What Is Settled, and What Is Not
It is worth being clear about the state of this question before going into detail, because it is often presented as more resolved than it is.
Reasonably settled: bone-marrow MSC number and function decline with donor age; adipose tissue yields MSCs more abundantly and less invasively than marrow; cord tissue is collected without any procedure on a donor; and MSCs from all these tissues meet the same defining criteria and suppress immune responses.
Not settled: whether any source produces better clinical outcomes in patients. No trial has randomised patients between cell sources for any indication. Every comparison below is a laboratory comparison, and laboratory measures — growth rate, marker expression, T-cell suppression in co-culture — have not been shown to predict how a patient will do.
Anyone presenting a definitive ranking of cell sources for treatment is going beyond what has been demonstrated.
The Sources
⚠️ The distinction that causes most of the confusion. "Umbilical cord MSCs" is used for two quite different materials: Wharton's jelly (the gelatinous tissue of the cord itself) and cord blood. MSCs are abundant and reliably isolated from Wharton's jelly; from cord blood they are scarce and often cannot be isolated at all. Comparative studies of the two reach different conclusions, and citing one as evidence for the other is a common and material error. Where the studies below used cord blood, it is stated.
A note on history
Alexander Friedenstein and colleagues identified the marrow stromal cells that became known as MSCs in work published through the 1960s and 1970s [2]. Adipose tissue was established as a practical source in 2001, when Zuk and colleagues showed that liposuction aspirate contained multilineage progenitor cells [7].
For umbilical cord, priority is genuinely contested rather than settled: Romanov and colleagues reported candidate MSC-like cells from umbilical cord in 2003 [11], and Wang and colleagues reported MSCs specifically from Wharton's jelly in 2004 [12].
Autologous or Allogeneic
This choice is separate from the choice of tissue, and often matters more.
Autologous — your own cells, harvested from your marrow or fat, expanded, and returned. No donor-matching question. Requires a procedure, and the cells are the age you are.
Allogeneic — donor cells, most often from cord tissue, manufactured in batches and banked. No procedure for you, immediate availability, consistent batch characterisation. The cells come from someone else.
On the immune question. MSCs express low levels of MHC class I and little or no class II, and they actively suppress immune responses — which is why allogeneic administration is feasible at all [17][18]. That is not the same as immune privilege. The often-repeated claim that MSCs are "immune-privileged" overstates the evidence: the underlying laboratory work describes these cells as having properties that "would be permissive to allogeneic transplantation" [17]— an in vitro characterisation. Allogeneic MSCs can provoke donor-specific antibody responses, and repeat dosing is where that matters most.
What the Comparative Studies Actually Found
The four-source comparison: Wharton's jelly ranked first on two measures
Li and colleagues compared MSCs from bone marrow, adipose tissue, Wharton's jelly and placenta in the same laboratory [14]. Their findings:
- Mitogen-induced T-cell proliferation was effectively suppressed by all four populations, through soluble factors rather than contact inhibition
- "Among the 4 populations of MSCs, the WJ-MSC has the strongest suppression effects"
- Wharton's jelly showed the weakest expression of MHC class II, TLR4, TLR3, JAG1, NOTCH2 and NOTCH3
- On proliferation: "WJ-MSCs showed the most rapid growth rate followed by the AT-, PL- and BM-MSCs"
- On osteogenic differentiation: "the WJ-MSCs had the strongest osteogenetic ability followed by PL, AT and BM-MSC"
- On adipogenic differentiation: adipose MSCs strongest, followed by Wharton's jelly, bone marrow and placenta
Their conclusion: WJ-MSCs "are the most attractive cell population for use in immune cellular therapy when immunosuppressive action is required" [14].
This is real evidence and it does favour cord tissue on those measures. It is also a single in vitro study, and its own conclusion is conditional on the indication.
Note also that the ranking is not uniform: adipose came first on adipogenic differentiation, and bone marrow last on osteogenic — a reversal of the conventional assumption that marrow is the bone-forming source. Which tissue looks best depends on which assay is run.
