How stem cells actually work — replace, renew, signal.

Most articles on stem cells skip the part about how they actually work. They jump from "stem cells are amazing" straight to "stem cells fix things." The mechanism in between gets glossed over because it used to be straightforward — and isn't anymore.

Here's what's actually happening in the body when stem cells are introduced. There are three modes. Two of them work the way you'd expect. The third turned out to be the most important — and it took the field about a decade to fully accept that.

Mode one — replace

The original story, and the one most people still default to: stem cells differentiate. They become the tissue you need. Cartilage damage? They turn into chondrocytes. Muscle injury? Myocytes. Brain damage? Neurons.

This was Arnold Caplan's original framing when he first named "mesenchymal stem cells" in 1991. The premise was elegant — give the body more of these multipotent cells, and they'll fill in for the damaged ones.^[1]

And it does happen. Sort of. In small amounts. But not nearly to the extent the original theory predicted. When researchers tracked transplanted MSCs in the body, they found the cells often didn't survive long, didn't reach the injury site in large numbers, and didn't differentiate into target tissue at the rate needed to explain the clinical improvement patients were experiencing.

Something else was going on.

Mode two — renew

The second mode is self-renewal: stem cells can divide to make more of themselves while also producing differentiated daughter cells. Think of it as both replacing tissue AND keeping the stem cell pool topped up.

For tissue maintenance over time, this matters. It's why stem cell populations in the body can keep functioning across decades — even though, as we've covered elsewhere in the Journal, those populations decline significantly with age.^[2]

But again, this didn't fully explain what was happening clinically. Patients were getting better faster than the cell-replacement and self-renewal models could account for.

Mode three — signal

Here's the breakthrough. Stem cells release a constant stream of biological signals — growth factors, cytokines, chemokines, and small lipid-bound packages called exosomes. These signals travel to surrounding cells and tell them what to do.

This is called paracrine signalling. "Para" meaning "near," "crine" meaning "secretion" — secretion to nearby cells. And it turns out this is doing most of the work.

In a 2011 review titled "The MSC: an injury drugstore," Caplan and his collaborator Diego Correa proposed that MSCs primarily function as a kind of mobile pharmacy: arriving at injury sites and releasing the right molecules to calm inflammation, stimulate local repair, and modulate the immune response.^[3] The cells aren't replacing damaged tissue so much as orchestrating its repair.

By 2017, Caplan went further. In a paper called "Mesenchymal Stem Cells: Time to Change the Name!" he formally proposed renaming MSCs to "Medicinal Signaling Cells." Same acronym. Different mechanism.^[4] The argument: the field had been holding on to a name and a model that the data no longer supported.

Why this matters clinically

Two practical implications.

First — dose and timing logic shifts. If the cells are signalling rather than replacing, you don't necessarily need them to engraft permanently. You need them to be metabolically active long enough to broadcast their signals. That changes how clinicians think about cell counts, infusion intervals, and adjunct therapies.

Second — the immunomodulatory effect becomes central. A lot of MSC clinical wins — graft-versus-host disease, autoimmune conditions, inflammatory osteoarthritis — make more sense through the signalling lens. The cells are sensing inflammation and switching it off, not replacing inflamed tissue.^[5]

This also explains why allogeneic (donor) cells work even though they're not "yours." If the primary mechanism is signalling, and the cells don't need to permanently engraft, donor cells can do their work and leave. The signal is what matters.

The research arc

The field went from "stem cells become the tissue you need" (1990s) to "stem cells help repair via multiple mechanisms" (2000s) to "stem cells are mostly signalling cells that happen to have multipotent ability" (2010s onwards). It's a more nuanced model — and a more accurate one.

It's also why language matters. "Stem cell therapy" the marketing phrase implies replacement. "Cell-mediated paracrine therapy" — what's actually happening — sounds less impressive but is closer to true.

What this means for treatment

If you're considering stem cell therapy, the signalling model is good news for two reasons:

1. Lower bar for benefit. The cells don't need to engraft permanently to help. They need to arrive, signal, and then can clear naturally.
2. Wider applicability. Conditions that are fundamentally inflammatory — joint pain, autoimmune flare-ups, age-related "inflammaging" — are exactly the kind of thing immunomodulatory signalling can affect.

It's also a reason to be sceptical of clinics that still talk about MSCs in pure replacement terms — "the cells will become new cartilage" — without nuance. The field has moved on. The marketing in some places hasn't.

Sources

1. Caplan AI. (1991) "Mesenchymal stem cells." J Orthop Res 9(5):641-50.
2. Stolzing A, Jones E, McGonagle D, Scutt A. (2008) "Age-related changes in human bone marrow-derived mesenchymal stem cells." Mech Ageing Dev 129(3):163-73.
3. Caplan AI, Correa D. (2011) "The MSC: an injury drugstore." Cell Stem Cell 9(1):11-15.
4. Caplan AI. (2017) "Mesenchymal Stem Cells: Time to Change the Name!" Stem Cells Transl Med 6(6):1445-51.
5. Bernardo ME, Fibbe WE. (2013) "Mesenchymal stromal cells: sensors and switchers of inflammation." Cell Stem Cell 13(4):392-402.