Obesity has become a global pandemic. Between 1990 and 2022, the worldwide prevalence of adult obesity more than doubled, and projections indicate continued increases through 2040, with the burden shifting progressively toward lower-income countries^[1,2]. This trajectory is expected to substantially increase the global burden of adiposity-related multimorbidity^[3]. Beyond population trends, growing evidence indicates profound effects of obesity at the tissue and cellular levels, including on stem cell biology and regenerative potential^[4]. Several mechanisms have been implicated in how obesity interferes with the reparative functions of circulating and tissue-resident stem cells. Disruption of stem cell niches by obesity-associated tissue and cellular changes, including chronic metabolic stress, low-grade inflammation, oxidative stress, and altered nutrient signaling, has been postulated to reduce stem cell proliferation and differentiation^[5]. Consequently, individuals with obesity may have inadequate self-healing capacity and impaired recovery from insults. In addition, this functional deficiency of mesenchymal stem/stromal cells (MSCs) impedes their potential as a regenerative, autologous, cell-based strategy in individuals with obesity.

Throughout this review, we define “MSCs” according to the minimal criteria proposed by the International Society for Cellular Therapy (ISCT). These include: plastic adherence under standard culture conditions; expression of CD73, CD90, and CD105 with the absence of specified hematopoietic and endothelial markers; and trilineage differentiation in vitro^[6]. Because these are population-level criteria rather than markers of a single-cell identity, and because MSC phenotype and function vary by tissue of origin, we specify the source of MSCs wherever the primary reports allow. Most human data on obesity-induced MSC dysfunction derive from adipose tissue-derived MSCs (AT-MSCs), typically harvested from abdominal subcutaneous fat, whereas bone marrow-derived MSCs (BM-MSCs) and MSCs from other depots may respond differently to the obese milieu.

The reports comparing AT- vs. BM-MSCs are, however, occasionally inconsistent. Donor-matched comparisons usually show tissue-specific differentiation bias, with AT-MSCs favoring adipogenesis and exhibiting higher proliferative rates, and BM-MSCs favoring osteogenesis and chondrogenesis, alongside substantial inter-donor variability^[7,8]. Reported differences in immunomodulatory potency are less consistent. Studies have found AT-MSCs more potent, BM-MSCs to be superior, or the two broadly equivalent. These discrepancies may partly reflect donor matching, passage number, and assay design^[9]. We therefore treat tissue source as requiring explicit statement; unless a source is specified, statements about “obese-MSCs” should be read as pertaining mainly to adipose tissue-derived cells.

Prior reviews have often addressed obesity-induced MSC impairment through a single mechanistic lens. The goal of the present review is integration: we connect the self-amplifying nicotinamide adenine dinucleotide (NAD⁺)-sirtuin-epigenetic loop to epitranscriptomic regulation and to extracellular vesicle (EV) cargo remodeling, trace these mechanisms through to organ-level functional deficits, and appraise candidate rescue strategies according to the experimental model and level of evidence supporting each. We also highlight voids in the current literature, including the reversibility of the obese-MSC phenotype and the near-absence of clinical data. Interpreting this literature requires attention to species, tissue source, and how obesity is defined. This review thus provides an overview of the multi-layered impact of the obese microenvironment on MSC function and considers the challenges to restoring their reparative capacity.

Human obese-MSCs show reduced engraftment owing to diminished proliferation and increased senescence, marked by elevated p16, p53, IL-6, and MCP-1 expression^[67]. BMI correlates with p16/p21-mediated cell-cycle arrest, imposing cellular stress that limits the feasibility of autologous MSC therapy^[15]. BM-MSC from high-fat-diet-fed mice also show reduced expression of VEGF-A and basic fibroblast growth factor (bFGF), resulting in impaired pro-angiogenic properties^[68]. Notably, co-incubation with non-obese MSCs restored VEGF expression and tube formation that were blunted in injured HUVECs^[67].

