The Cell Isn't the Cure—Stem Cells Don't Become Tissue; They Speak to the Host

A patient's knee cartilage has worn away, so a doctor injects stem cells. The familiar picture goes like this. The injected cells settle at the site of injury, grow into the worn cartilage cells, and fill the gap with new tissue. A broken part is swapped for a fresh one. The stem cell is a universal regenerative part that can become anything, and the injection is the delivery of that part to the site of damage.

First, a few terms. A stem cell is one that can transform into several kinds of cells and copy itself. That transformation (a stem cell turning into a specific tissue cell such as bone, cartilage, or nerve) is called differentiation. And when an injected or transplanted cell settles in the body, survives, and does its job, that is engraftment. If a transplanted tree's taking root and surviving counts as engraftment, a sapling's growing into a particular species is differentiation. The common picture leans on these two words: the injected stem cell engrafts, then differentiates.

But stain the cells with a fluorescent dye or tag them with a genetic marker and track them through the body, and in most regenerative injections this picture falls apart. The injected cells vanish within days. And yet, in the cases where there is an effect, the patient improves. If the cells are gone and the effect remains, then whatever did the healing wasn't what the cells "became"—it was the result of the signals they released into their surroundings during their brief stay. The cell does not "become" tissue; it "speaks to" the house next door.

When a cell sends chemical signals only to the cells right beside it, that is called paracrine signaling. If a hormone circulating through the bloodstream—endocrine signaling—is a broadcast to the whole country, then paracrine signaling is a shout over the neighbor's fence. This piece asks one question. If stem cell therapy truly heals, is it the cell that heals, or the words the cell speaks?

The Cell That Compounds a Drug

One way toward this question is to change the question itself. Instead of asking what the cell "becomes," ask what the cell "releases." That means reading regenerative medicine in pharmacology's language, not transplant surgery's. A transplant swaps out a part, and its success or failure turns on whether the new part settles in and lives long—on whether it engrafts. A drug is different. A drug is not given so that it will stay in the body and become tissue. It acts briefly, passes on a signal, is metabolized, and disappears. A drug's effect is set not by how long it lingers in the body but by what it delivered.

The strongest evidence that this lens is no forced analogy is that the person who built the field picked up the lens himself. The workhorse of the commercial "stem cell injection" is the mesenchymal stem cell, or MSC, drawn from bone marrow or fat.

Why did this cell get the name "stem cell"? Grow it in the lab and the name makes sense. Take the cell pulled from bone marrow or fat, culture it, shift the conditions, and it turns into bone cells, into cartilage cells, into fat cells. One kind of cell shape-shifting into three separate tissues—this is the very picture a textbook holds up to say "this is what a stem cell is." The cultured MSC performed that trick, and so "stem" went into its name.

The trouble is that the lab and the body are not the same. Shape-shifting in culture is one thing; settling into the body, taking hold, and actually rebuilding that tissue is another. And as we'll see, the MSC that enters the body mostly does not do the latter—it neither takes hold nor becomes tissue, and within days it is gone. The name "stem cell" was earned in the lab, not confirmed in the body.

In 1991, Arnold Caplan first named the cell. And in 2017, Caplan wrote this about the name he had coined: to call the cell a "stem cell" invites misunderstanding, both scientifically and clinically. These are not cells that differentiate into tissue and rebuild the body; they are cells that travel to the site of injury, regulate immunity, and secrete a range of substances. So he proposed changing the name. Keep the abbreviation MSC, but change what it stands for to "Medicinal Signaling Cell."

He did not discard the name; he swapped out its meaning. Earlier, in 2011, Caplan had already called the cell an "injury drugstore." When damage occurs, the MSC is released from its perivascular niche and activated; it regulates the local immune response and secretes bioactive molecules to build an environment favorable to regeneration. Not a brick that stacks into tissue, but a local drugstore that compounds and dispenses medicine on the spot. In fact, in 2006 Caplan had defined the cell as a "trophic mediator" that secretes cytokines and growth factors, fixing the mechanism as paracrine long before he proposed the new name.

Of course this is only one researcher's renaming proposal, and the field still uses the abbreviation "MSC." So this piece does not rest on Caplan's word. The renaming is not an argument but a summary. It is simply the event of someone who built a field over more than thirty years publicly conceding that the picture he first drew (the stem cell becomes tissue and rebuilds the body) does not match the data. The data is where we have to start.

Track the Cells, and They're Gone

There are several ways to track living cells inside the body. Put a fluorescent protein into a cell to make it glow, label it with a radioactive isotope, or attach a genetic marker, and count how many are alive and where, hour by hour. The studies that tracked MSCs this way reach a conclusion that runs head-on against the common picture.

