After MSCs Enter the Bloodstream, Where Do the Cells Actually Go?

TL;DR — Suspended mesenchymal stem cells measure 15–19 micrometers in diameter; pulmonary capillaries trap particles larger than 10 micrometers. Human data: 27–33.5% of signal remains in the lungs immediately after infusion. Animal data: viable cells are unrecoverable from other organs within 24 hours, and no migration occurs even when injury signals are artificially induced. Therapeutic effects come from the immunomodulation triggered after the cells die — not from cells reaching a target lesion. This mechanism has a narrow range.

Cell therapy advertisements often feature a long list of diseases. The most recent one I came across listed over forty, including cerebral infarction, Parkinson’s disease, Alzheimer’s disease, ALS, diabetes, myocardial infarction, chronic kidney disease, and pulmonary fibrosis.

The list also included periodontitis, androgenetic alopecia, post-breast-cancer recovery, and COVID-19.

In the previous two articles, I approached this from a regulatory angle: filing for notification is not the same as approval, and each treatment plan is tied to a specific disease — one comparison is enough to see the gap.

This article takes a different approach. Even setting every regulatory document aside, one basic biological question is enough to make that list collapse:

After a dose of mesenchymal stem cells enters the bloodstream, where do the cells actually go?

I recently started researching the feasibility of bringing a cell therapy to Taiwan, so I read through the literature. Some of what I found was quite different from my initial assumptions — different enough to be worth writing down.

Can intravenously infused cells reach their target organ?

Start with one unglamorous fact. The answer to this question is largely determined by fluid mechanics.

Suspended mesenchymal stem cells measure approximately 15 to 19 micrometers in diameter. This figure comes from a study that used fluorescent microspheres of varying sizes to test the caliber of pulmonary capillaries. Microspheres of 4 micrometers passed through with almost no retention; those of 10 and 15 micrometers were retained significantly. The same study documented that “the vast majority of cells appeared in the lungs within seconds” of intravenous injection. (Schrepfer et al., Transplant Proc 2007, PMID 17362785)

In other words, the caliber of pulmonary capillaries happens to fall right at the threshold of the cells’ size. The cells are wider than the vessels, so they get stuck. This has nothing to do with where the cells might “want” to go.

Quantitative human data are also available. One study infused radiolabeled autologous bone marrow MSCs into four patients with advanced cirrhosis and tracked distribution by whole-body planar scanning over ten days:

(Gholamrezanezhad et al., Nucl Med Biol 2011, PMID 21810549)

This table requires careful reading. What it measures is the distribution of radioactivity, not the distribution of viable cells. After cells die, the label travels with cell debris and is cleared by the mononuclear phagocyte system in the liver and spleen — the body’s mechanism for removing cellular debris — which explains why hepatic and splenic signals rise over time.

“30–42% in the spleen by day 10” cannot be read as “30–42% of viable cells in the spleen.” Those signals most likely represent cell remnants.

I include this caveat because the table, taken out of context, could be misread as evidence that “cells distribute to organs throughout the body” — the opposite of what the data support.

How long do these cells survive in the body?

The answer to this question is even shorter: hours.

In a mouse study, cells were infused via the tail vein and animals were sacrificed at 5 minutes, 1 hour, 24 hours, and 72 hours. Researchers examined blood, lungs, liver, spleen, kidneys, and bone marrow. The findings:

• At 1 hour, signal was concentrated in the lungs.
• At 24 hours, signal was concentrated in the liver.
• Culture recovery: viable donor cells could be recovered from the lungs up to 24 hours after infusion, then disappeared.
• At no time point, in any other organ, were viable cells recovered.

The study also included a critical additional experiment: even when hepatic ischemia-reperfusion injury was deliberately induced, no migration of viable cells to the liver occurred. (Eggenhofer et al., Front Immunol 2012, PMID 23056000)

That control directly addresses the claim most common in advertisements: “stem cells find their way to injured tissue.” The researchers deliberately created an injury signal. The cells still did not go.