A dedicated comparison of adipose against bone marrow reaches the same conclusion about the difficulty of ranking: the two share far more than they differ on [9]. Broader reviews of adult versus neonatal sources describe the same picture — real differences in cell-biological properties and marker expression, without a single ordering that holds across outcomes [3][16].
The comparison that found no difference
Heo and colleagues compared MSCs from bone marrow, adipose tissue, placenta and umbilical cord blood on immunophenotype, proliferative potential, differentiation and immunomodulatory capacity [15]. Their finding:
"No significant differences in growth rate, colony-forming efficiency and immunophenotype were observed."
They also reported that bone-marrow and adipose MSCs shared not only tri-lineage differentiation potential but gene-expression profiles.
These two results do not contradict each other. Li studied Wharton's jelly; Heo studied cord blood. Against cord blood, no significant difference. Against cord tissue, cord ranked first on immunosuppression and growth rate. That is the honest reconciliation, and it is why the tissue/blood distinction matters so much.
The review most often cited for cord superiority
Liau and colleagues' review supports the qualitative claim: Wharton's jelly MSCs "have excellent proliferative potential and a faster growth rate and can retain their multipotency for more passages in vitro compared to adult MSCs from bone marrow or adipose tissue" [13].
It ends less often quoted:
"Most clinical trials/studies reported thus far are relatively smaller in scale. The collected evidence is insufficient to support the routine use of WJ-MSC therapy in the clinic. Thus, rigorous clinical trials are needed in the future to obtain more information on WJ-MSC therapy safety and efficacy." [13]
Where each source has a recognised strength
Donor Age
This is the best-supported source-related claim on this page, and it applies specifically to autologous therapy.
Bone-marrow MSC number and quality decline with donor age. Stolzing and colleagues found reduced colony-forming units and reduced MSC numbers in adulthood relative to childhood, alongside increased oxidative damage, reactive oxygen species and the senescence markers p21 and p53 — "an age-related decline in overall BM MSC 'fitness' which might lead to problems when using autologous aged MSC for cell-based therapies" [4]. They note that earlier work on this question had been "contradictory", and set out to resolve it with a wider age range.
The pattern is corroborated across tissue-resident MSCs generally [5], in bone-marrow MSCs specifically [6], and sits within a broader literature on how stem-cell function changes with age [1]. Adipose MSCs are also affected: donor age negatively impacts adipose MSC expansion and differentiation [8], and adipose tissue from people with obesity yields cells with an altered immunophenotype and function [10].
What this supports: if you are older and considering autologous therapy, the cells available to you are likely to be fewer and less proliferative than a younger person's. That is a genuine consideration.
What it does not support: a specific rate of decline. Figures such as "a ten-fold fall from birth to age eighty" or precise MSC-per-marrow-cell ratios at given ages are widely repeated and are not found in the source they are usually attributed to.
What the Clinical Evidence Says About Source — and What It Cannot
Several clinical reviews and trials are routinely cited in discussions of cell source. Read carefully, none of them establishes a source ranking, and two are frequently misreported.
Knee osteoarthritis. Copp and colleagues evaluated 15 randomised and 11 non-randomised controlled trials of culture-expanded MSCs [22]. They found net positive effects on pain and symptoms — function improved in 12 of 15 RCTs relative to baseline and in 11 of 15 relative to controls — and reported cartilage protection or repair in 18 of 21 clinical studies. They examined dose, tissue of origin, and autologous versus allogeneic as parameters. Their own caveat: "the relatively small sample size of 610 patients limited the drawing of definitive conclusions." The review does not conclude that any tissue of origin outperformed another.
COPD. Calzetta and colleagues pooled 11 studies and 371 patients [23]. The 6-minute walk test improved significantly: +52 m (95% CI 18–87). Lung function did not: FEV₁ improved by +71 mL (95% CI −2 to 145), p = 0.056 — which the authors describe as "a strong tendency towards significance", not a significant result. Their conclusion is that stem-cell therapy "may be effective" but that "the current evidence comes from small clinical trials" and large randomised trials are needed. The analysis does not stratify safety by cell source.