One study injected 500,000 MSCs, tagged with fluorescent and radioactive labels, into the tail vein of mice, then opened the body at five minutes, one hour, one day, and three days to count the living cells. Living MSCs were detected in the lungs for at most one day. After that they were gone. Through that first day, no living MSCs turned up in the blood, nor in the liver, spleen, kidney, or bone marrow. The paper's title wrote the conclusion plainly: they are short-lived and do not migrate beyond the lungs.

There is a reason cells put into a vein get caught in the lungs. A cell injected intravenously passes through the heart and must first cross the lungs' capillary mesh, and the MSC is too large for that mesh, so many are trapped on the first pass. The larger the cell, the more get caught. This lung trapping is confined to intravenous injection, but the study notes that the number of cells clearing the lungs to reach the target tissue in the acute phase is too low to expect a therapeutic effect. The time to clearance varies by route of administration and by study, but it generally falls within a day, or a few days at most.

This is not to say that not a single cell remains, or that MSCs can never differentiate. How many cells persist and how much they differentiate depends on where and how they are placed. Injected directly into tissue, or transplanted locally on a scaffold, they last longer than with intravenous injection. But in the regenerative injections placed into a joint or a vein, culture-expanded MSCs barely engraft and disappear quickly—the consistent finding across many tracking studies. The rare cells observed to settle into tissue and differentiate are far too few to account for the functional recovery actually documented. So the issue is not that no cells differentiate. It is that the healing does not come from differentiation.

The cells are gone, and yet in some cases a myocardial infarction heals and inflammation subsides. The cells are absent, but the effect remains. So what does the healing?

Then What Does the Healing?

One experiment answers this cleanly. When two million human MSCs were put into the vein of a mouse, many lodged in the lungs. So far, this is the same disappearance story as before. But those cells trapped in the lungs released large amounts of the anti-inflammatory protein TSG-6. In myocardial infarction, mice given these MSCs intravenously showed reduced inflammation and infarct size and improved heart function. The decisive part is the control. When MSCs suppressed so they could not make TSG-6 were given, the cells lodged in the lungs just the same, yet the effect vanished. What improved the heart was not the cells trapped in the lungs but the signal those cells released while trapped.

Follow what this signal does, and you can see where the subject that heals moves. The growth factors and cytokines the cell secretes (chemical letters that recruit other cells, make them grow, and regulate inflammation) call the host's cells to the site of injury, block the suicide of dying host cells, and coax new blood vessels to sprout. The cell also releases exosomes (the best-known of the extracellular vesicles), tiny membrane sacs. Inside them, signals such as proteins and RNA are packed and delivered to neighboring cells. If a cytokine is a chemical letter, an exosome is the parcel that ships its contents.

The most important target is the host's macrophage. The macrophage is the body's cleanup and security cell; it eats invaders and can both ignite inflammation and quell it. The same macrophage shifts character, depending on the situation, between a destroy-and-inflame mode and a repair-and-calm mode; this shift of character is called polarization, and turning that switch is called immunomodulation. The signals the MSC secretes turn the host's macrophage from the destroy-and-inflame type to the repair-and-calm type. The fire crew that was putting out the blaze is reassigned as the recovery crew.

And here a paradox begins. As we saw, most injected MSCs die. A leading recent explanation holds that this death may be not a failure but the mechanism itself. The process by which cleanup cells, macrophages among them, devour dead cells is called efferocytosis, literally "eating the dead." The act of eating dead cells itself flips on a signal: "crisis over, repair begins." Putting out the fire and clearing the debris becomes the siren that announces the blaze is out. In a mouse model of graft-versus-host disease (a condition in which transplanted immune cells instead attack the patient's body), the recipient's cytotoxic cells (immune cells that kill other cells) drove the injected MSCs straight into death, and that death was exactly what triggered immunomodulation within the body. In animals lacking the activity that kills MSCs, no immunosuppression occurred either. And a similar picture emerges with intravenous injection. When MSCs put into a vein are eaten by monocytes and macrophages in the lungs, one study observes that the cells that ate them turn anti-inflammatory and even summon other cells that calm immunity. The healing comes not from cells surviving and becoming tissue, but from their being killed and eaten.

That said, "healing by dying" is not yet settled science but, for now, the most persuasive hypothesis. A comprehensive review that sets the paracrine activity of the living cell alongside the efferocytosis of the dead one also labels this contribution of death "powerful but under debate." Should evidence accumulate that cell death instead diminishes the effect, this layer may weaken. But the larger picture—that the signals a cell releases while alive do the healing—still stands.

Step back from this chain, and the subject that heals moves from beginning to end. The injected cell, arriving as a guest, releases signals; those signals move the cells and immunity of the host, the homeowner; and what actually carries out regeneration and anti-inflammation is, in the end, the host itself. The injected cell pulls the trigger, spends itself, and disappears. Back to the image from the opening: the cell does not "become" tissue; it "speaks to" the house next door and sets that house moving on its own.