Obese-MSCs exhibit impaired immunomodulation with a shift toward pro-inflammatory activity. Whereas lean human MSCs promote an anti-inflammatory macrophage phenotype, this capacity is blunted in MSCs from obese donors, which are less effective at activating M2 polarization and suppressing lymphocyte proliferation, and which show increased expression of inflammatory markers and inflammasome activation^[69]. In high-fat-diet-induced obesity, visceral adipose MSCs activate the IL-1 and granzyme-A pathways, reversing their immunoregulatory effects on leukocyte activation, inflammation, and chemokine signaling^[70]. Moreover, obesity reverses the anti-fibrotic activity of MSCs toward pro-fibrotic signaling, accompanied by impaired differentiation and altered paracrine signaling. The senescence-associated secretory phenotype of obese-MSCs includes pro-fibrotic cytokines that promote extracellular matrix deposition and tissue scarring^[70].

The diminished reparative potency of obese-MSCs has been demonstrated in several organ systems.

a) Kidney repairFunctionally, human obese-MSCs fail to protect against renal ischemia-reperfusion injury or to ameliorate chronic ischemic injury, as they develop cellular senescence^[71] and blunted immunomodulatory function^[69], while elevated IL-1β expression shifts their function from therapeutic to pathogenic^[4]. Similarly, porcine obese-EVs fail to restore renal blood flow, glomerular filtration rate, mitochondrial integrity, and oxidative balance, or to reduce tubular injury and fibrosis in renal artery stenosis models^[66,72]. EVs derived from obese adipose-tissue MSCs show impaired reparative activity, failing to suppress inflammation and exhibiting reduced anti-apoptotic and MAPK-modulatory effects compared with EVs from lean MSCs; obesity-associated alterations in EV miRNA cargo also involve regulators of Wnt and MAPK signaling^[64].

b) Cardiac repairObesity markedly attenuates the cardioprotective efficacy of adipose-derived MSCs in vivo. In a murine model of renovascular hypertension (RVH), obese-MSCs only modestly reduced left ventricular (LV) mass, failed to correct systemic hemodynamics, and exacerbated LV systolic dysfunction, indicating a deleterious effect on contractile performance. These functional outcomes were mirrored by tissue-level responses: obese-MSCs displayed a pro-inflammatory phenotype that failed to mitigate oxidative stress and myocardial structural injury^[73] [Figure 3]. Similarly, obese-MSC-EVs fail to prevent cardiac hypertrophy, exerting minimal effects on LV wall thickening and cardiomyocyte hypertrophy and reflecting a compromised cardioprotective phenotype. Notably, obese-EVs modestly reduced myocardial fibrosis, indicating selective preservation of anti-fibrotic signaling pathways^[74].

Obese-MSCs show impaired paracrine signaling marked by increased pro-inflammatory activity and reduced antioxidant responses, consistent with obesity-associated senescence and metabolic dysfunction. As a result, RVH mice treated with obese-MSCs retain a heightened oxidative burden and show only partial reversal of adverse cardiac remodeling, whereas lean-MSC therapy restores redox homeostasis and supports more effective cardiac protection^[73] [Figures 3 and 4].

c) Hepatic repairMetabolic dysfunction-associated steatotic liver disease (MASLD) has become a leading indication for liver transplantation and is now the top indication in the United States among women and patients with hepatocellular carcinoma^[75]. MSCs are postulated as central regulators of liver inflammation, fibrosis, and regeneration^[76,77]. MSCs decrease insulin resistance, oxidized low-density lipoprotein (LDL), and IL-6 in high-fat-diet-fed obese mice, and injection of their suspension improved glucose tolerance and reduced hepatic steatosis, macrophage infiltration, and pancreatic islet hypertrophy^[78]. Dysfunctional endogenous MSCs likely worsen disease progression by failing to control fibrosis and inflammation; however, data directly demonstrating the effect of dysfunctional obese-MSCs on liver disease progression remain limited.