Another study put more specific numbers on this. After 2 million human MSCs were infused intravenously into mice, most formed emboli in the lungs, with a pulmonary half-life of approximately 24 hours, and fewer than 1,000 cells in total appeared across six other tissues combined. (Lee et al., Cell Stem Cell 2009, PMID 19570514)

Long-term human data tell the same story. One study examined autopsy material from 18 patients who had received allogeneic MSC therapy. 108 tissue samples from 15 patients were analyzed by PCR. Macroscopic and histological examination revealed no ectopic tissue formation and no donor-derived malignancies. Donor DNA was detected in the lungs, lymph nodes, and gut of 8 patients, at levels of one percent or below. Among 13 samples taken within 50 days of infusion, 9 were positive; among 8 samples taken beyond 50 days, only 2 were. The researchers described the mechanism with a single phrase: hit and run. (von Bahr et al., Stem Cells 2012, PMID 22553154)

The same paper also noted that the absence of sustained engraftment limits long-term risk. That is an honest observation with two sides, and I am not going to cite only the half that suits my argument.

Then why does it work at all?

A reasonable question arises at this point: if the cells never arrive and do not survive, why have clinical trials produced positive results, and why have any products received regulatory approval?

The answer is that the mechanism is not what most people imagine.

One study found that the recipient’s cytotoxic cells induce apoptosis in MSCs via a perforin-dependent pathway. Apoptosis means controlled cell death, and this apoptotic process is a necessary condition for immunosuppression to be activated. A clinical parallel was also observed: among graft-versus-host disease patients, only those with high cytotoxic activity against the infused cells responded to treatment; those with low activity did not. Phagocytes that engulf apoptotic cells subsequently produce immunosuppressive enzymes, and that is how the effect is propagated. (Galleu et al., Sci Transl Med 2017, PMID 29141887)

A second study filled in the downstream picture. By 24 hours post-infusion, most cells had died and been phagocytosed by monocytes in the lungs and liver. These monocytes, having engulfed the cells, shifted their functional state: they upregulated immunosuppressive molecules, reduced inflammatory cytokine output, and induced regulatory T cells. The monocytes are the true carriers of the immunomodulatory effect. (de Witte et al., Stem Cells 2018, PMID 29341339)

The myocardial infarction mouse study mentioned earlier included one more direct experiment. Cells trapped in the lungs upregulate an anti-inflammatory protein. When the researchers used interference techniques to suppress expression of that protein, the cells lost their protective effect. Conversely, intravenous injection of the recombinant protein alone was sufficient to replicate the therapeutic effect. (Lee et al., 2009)

Taken together, the picture is clear: MSCs behave more like a rapidly metabolized anti-inflammatory agent than a piece of implanted tissue.

The cells dying is part of the process, not a flaw in it.

How many diseases can this mechanism actually reach?

Once the mechanism is understood, that list of forty-plus diseases can be addressed directly.

If the therapeutic logic is “cells reach the lesion and repair the tissue,” the evidence above closes that path. The only remaining pathway is this: cells die and are phagocytosed in the lungs, and systemic immunomodulation then indirectly influences distant organs.

This pathway is mechanistically coherent. But it changes the nature of the claim, and it brings three constraints.

First, what this mechanism does is immunomodulation — it cannot directly repair organs. Any claim must be reframed as “modulating inflammation,” not “repairing damaged tissue.”

Second, the ceiling of the effect depends on how much of a given disease’s pathology can be modified by immunomodulation. The amyloid and tau pathology of dementia, the loss of islet cells in diabetes, the necrotic myocardium in myocardial infarction — none of these primary pathologies can be reversed by a brief, transient episode of immunomodulation.

Third, each indication must be demonstrated independently through controlled trials. “Having immunomodulatory properties” and “producing clinical benefit in this specific disease” are two different things.

