against their own starting condition. Follow-up beyond twelve months is limited. The company that ran these trials stopped developing the product in February 2023.
A seventh trial, in COVID-19 ARDS, was registered and started but has never published its results. It is not in the table above because there is nothing to report.
A safety record built on one specific manufactured product doesn't automatically carry over to a different one. A differently made preparation has its own, separately unproven safety profile.
Safety
In plain terms: across every published study, nobody developed a tumour from these cells, and nobody's body rejected them — even patients who received cells from an unmatched donor, with no anti-rejection medication. That's a genuinely good safety signal, within the limits described below.
No tumour or teratoma formation has been reported across published Muse cell studies, and no immune rejection events have been reported in HLA-mismatched allogeneic recipients. No published trial used immunosuppression.
Two properties underlie this. Muse cells are endogenous and have intrinsically self-limited proliferation, unlike induced or embryonic pluripotent cells. And the immune-privileged phenotype permits allogeneic delivery without the separate risk profile that accompanies immunosuppressive regimens.
Three boundaries on that record
Small numbers. Sixty-seven trial patients isn't enough to catch a rare side effect. Nobody has studied this in a large population over a long time yet.
One specific product. This safety record belongs to the exact preparations that were studied — not automatically to “Muse cells” as a whole category.
It isn't a completely clean sheet. The spinal cord injury trial did record two serious adverse events; both were judged unrelated to treatment. “Nothing was caused by the treatment” is a different statement from “nothing happened.”
