Discovery and definition
In 2010, Mari Dezawa's laboratory at Tohoku University reported something unexpected hiding inside two of the most ordinary cell preparations in regenerative medicine: cultured skin fibroblasts and bone marrow stromal cells. When Kuroda and colleagues subjected these cultures to severe cellular stress, including long incubation in the digestive enzyme trypsin, most cells died. A small fraction survived, and that surviving fraction behaved unlike anything expected of mesenchymal tissue. Single surviving cells formed clusters resembling embryonic stem cell colonies, expressed pluripotency-associated genes, and gave rise to cells with characteristics of all three germ layers: ectoderm, mesoderm, and endoderm.[1]
The group named the population multilineage-differentiating stress-enduring cells, or Muse cells, after the property that revealed them: they endure stress that kills their neighbors. Practically, Muse cells are identified and sorted by a surface marker, stage-specific embryonic antigen-3 (SSEA-3), a glycolipid otherwise associated with human embryonic stem cells. They constitute a small percentage of cultured mesenchymal stromal cells and a far smaller fraction of fresh bone marrow, on the order of one in several thousand mononuclear cells.[1]
A follow-up study from the same group the next year added a provocative finding: when human fibroblasts were reprogrammed into induced pluripotent stem (iPS) cells using the standard Yamanaka factors, the iPS colonies arose almost exclusively from the Muse fraction, suggesting that Muse cells are the subpopulation within ordinary tissue most primed toward pluripotency.[2] That claim remains associated mainly with Dezawa's group and has been debated, but it framed the way the field has understood these cells since: not as a new cell type engineered in a dish, but as a pre-existing, endogenous population that ordinary mesenchymal cultures already contain.[3]
- In one sentence
- A Muse cell is a rare, stress-tolerant, SSEA-3-positive cell found within ordinary mesenchymal tissue that can differentiate into cell types of all three germ layers from a single cell, without forming teratomas.
Defining properties
Three properties, taken together, distinguish Muse cells from both their parent population and from engineered pluripotent cells. The first is broad differentiation capacity. Standard mesenchymal stromal cells, described in depth in this journal's MSC explainer, reliably form bone, cartilage, and fat: mesodermal tissues. Muse cells, in the published work, additionally generate cells with hepatocyte, neuronal, and epidermal characteristics, which belong to the endodermal and ectodermal lineages, and they do so clonally, from a single cell.[1] Reviews of the field treat this triploblastic capacity, verified by marker expression and in vivo integration rather than by full organ formation, as the population's defining feature.[3]
The second is non-tumorigenicity. Pluripotency normally carries a price: embryonic stem cells and iPS cells form teratomas when transplanted, which is why raw pluripotent cells are never injected into patients. Muse cells, despite their pluripotency-like differentiation, did not form teratomas in the testes of immunodeficient mice over six months of observation in the original report. The proposed explanation is instructive: Muse cells show low telomerase activity, closer to somatic fibroblasts than to embryonic stem cells, meaning they lack the unlimited replicative drive that makes true pluripotent cells dangerous.[2] They appear to occupy a middle position: broader potential than a standard MSC, without the proliferative machinery of an iPS cell.
The third is the property in the name: stress endurance. Muse cells survive prolonged protease exposure, hypoxia, and serum deprivation that eliminate other cells, which is both the laboratory method for enriching them and, in the field's working model, the reason they might function as a reservoir for tissue repair: the cells built to survive injury are the ones present at the site of it.[3]
The comparison with iPS cells is worth making explicit, because both are described with the word pluripotent and the practical differences are large. An iPS cell is manufactured: a somatic cell is genetically reprogrammed, expanded, differentiated into the desired cell type, and screened for residual undifferentiated cells before any clinical use, a process that takes months and carries genomic-integrity risks the field manages with extensive testing. A Muse cell is selected rather than made: it is sorted from tissue that already contains it, requires no reprogramming step, and in the published protocols is administered as-is, relying on the damaged tissue itself to direct differentiation.[3] The trade is capability for simplicity. An iPS-derived product can be driven to a specific, mature cell type at scale; a Muse infusion delivers a smaller number of less committed cells and depends on in vivo cues that remain incompletely characterized.