The one thing the forty-plus diseases on that list have in common is that inflammation is present somewhere in the disease process. But there is a long distance between “inflammation is present” and “suppressing inflammation will cure this disease.” By the same logic, aspirin should be able to treat forty-plus conditions.

Sharing a downstream phenomenon is not the same as sharing a tractable upstream mechanism.

A persuasive case of failure

If the reasoning above still feels abstract, one case study makes the whole thing concrete.

Alofisel (darvadstrocel) is an allogeneic adipose-derived MSC product for complex perianal fistulas in patients with Crohn’s disease. Its starting conditions were far better than anything typically seen in direct-to-consumer advertisements:

• Administration was by local injection into the fistula tract, not intravenous infusion — bypassing the pulmonary trap entirely.
• It targeted a single indication, not a broad disease spectrum.
• It completed full regulatory review and received EU marketing authorization on March 23, 2018.

Then a randomized, placebo-controlled confirmatory trial in 568 patients failed to meet its primary endpoint. The European Medicines Agency determined that the clinical benefit no longer supported continued use, and the marketing authorization was withdrawn on December 13, 2024. (European Medicines Agency official product page)

This product was administered directly at the site of disease, targeted a single indication, and passed full Phase III review to receive EU approval. A larger controlled trial still overturned it.

With that in mind: what is the evidentiary basis for a preparation that has never been tested in a controlled trial, is administered intravenously, and claims to cover more than forty diseases?

Why have 1,894 trials produced only 5 approved indications?

Consider one more set of numbers.

ClinicalTrials.gov lists 1,894 interventional trials using MSCs as the intervention.

The number of MSC indications that have received approval across major global markets is approximately five:

Among these, the indications that are fully approved (not conditional) and use the intravenous route effectively reduce to one: acute graft-versus-host disease.

The U.S. FDA has approved only one MSC product to date — RYONCIL. This is the result of reviewing the full text of the FDA’s official list of approved cellular and gene therapy products, which contained 52 products at the time. The rest are CAR-T therapies, adeno-associated virus gene therapies, umbilical cord blood hematopoietic stem cells, and others. Only one is an MSC product.

1,894 trials. Five approved indications. More than a thousand studies, and very few have cleared the threshold.

If claims covering forty-plus diseases held up, regulatory agencies around the world would not have approved only five.

”Meeting ISCT criteria” only proves cell identity

Another common advertising claim is that the cells “meet the International Society for Cell & Gene Therapy (ISCT) standards.” This phrase deserves to be unpacked.

In 2006, ISCT established three minimum criteria for mesenchymal stromal cells: plastic adherence under standard culture conditions; expression of specified surface markers and absence of another set; and in vitro differentiation capacity into osteoblasts, adipocytes, and chondroblasts. (Dominici et al., Cytotherapy 2006, PMID 16923606)

These three criteria answer the question: “Is this batch of cells mesenchymal stromal cells?” None of them address efficacy.

ISCT’s 2019 position paper was explicit on this point: functional assays “are not defined universally, but by the intended mechanism of action.” One disease requires its own assay; a single assay cannot serve every indication. (Viswanathan et al., Cytotherapy 2019, PMID 31526643)

There is also a safety dimension worth noting. When those 2006 criteria were established, nearly all MSCs used clinically came from bone marrow. The market now relies heavily on umbilical cord and adipose sources.

One study compared all three sources directly. It found that bone marrow-derived cells include a subpopulation with significantly lower tissue factor expression, corresponding to lower procoagulant activity. Tissue factor influences the coagulation cascade. Umbilical cord and white adipose tissue-derived cells, by contrast, showed widespread tissue factor expression and sustained clot formation.