Graft-versus-host disease. Zhao and Liu review MSC use in haematopoietic stem cell transplantation, covering engraftment failure and GVHD [24]. It is a general account of the field and does not compare response rates between cell sources.
Aging frailty — and a distinction that matters. Two randomised trials are often cited together here, and they used different cells:
- Tompkins and colleagues (2017) infused allogeneic bone-marrow-derived MSCs intravenously at 100 or 200 million cells in 30 patients, with no therapy-related serious adverse events at one month and physical-performance improvement preferentially in the 100-million group [19]
- Zhu and colleagues (2024) infused umbilical-cord-derived MSCs in a randomised, double-blind, placebo-controlled phase I/II trial, reporting improvement in the SF-36 physical component score at 6 months (p = 0.042), in EQ-VAS (p = 0.023 at 2 months, p = 0.002 at 6 months), and on the Timed Up and Go test [20]
Both are positive, and they used different tissue sources — which is a reason to treat them as two separate pieces of evidence rather than one trend. Neither trial compared sources.
Laboratory rejuvenation work. Lei and colleagues showed that extracellular vesicles from umbilical cord MSCs rejuvenated senescing adult bone-marrow MSCs in culture — increasing self-renewal capacity and telomere length — by transferring PCNA, and reduced bone and kidney degeneration in aged mice [21]. This is striking cell biology about vesicles rather than cells, in culture and in mice, and it is not a clinical finding about cell source.
Route Matters, and Here the Ranking Reverses
For intravenous administration, the source-related evidence points the opposite way to the laboratory rankings above. This is the part most often left out of comparisons.
Different MSC products carry different amounts of tissue factor (TF/CD142), which initiates clotting on contact with blood and can trigger the instant blood-mediated inflammatory reaction (IBMIR).
- Moll and colleagues put the problem directly: "Meta-analysis has demonstrated that bone marrow MSC infusion is safe. It is not clear if this also applies to diverse new cell products derived from other sources, such as adipose and perinatal tissues" [25]. Umbilical cord is a perinatal tissue.
- Hoang and colleagues measured it: "Xeno- and serum-free cultured adipose tissue and UC-derived MSCs expressed the highest level of TF, followed by those from dental pulp, and the lowest expression was observed in MSCs of bone marrow origin" [26]. In the same study, patients receiving intravenous UC-MSC infusion showed a transient rise in D-dimer, while patients receiving cells intrathecally did not.
⚠️ This must not be overstated, and the same researchers are careful about it. Hoang also found no correlation between tissue-factor expression and D-dimer or other coagulation measures, concluding that TF "cannot be used as a solid biomarker to predict MSC-induced coagulant activity" [26]. And pooled randomised data show no increase in thrombotic or embolic events with intravascular MSC therapy.
The honest position: this is an open question about product characterisation for the intravenous route, not a demonstrated harm of cord cells. But it is the one source-specific safety consideration that bears directly on a decision, and any comparison that ranks cord safest without mentioning it is incomplete.
What Is Actually Approved
One MSC product has FDA approval. In December 2024 the FDA approved remestemcel-L-rknd (Ryoncil) for steroid-refractory acute graft-versus-host disease in patients aged 18 and under, reporting "70% response rates and ~70% 6-month survival" [27]. It is an allogeneic bone-marrow-derived mesenchymal stromal cell therapy [28].
That is worth stating plainly on a page about cell sources: the only licensed MSC product in the United States uses bone marrow — the source most often described in commercial material as the weakest.
For every other indication discussed here, MSC therapy of any source remains investigational in most jurisdictions.
Source Is One Variable Among Several
The premise that tissue source is the dominant factor in how a cell product behaves is not supported by the manufacturing literature.
Reviewing why MSC products have produced inconsistent clinical outcomes, Česnik and Švajger identify "inherent and extensive variability" arising from "differences in tissue sources, donor attributes, variances in manufacturing protocols, as well as modes of administration" [29]. Donor age, sex, body mass index and health status all influence phenotype and function; so do culture medium composition, growth-factor supplementation, serum choice and passage number.