One further characteristic matters for how the clinical program was designed. Muse cells express HLA-G, an immunomodulatory molecule associated with the placenta's tolerance of the fetus. Partly on this basis, every Muse trial to date has administered donor cells intravenously without HLA matching and without immunosuppressive drugs, and the trials have reported that the cells were tolerated under those conditions over their observation periods.[4] Whether donor Muse cells nonetheless provoke slower immune clearance over longer horizons is one of the open questions the small trials cannot settle.
Homing and engraftment biology
The most cited practical argument for Muse cells over bulk MSCs concerns what happens after an intravenous infusion. Standard MSCs infused into a vein are, in most published tracking studies, trapped predominantly in the lung's capillary bed within minutes, and comparatively few reach injured tissue; much of the benefit attributed to MSC infusions is therefore ascribed to secreted factors rather than to the cells arriving and staying, a point discussed in the MSC explainer. Muse cells appear to behave differently, and the proposed mechanism is specific.
Damaged tissue releases sphingosine-1-phosphate (S1P), a lipid signaling molecule that rises at sites of cellular injury. Muse cells express the receptor S1PR2, and in a rabbit model of acute myocardial infarction, Yamada and colleagues showed that intravenously infused Muse cells migrated selectively to the infarcted heart along this S1P gradient. Roughly 14.5 percent of the injected cells were estimated to be engrafted in the heart at day 3, a high figure by the standards of intravenous cell delivery, and the homing was abolished when the S1P-S1PR2 axis was blocked pharmacologically or genetically. Engrafted cells differentiated into cardiomyocyte-marker-positive and vessel-associated cells, infarct size fell by around half versus vehicle, and functional improvement persisted for months.[5]
The same group has since taken the logic one step further: if injured tissue recruits Muse cells through S1PR2, a drug stimulating that receptor might mobilize a patient's own endogenous Muse cells without any infusion at all. In rabbits, an S1PR2 agonist increased circulating Muse cell numbers after myocardial infarction, reduced infarct size, and improved ventricular function.[6] This remains entirely preclinical, but it illustrates where the field believes the biology points: toward repair driven by targeted arrival at damaged tissue rather than by transient signaling from cells lodged in the lung.
A caution belongs here. These homing and engraftment figures come from animal models, largely from laboratories associated with the cells' discoverers, and equivalent cell-tracking data in humans do not exist. The human trials described next measured clinical outcomes, not engraftment.
The clinical trial program
Muse cells are unusual among newer cell types in having an organized clinical development program. A Japanese company, Life Science Institute, manufactured a standardized allogeneic Muse cell product designated CL2020, derived from donor bone marrow, and sponsored a series of investigator-led trials across strikingly different diseases, all using intravenous infusion without HLA matching or immunosuppression, at doses well below the typical mesenchymal cell dose and route reported in registered trials.[4] The individual studies deserve to be described at their actual size.
Acute myocardial infarction. The first-in-human study enrolled 3 patients with ST-elevation myocardial infarction and reduced ejection fraction after successful stenting. Each received a single intravenous dose of 15 million cells. No adverse drug reactions were reported, and mean ejection fraction improved from roughly 41 percent to 52 percent over 12 weeks.[7] With three patients and no control group, the safety observation is the meaningful result; the efficacy signal, however encouraging, cannot be separated from natural recovery after reperfusion. A larger randomized trial in this indication was subsequently initiated.
Subacute ischemic stroke. The stroke study is the largest and methodologically strongest Muse trial published to date: a randomized, double-blind, placebo-controlled trial in 35 patients, 14 to 28 days after stroke with significant residual disability, allocated roughly 2.5 to 1 between a single CL2020 infusion and placebo. The primary endpoint was safety through week 12. On the prespecified efficacy measure, 40 percent of treated patients reached a modified Rankin Scale score of 2 or better at week 12, versus 10 percent of placebo patients, and the lower confidence bound exceeded a threshold derived from registry data. Adverse events were nearly universal in both arms, as expected in a stroke population, with adverse reactions in 28 percent of the treated group including one grade 4 seizure.[8] This is a genuinely positive controlled signal, and also a 35-patient study whose result requires confirmation at phase 3 scale.