In a rat intravenous injection experiment, tissue factor-deficient bone marrow cells produced no thromboembolism, while tissue factor-expressing stromal cells induced extensive intravascular thrombosis. The researchers described this as sounding “an additional note of caution for the uncritical systemic administration of stromal cells, especially from non-hematopoietic, extravascular sources.” (Oeller et al., Theranostics 2018, PMID 29507631)

The ISCT minimum criteria still do not include assessment of tissue factor or hemocompatibility. Research teams have called for revision. (Moll et al., Stem Cells Transl Med 2022, PMID 35641163)

“Meeting ISCT standards” is using an identity card as a license to operate. It proves only that the cells are mesenchymal stromal cells. Efficacy must be established separately. And as of 2026, those criteria are insufficient to guarantee the safety of intravenous administration.

Even the person who named them wants to rename them

This is the point I most wanted to put on record.

Arnold Caplan, the researcher who originally coined the term “mesenchymal stem cell,” published an article in 2017 with a title that translates directly as: “Mesenchymal Stem Cells: Time to Change the Name!”

His reasons had nothing to do with academic convention. The article argues directly that calling these cells “stem cells” is being used to imply to patients that they will receive direct medical benefit, because patients imagine the cells will differentiate into cells that regenerate tissue. He named specific claims, from “bone-on-bone knees” to “neurological diseases including dementia.” He proposed renaming them Medicinal Signaling Cells. (Caplan, Stem Cells Transl Med 2017, PMID 28452204)

And even earlier, ISCT’s 2005 nomenclature position statement noted that the biological properties of these cells “do not appear to meet the generally accepted criteria for stem cell activity, making the name scientifically inaccurate and potentially misleading to the general public.” (Horwitz et al., Cytotherapy 2005, PMID 16236628)

Both of those statements came from inside the field itself.

So do MSCs actually work?

I do not want to overcorrect in the course of dispelling a misconception, so this section will lay out the positive evidence clearly.

MSCs are effective in steroid-refractory acute graft-versus-host disease, and the effect is not trivial. A multicenter Phase II trial enrolled 55 patients with severe disease. Thirty achieved complete response and nine showed improvement. Among complete responders, one-year transplant-related mortality was significantly lower (37% vs. 72%) and two-year overall survival was significantly higher (53% vs. 16%). (Le Blanc et al., Lancet 2008, PMID 18468541. To be clear: this was a single-arm Phase II trial, not a randomized controlled trial.)

The stronger signal is at the regulatory level: two independent, stringent regulatory agencies — the U.S. FDA and Japan’s PMDA — each approved a product for the same indication at different points in time. That kind of cross-jurisdictional convergence carries more weight than any single trial.

Why MSCs work in this particular indication is precisely what the mechanistic picture above would predict. Graft-versus-host disease is a systemic, T-cell-mediated hyperimmune response — squarely within the range of the mechanism MSCs have been shown to exert (induction of regulatory T cells, immunosuppressive enzymes, monocyte polarization).

When the mechanism matches the disease, the therapy works. When it does not, it does not.

That said, a mechanistic match does not guarantee success. COVID-19-associated acute respiratory distress syndrome is inflammation-driven and, in theory, should fall within range. Yet in a French multicenter randomized double-blind placebo-controlled trial, the primary endpoint showed no significant difference between groups. (Monsel et al., Crit Care 2022, PMID 35189925)

Even the most theoretically well-matched direction can fail. That is why the claim that “one cell type treats forty-plus diseases” runs counter to everything this field has established.