Tissue source is the first item on a list of at least four. Two products from the same tissue, made differently, can differ more than two products from different tissues made the same way.
The practical consequence for a reader: "what tissue are the cells from?" is a reasonable question, but on its own it does not tell you much. "How were they made, at what passage, and how will they be given?" tells you more.
Quality Control: What to Look For
Regardless of source:
Identity — verification against ISCT criteria: CD73+, CD90+, CD105+ at ≥95%; CD45, CD34, CD14 negative at ≤2%; plastic adherence; tri-lineage differentiation.
Viability — a stated percentage, measured by flow cytometry rather than manual counting, with the figure on a certificate of analysis. Note that no international body sets a viability threshold; any number quoted is a manufacturer's own specification.
Safety testing — sterility, mycoplasma, endotoxin, and donor infectious-disease screening.
Manufacturing — GMP or equivalent, documented procedures, batch records and traceability, and a stated passage number.
For intravenous products specifically — whether the manufacturer assesses effects on coagulation. No validated release assay for this exists yet, which is precisely why it is worth asking.
Questions worth asking
- What tissue are the cells from — and if umbilical cord, is it cord tissue or cord blood?
- Autologous or allogeneic, and why that choice for my situation?
- At what passage are the cells administered?
- What is the viability, and how is it measured?
- Is the product licensed for this use, or investigational?
- What evidence is there that this specific product helps my condition? — the source question does not answer this one.
Frequently Asked Questions
What is Wharton's jelly?
The gelatinous connective tissue inside the umbilical cord that surrounds and protects the blood vessels, rich in hyaluronic acid and collagen. It is a reliable source of MSCs and is distinct from cord blood. Named after Thomas Wharton, who described it in 1656.
Is one source definitively better?
For patient outcomes, no — that comparison has not been made. In the laboratory, Wharton's jelly ranks first on T-cell suppression and growth rate among four sources [14], adipose ranks first on adipogenic differentiation [14], and against cord blood no significant differences were found at all [15]. The only approved product is bone-marrow-derived [27].
Are my own cells too old to be useful?
They are likely to be fewer and less proliferative than a younger donor's, and that effect is real and well documented [4][5][6]. Whether it changes your clinical result has not been established, and many patients have been treated with their own cells.
Are donor cells rejected?
MSCs express low MHC class I and little class II and suppress immune responses, which is why allogeneic use is feasible and why immunosuppressive drugs are not routinely given [17][18]. They are immune-evasive rather than immune-privileged, and donor-specific antibody responses have been described. Allogeneic MSCs have been given to large numbers of patients in trials with a good safety record [18].
Are MSC therapies FDA approved?
One is. Ryoncil, a bone-marrow-derived product, for steroid-refractory acute GVHD in children [27][28]. For orthopaedic, respiratory, longevity and most other indications, MSC therapy remains investigational.
How long do effects last?
This depends on the indication and is not a property of the cell source. Where duration has been measured it is reported per trial, and this page is about sources rather than outcomes.
The Bottom Line
Tissue source is a real variable. Wharton's jelly MSCs are, in the most direct four-source laboratory comparison available, the strongest immunosuppressors and the fastest growers [14]. Cord tissue is collected without any procedure on a donor, has no donor-age variation, and can be banked and characterised in advance — genuine, practical advantages that need no evidentiary inflation.
Against that: the same claim tested against cord blood found no significant difference [15]; the review most often cited for cord superiority concludes the evidence is insufficient for routine clinical use [13]; the only approved MSC product is bone-marrow-derived [27]; the intravenous safety data apply most directly to bone marrow, and cord and adipose are the higher tissue-factor sources [25][26]; and manufacturing, donor and route variables contribute at least as much as tissue of origin [29].
No trial has compared cell sources head to head in patients. Until one does, the useful questions are about the specific product — how it was made, at what passage, by what route, and what evidence exists for it in your condition — rather than about which tissue it came from.