Dystrophic epidermolysis bullosa. Preclinical work showed that infused Muse cells homed to the damaged skin of a mouse model of this inherited blistering disease and produced type VII collagen, the protein the disease lacks.[9] The subsequent phase 1/2 open-label study treated 5 adult patients with a single infusion. Adverse effects were mild and self-limited, and ulcer size and pain scores improved, but a skin biopsy did not demonstrate increased type VII collagen or new anchoring fibrils, so the mechanism the preclinical work predicted was not confirmed in patients.[10]
Cervical spinal cord injury. A trial in 10 patients with traumatic cervical injury administered a single 15-million-cell infusion. Tolerability was good; two severe adverse events were judged unrelated to the cells. Upper-extremity motor scores improved significantly against baseline through 52 weeks, but the study had no control arm, and some motor recovery after incomplete cervical injury is expected without any intervention.[11]
Amyotrophic lateral sclerosis. A phase 2 open-label study gave 5 ALS patients monthly infusions for six months, the first repeated-dose Muse protocol. The regimen was tolerated without serious adverse effects attributed to the cells, and functional rating scores remained stable in most patients for around ten months.[12] In a disease as variable as ALS, five uncontrolled patients cannot establish that the stability was caused by the treatment.
Read together, the program supports two statements with reasonable confidence: single and repeated intravenous doses of allogeneic Muse cells, given without matching or immunosuppression, were well tolerated in roughly sixty published trial participants across five diseases, and one small randomized trial in stroke produced a positive efficacy signal.[4] Everything beyond those two statements remains open.
Current clinical status and limitations
No Muse cell product is approved by any major regulator. CL2020 has not received marketing authorization in Japan, the United States, or Europe, and the published human evidence consists of the early-phase studies described above: one randomized trial of 35 patients and a handful of open-label studies of 3 to 10 patients each. The honest summary is that Muse cells are a scientifically interesting cell population with a credible mechanistic story, a consistent early safety record, and an efficacy case that rests, so far, on one small controlled trial and several uncontrolled ones.
The limitations deserve equal weight. Much of the foundational biology, including the pluripotency claims and the homing quantification, originates from the discovering laboratory and its collaborators, and independent replication outside that network is thinner than the citation counts suggest. Human engraftment has never been directly measured. The epidermolysis bullosa trial failed to confirm its predicted mechanism in patient tissue. And the two most encouraging efficacy results, in stroke and myocardial infarction, both await the larger randomized trials that would ordinarily precede any approval. It is also worth stating what a definitive program would require: a phase 3 trial powered for a clinical endpoint, run across multiple centers with investigators independent of the product's originators, with prespecified analysis and long-term follow-up. Nothing published on Muse cells yet meets that description, which is normal for a cell type sixteen years from discovery and is also the reason firm efficacy claims would be premature.
Within that context, TrueCell offers Muse cell protocols in México as physician-directed treatments under the Mexican regulatory framework, with candidacy and dosing determined case by case after medical review; the Muse treatment page describes the protocol and the treatments hub places it alongside the clinic's other programs. Patients considering this option should understand that they are choosing a therapy whose evidence base is early, and any provider unwilling to say so plainly is not describing the field accurately. Because Muse cells are isolated from mesenchymal tissue by SSEA-3 sorting, the questions worth asking a provider parallel those in our cell-sourcing article: where the source tissue comes from, how the Muse fraction is verified, and what viability and identity testing accompanies each dose.
Common questions
How do Muse cells differ from regular stem cells?
Muse cells are a small subpopulation found inside ordinary mesenchymal stromal cell preparations, identified by the marker SSEA-3. They differ from the bulk population in three documented ways: they can generate cells of all three germ layers from a single cell rather than only bone, cartilage, and fat; they survive severe stress that kills other cells; and in animal models they home to damaged tissue through the S1P-S1PR2 axis instead of lodging mainly in the lungs.[1][5] Unlike iPS or embryonic stem cells, they have not formed teratomas in published transplantation studies, consistent with their low telomerase activity.[2]
Is Muse cell therapy proven?
No. There is no approved Muse cell product anywhere, and the human evidence comes from small early-phase trials: one randomized study of 35 stroke patients with a positive result, and open-label studies of 3 to 10 patients in heart attack, epidermolysis bullosa, spinal cord injury, and ALS.[7][8][12] Safety across those studies has been consistently good, which is meaningful, but efficacy claims beyond the stroke signal are not yet supported by controlled data.
How does TrueCell use Muse cells?
TrueCell offers Muse cell protocols in México under the COFEPRIS regulatory framework, as physician-directed treatment rather than as part of a registered trial. A board-certified physician reviews each patient's history, imaging, and laboratory results before determining candidacy and dosing, and the consultation includes a plain statement of what the evidence supports and what it does not. Details are on the Muse treatment page.