Key points from this article

1. Suspended MSCs measure approximately 15–19 micrometers in diameter; pulmonary capillaries trap particles larger than 10 micrometers. Human data show that 27–33.5% of signal remains in the lungs immediately after infusion.
2. Viable cells disappear from the lungs within approximately 24 hours and are unrecoverable from all other organs at any time point. Artificially induced organ injury signals do not trigger migration.
3. The therapeutic mechanism operates through the brief immunomodulation triggered when cells are killed and phagocytosed — not through differentiation or replacement. Cellular death is part of the mechanism.
4. The mechanism has a narrow range: it can modulate inflammation, but it cannot address amyloid deposits, cell loss, or tissue necrosis.
5. Alofisel was administered locally, targeted a single indication, and received full regulatory approval — and still failed in a 568-patient confirmatory trial and was withdrawn in December 2024.
6. ClinicalTrials.gov lists 1,894 MSC trials; approximately five indications have received full approval across major global markets. The U.S. FDA has approved only one MSC product to date.
7. “Meeting ISCT criteria” answers the question of cell identity, not efficacy. Those criteria were established for bone marrow-derived cells and do not include tissue factor or hemocompatibility assessment.
8. MSCs demonstrably work in acute graft-versus-host disease — both the U.S. FDA and Japan’s PMDA have independently approved products for that indication. This reflects a genuine mechanistic match.

What this article does not claim

There is a risk in writing this kind of piece of overstating the case for rhetorical effect. The following are points I investigated where the evidence is insufficient to support a strong claim, and I want to be clear about that:

• I did not find any peer-reviewed case reports clearly attributing a death to intravenous MSC infusion. Mechanistic evidence for thrombotic risk exists in animal models (extensive intravascular thrombosis in rats), and review articles mention individual case reports of thromboembolic complications in clinical settings — but I cannot confirm fatal cases from primary literature, and so this article makes no such claim.
• Tumor formation risk is not supported by current clinical evidence. Autopsy studies found no donor-derived tumors, and meta-analyses have not identified elevated malignancy risk. The FDA’s approval letter does, however, require expedited reporting of ectopic tissue formation events for three years, indicating that long-term uncertainty has not been fully resolved.
• In controlled trials using well-manufactured products, the safety record of intravenous MSC infusion is acceptable. A meta-analysis of 55 randomized controlled trials involving 2,696 patients found that the only significantly elevated adverse event was transient fever. This conclusion does not extend to products of unknown origin or unregulated manufacture — but it should not be omitted.

References

Listed in order of appearance. All cited studies are peer-reviewed journal articles; all regulatory documents are from official agency pages.

Biodistribution and survival

• Schrepfer S, et al. Stem cell transplantation: the lung barrier. Transplant Proc. 2007. PMID 17362785. https://doi.org/10.1016/j.transproceed.2006.12.019 (cell diameter 15–19 µm; pulmonary capillaries trap particles >10 µm)
• Gholamrezanezhad A, et al. In vivo tracking of ¹¹¹In-oxine labeled mesenchymal stem cells following infusion in patients with advanced cirrhosis. Nucl Med Biol. 2011. PMID 21810549. https://doi.org/10.1016/j.nucmedbio.2011.03.008 (quantitative human data; 27.0–33.5% in lungs immediately post-infusion)
• Eggenhofer E, et al. Mesenchymal stem cells are short-lived and do not migrate beyond the lungs after intravenous infusion. Front Immunol. 2012. PMID 23056000. https://doi.org/10.3389/fimmu.2012.00297 (no viable cells recoverable from other organs at 24 hours; no migration despite induced liver injury)
• Lee RH, et al. Intravenous hMSCs improve myocardial infarction in mice because cells embolized in lung are activated to secrete the anti-inflammatory protein TSG-6. Cell Stem Cell. 2009. PMID 19570514. https://doi.org/10.1016/j.stem.2009.05.003 (pulmonary half-life ~24 hours; recombinant protein alone replicates therapeutic effect)
• von Bahr L, et al. Analysis of tissues following mesenchymal stromal cell therapy in humans indicates limited long-term engraftment and no ectopic tissue formation. Stem Cells. 2012. PMID 22553154. https://doi.org/10.1002/stem.1118 (autopsy material from 18 patients; hit and run)

Mechanism of action

• Galleu A, et al. Apoptosis in mesenchymal stromal cells induces in vivo recipient-mediated immunomodulation. Sci Transl Med. 2017. PMID 29141887. https://doi.org/10.1126/scitranslmed.aam7828 (apoptosis is a necessary condition for immunosuppression)
• de Witte SFH, et al. Immunomodulation by therapeutic mesenchymal stromal cells (MSC) is triggered through phagocytosis of MSC by monocytic cells. Stem Cells. 2018. PMID 29341339. https://doi.org/10.1002/stem.2779 (monocytes as the actual carriers of immunomodulation)