Take the cells out entirely and inject only the signals they made—does that still work? It does. Filter and inject only the culture medium in which cultured MSCs had secreted their signals, and, with no cells, acutely infarcted myocardium was protected. But what this experiment used was the concentrated culture medium of a special MSC engineered to overexpress a gene called Akt, so to stretch it to "ordinary MSCs heal without cells too" would overstate it. The cleaner evidence, with the cells removed altogether, comes later. What to note here is the direction.

This picture does not clash with the old complaint that MSC clinical results are erratic. It meshes with it. If engraftment is what heals, then how many cells survive and settle would govern the effect. If the signal is what heals, then how the cells were grown, how actively they are releasing signals, and whether the host's inflammatory environment is ready to receive those signals all shake the result. That the results are hard to reproduce points to the latter. This piece does not claim that MSC injections have a large effect. It claims only that when healing happens, its share lies not in what the cell became but in the signal the cell sent—a matter of the mechanism's attribution, not the size of the efficacy.

The Misunderstanding of Lumping Three Strands Into One

One thing has to be made clear here. Push "the cell does not heal" into the conclusion "stem cell therapy is all a sham," and you miss the mark just as badly from the opposite direction. In truth, the common picture—"stem cells become tissue and heal"—is not simply wrong. In one strand it holds exactly; in another it fails. The misunderstanding comes from lumping the different things called "stem cell therapy" into one. What this piece has dissected so far is just one of those strands—MSCs, the mainstay of the "stem cell injections" sold in clinics. Beside the MSC lie two more strands of an entirely different character.

The MSC regenerative injection, this piece's protagonist, is just as we saw. The cells mostly disappear, and if healing happens, its share comes from the signal the cell sent. The two strands set beside it draw the boundary of this diagnosis.

The first contrast is hematopoietic stem cell transplantation, the treatment we usually call a bone marrow transplant. In a disease such as leukemia, the patient's blood-making stem cells are swapped out for those of a healthy donor. These cells actually engraft, take hold, and remake the patient's blood for the rest of their life. The picture of healing by swapping cells holds exactly here. This is precisely where "stem cells become tissue and heal" genuinely holds, and the bone marrow transplant is the most established success of orthodox stem cell therapy.

The second contrast comes from iPSCs, or induced pluripotent stem cells: cells made by winding a fully grown body's cells back into a state where they can become anything. Differentiate these into the desired tissue cells and transplant them, and, unlike the MSC injection, you can aim for genuine engraftment and integration. In a case where a sheet of retinal pigment epithelium cells made from the patient's own cells was transplanted beneath the retina of a patient with age-related macular degeneration, the sheet took hold without rejection and the decline in vision stopped. But this transplant is still experimental. Only one person has received this autologous-sheet approach; for the second patient, a genomic change was found in the prepared iPSCs and the surgery was called off. The attempt to transplant iPSC-derived dopaminergic neurons for Parkinson's disease showed engraftment and functional recovery in primate experiments, and has since moved on to early-phase human clinical trials. In this strand, "the cell becomes tissue" is the goal, and a frontier not yet reached.

The three strands share one name—"stem cell therapy"—but what they do in the body differs. In hematopoietic transplantation the cell becomes tissue; in the MSC injection the cell speaks; in the iPSC transplant the cell is still trying to become. Only the bone marrow transplant is established; the signals the MSC sends and the engraftment iPSCs aim for are both still fields where the evidence is accumulating. What Caplan's renaming took aim at, too, was not all three but one of them—the MSC he himself had named. And that renaming is not one man's verbal tic but a signature summarizing in a single line an independent pile of evidence: disappearance tracking, the TSG-6 experiment, paracrine factors and exosomes, and macrophage conversion.

Not the Cell, but the Cell's Language

If what heals is not the cell but the cell's message, the next question is natural. Why not just refine the message and use that? Regenerative medicine's center of gravity is in fact shifting that way. Filter out only the exosomes an MSC made—not a single cell among them—and inject them into a mouse, and the infarct size of myocardial ischemic injury fell. The sum of these signals a cell sends out is called the secretome, and the direction of treating with this secretome or the exosomes alone, cell-free, is under study under the name "cell-free therapy." But this is, for the most part, animal experiments and early stages, not a proven standard of care. The direction is that way; it does not mean the destination is reached.

Getting the mechanism right is no idle matter of precision. The false picture that "the injected stem cell grows into the damaged tissue" makes the promise that "one shot regenerates you" easy to sell. Erase the fact that the cells mostly disappear and that any effect comes from the signal, and you open a gap in which an unproven injection is easily packaged as universal regeneration. That is why getting the mechanism right matters in practical terms too.

So the question to ask of cell therapy in regenerative medicine is not "what does this cell become." It is "what does this cell say, and what in the host does it switch on," and the answer differs by layer. The cell does not do the healing. The cell speaks to the host to switch on the host's own repair, spends itself and disappears, and in some layers works by dying. The next chapter of regenerative medicine lies not in the cell, but in handling the cell's language.