A cell first found in bone marrow.
The mesenchymal stem cell was not discovered in a search for a therapy. In the 1960s and 70s the Soviet scientist Alexander Friedenstein noticed that a small, rare fraction of bone-marrow cells would stick to plastic culture dishes, grow into colonies, and, under the right conditions, turn into bone and cartilage. These were plainly not the blood-forming (hematopoietic) stem cells everyone was studying at the time. They were something else: a stromal, connective-tissue-forming population.
Two decades later, Arnold Caplan formalized and popularized the term "mesenchymal stem cell" for this population, framing it as a multipotent progenitor capable of becoming the tissues of the musculoskeletal system: bone, cartilage, tendon, ligament, muscle, marrow stroma, and fat.[1] That framing shaped a generation of research and the entire commercial regenerative-medicine field that followed.
Since then, cells with the same defining behavior have been isolated from far more than bone marrow: adipose (fat) tissue, umbilical cord (Wharton's jelly), placenta, dental pulp, and synovium, among others. They are, in the loose clinical sense, everywhere connective tissue is. What they are for, and what they can be relied upon to do once injected, is a more careful question than the confident name suggests.
- In one sentence
- A mesenchymal stem cell is a plastic-adherent, self-renewing stromal cell that can differentiate into bone, cartilage, and fat in the laboratory — but whose main therapeutic action appears to be signaling to other cells rather than replacing them.
What has to be true to call it one.
Because "MSC" was being applied inconsistently across hundreds of labs, in 2006 the International Society for Cellular Therapy (ISCT) published a set of minimal criteria. They remain the reference standard for what may legitimately be called a mesenchymal stromal cell.[2] There are three:
- Plastic adherence. Under standard culture conditions the cells must adhere to tissue-culture plastic — Friedenstein's original observation, now a gatekeeping test.
- A specific surface-marker profile. At least 95% of the population must express CD105, CD73, and CD90, while lacking (≤2%) the hematopoietic and immune markers CD45, CD34, CD14/CD11b, CD79α/CD19, and HLA-DR. This is what separates them from blood and immune cells.
- Trilineage differentiation. The cells must be able to become osteoblasts (bone), adipocytes (fat), and chondroblasts (cartilage) when pushed to in vitro.
It is worth being precise about what this definition does and does not establish. It sets a floor for identity and reproducibility so that two labs mean the same thing by "MSC." It says nothing, on its own, about potency, clinical effect, or which secreted factors a given batch produces. Two preparations can both satisfy the ISCT criteria and still behave quite differently in a patient — a point the field has spent the last decade taking seriously.
The paracrine hypothesis.
The original, intuitive story was straightforward: injected MSCs travel to damaged tissue, engraft, and differentiate into replacement cells — new cartilage for a worn joint, new neurons for a damaged nerve. It is a compelling picture. It is also, for most indications, not what the evidence shows happens.
When researchers tracked labeled cells after injection, they found that the great majority do not durably engraft. Many are cleared within days, and a large fraction of intravenously delivered cells are trapped in the lungs on first pass. And yet functional benefit is still observed. That paradox reshaped the field's understanding of the mechanism.[3]
The cells appear to work less like replacement parts and more like a temporary pharmacy — releasing signals that instruct the tissue already present to behave differently.
This is the paracrine hypothesis: the therapeutic effect is driven mainly by what the cells secrete, not by what they become. That secretome is a complex mixture — growth factors, cytokines, and extracellular vesicles (exosomes) — that can dampen inflammation, promote local blood-vessel formation, reduce programmed cell death in stressed tissue, and recruit the body's own resident repair cells.[4] Caplan and Correa memorably described the injected MSC as behaving like an injury drugstore, dispensing site-specific factors in response to the local environment it lands in.[4]
This distinction is not academic hair-splitting. It reframes what a realistic outcome looks like. If the mechanism were cell replacement, one would expect visible tissue regrowth on imaging. If the mechanism is signaling, the expected outcome is a shift in the tissue environment — less inflammation, better function, slower degradation — which is exactly the pattern the better clinical trials report.
Talking to the immune system.
Perhaps the most clinically consequential property of MSCs is their effect on immunity. In inflamed tissue they release mediators — among them prostaglandin E2, indoleamine 2,3-dioxygenase (IDO), TGF-β, and nitric oxide — that suppress the proliferation of activated T cells, dampen inflammatory macrophages, and nudge the local immune response toward tolerance and resolution rather than chronic activation.[5]
Two features make this practically important. First, the effect is conditional: MSCs are largely inert until "licensed" by an inflammatory environment, which is part of why they are generally well tolerated. Second, their own low immunogenicity — modest HLA class II expression and a lack of the co-stimulatory signals that normally trigger rejection — is what makes allogeneic (donor-derived) therapy feasible at all. It is the reason a well-screened umbilical-cord cell line can be given to an unrelated recipient without tissue matching, the model most modern clinics use.
This immunomodulatory capacity is also why MSCs are studied well beyond orthopedics — in graft-versus-host disease, autoimmune conditions, and inflammatory organ injury. The mechanism that quiets an inflamed knee is, in principle, the same one being tested against systemic inflammation.
Why the name may be wrong.
Here is the uncomfortable part, and it comes from the person who coined the term. In 2017 Arnold Caplan published a paper titled, plainly, "Mesenchymal Stem Cells: Time to Change the Name!"[6] His argument: the phrase "stem cell" implies the cells routinely differentiate into replacement tissue in the body, which — as the tracking studies showed — is mostly not how they help. The word oversells the mechanism and, worse, fuels marketing claims the science does not support.
Caplan proposed renaming them "medicinal signaling cells" — keeping the initials MSC while relocating the emphasis to where the evidence actually points: signaling. The ISCT, for its part, has recommended reserving "mesenchymal stromal cell" (not stem cell) for the bulk, unfractionated populations used clinically, restricting "stem cell" to preparations where genuine stem-cell activity has been demonstrated.[7]
Why should a patient care about a terminology dispute? Because the name sets the expectation. "Stem cell therapy" invites the belief that damaged tissue will be regrown. The honest, evidence-based promise is different and more modest: a well-characterized dose of signaling cells can, in the right candidate, meaningfully reduce inflammation and improve function — and in doing so buy time, defer surgery, and improve quality of life. That is a real benefit. It is simply not the miracle the label sometimes implies, and a serious clinic should be the first to say so.
If you want to see how this mechanism translates into a specific condition, the companion pieces on arthritis and neuropathy walk through the trial evidence, candidacy, and realistic timelines. For where these cells come from and why cord tissue is favored, see the article on cell sourcing.