Regulatory record

• U.S. FDA, Approved Cellular and Gene Therapy Products: https://www.fda.gov/vaccines-blood-biologics/cellular-gene-therapy-products/approved-cellular-and-gene-therapy-products
• U.S. FDA, RYONCIL (remestemcel-L-rknd) BLA approval letter, December 18, 2024: https://www.fda.gov/media/184605/download
• European Medicines Agency, Alofisel (darvadstrocel) product page (including December 13, 2024 withdrawal and ADMIRE-CD II trial results): https://www.ema.europa.eu/en/medicines/human/EPAR/alofisel
• U.S. FDA, Statement on stem cell clinic permanent injunction, June 25, 2019: https://www.fda.gov/news-events/press-announcements/statement-stem-cell-clinic-permanent-injunction-and-fdas-ongoing-efforts-protect-patients-risks
• ClinicalTrials.gov: interventional trials using MSCs as the intervention, 1,894 results (queried August 28, 2026).

Nomenclature and standards

• Horwitz EM, et al. Clarification of the nomenclature for MSC: The International Society for Cellular Therapy position statement. Cytotherapy. 2005. PMID 16236628. https://doi.org/10.1080/14653240500319234 (“scientifically inaccurate and potentially misleading to the general public”)
• Dominici M, et al. Minimal criteria for defining multipotent mesenchymal stromal cells. Cytotherapy. 2006. PMID 16923606. https://doi.org/10.1080/14653240600855905 (three minimum criteria)
• Viswanathan S, et al. Mesenchymal stem versus stromal cells: ISCT MSC committee position statement on nomenclature. Cytotherapy. 2019. PMID 31526643. https://doi.org/10.1016/j.jcyt.2019.08.002
• Caplan AI. Mesenchymal Stem Cells: Time to Change the Name! Stem Cells Transl Med. 2017. PMID 28452204. https://doi.org/10.1002/sctm.17-0051 (the original naming researcher calls for renaming)

Safety

• Oeller M, et al. Selection of tissue factor-deficient cell transplants as a novel strategy for improving hemocompatibility of human bone marrow stromal cells. Theranostics. 2018. PMID 29507631. https://doi.org/10.7150/thno.21906 (tissue factor expression and thrombotic risk in umbilical cord and adipose-derived cells)
• Moll G, et al. Improved MSC minimal criteria to maximize patient safety. Stem Cells Transl Med. 2022. PMID 35641163. https://doi.org/10.1093/stcltm/szab005 (call to incorporate tissue factor and hemocompatibility assessment into minimum criteria)
• Thompson M, et al. Cell therapy with intravascular administration of mesenchymal stromal cells continues to appear safe: an updated systematic review and meta-analysis. EClinicalMedicine. 2020. PMID 31989101. https://doi.org/10.1016/j.eclinm.2019.100249 (55 RCTs, 2,696 patients; only significantly elevated adverse event was transient fever)

Efficacy evidence

• Le Blanc K, et al. Mesenchymal stem cells for treatment of steroid-resistant, severe, acute graft-versus-host disease: a phase II study. Lancet. 2008. PMID 18468541. https://doi.org/10.1016/S0140-6736(08)60690-X (single-arm Phase II trial)
• Monsel A, et al. Treatment of COVID-19-associated ARDS with mesenchymal stromal cells: a multicenter randomized double-blind trial (STROMA-CoV-2). Crit Care. 2022. PMID 35189925. https://doi.org/10.1186/s13054-022-03930-4 (primary endpoint did not reach statistical significance)

Series: Filing for Notification Is Not the Same as Approval · What to Do When You See a Stem Cell Advertisement