d) Central nervous system repairDirect evidence that obesity compromises MSC-mediated central nervous system (CNS) repair is sparse. MSCs and their EVs improve functional outcomes in preclinical stroke and traumatic brain injury models^[79,80], yet no study has tested cells from donors with obesity in a CNS injury model, and donor adiposity is rarely reported. The most direct evidence is our own: human obese-MSCs displayed altered hydroxymethylation of nuclear-encoded mitochondrial genes governing neuronal development, together with impaired neurogenic differentiation in vitro that improved after epigenetic modulation^[18]. This establishes diminished neurogenic capacity rather than failure of CNS repair, which remains untested in vivo. The concern therefore rests on extrapolation from the senescent, pro-inflammatory phenotype that blunts obese-MSC efficacy in renal and cardiac models, supported by evidence that donor characteristics predict MSC potency after brain injury^[81]. Separately, obesity suppresses endogenous adult neurogenesis^[82,83], though this reflects host rather than graft dysfunction.

A central challenge in synthesizing existing literature is that the underlying studies differ substantially in species, tissue source, and how obesity is defined, and these differences constrain direct comparison. Mechanistic work on ‘obese’ MSCs spans different species, including human, porcine, and murine MSCs. Human studies most often used AT-MSCs from abdominal subcutaneous fat of bariatric or surgical patients, offering clinical relevance but limited sample sizes, heterogeneous comorbidity and medication profiles, and cross-sectional designs^[4,18,67]. Porcine models of diet-induced obesity and metabolic syndrome permit controlled dietary exposure and standardized harvesting and have helped define EV and mitochondrial phenotypes. However, swine metabolic syndrome also encompasses hypertension and dyslipidemia, making obesity-specific effects difficult to isolate^{[19,25,65,66,72]}. Murine high-fat-diet models afford genetic tractability and mechanistic dissection but differ from humans in adipose biology and MSC behavior^[23,68,74].

Several additional factors complicate comparisons among research studies. The tissue of origin is an important variable: AT- and BM-MSCs differ in secretome, immunomodulatory potency, and susceptibility to lipotoxic stress, so findings in one compartment do not necessarily generalize to another^[7,8,11]. Definitions of obesity are similarly inconsistent. Human studies use varying BMI thresholds, whereas animal studies generally rely on diet duration and weight gain. Moreover, few studies account for coexisting insulin resistance, dyslipidemia, or hypertension. Reporting of MSC characterization and EV isolation also varies, further confounding comparisons. Protocol harmonization does not eliminate inter-donor or inter-laboratory variability. In a multicenter study using a shared workflow, broad source-specific patterns were reproducible, but quantitative readouts still varied by laboratory and operator^[9]. Donor metabolic status is increasingly argued to determine adipose-derived MSC potency, which primary reports seldom document^[84]. We therefore outline the model underlying key findings throughout this review and specify the source of supporting evidence in the subsequent summary tables. Standardized characterization of MSCs (per ISCT criteria) and EVs (per MISEV2023), together with harmonized definitions of obesity, would substantially strengthen the comparability and translational value of future work^[6,62].

To connect the mechanistic sections above with the therapeutic strategies that follow, Table 1 maps each principal molecular derangement in obese-MSCs to its downstream functional consequence and the candidate intervention that targets it.

Increasing understanding of the mechanisms underlying obese-MSC dysfunction has driven development of therapeutic strategies to blunt it [Table 2]. Metabolic and genetic reprogramming, mitochondrial protection, preconditioning, and allogeneic or EV-based therapies are potentially promising avenues for restoring function to obesity-impaired MSCs. Importantly, evidence for these strategies remains predominantly preclinical, derived from in vitro MSC cultures and animal models. None has been tested in a clinical trial specifically designed to rescue obesity-induced MSC dysfunction. Table 2 therefore identifies the experimental model and evidence level for each approach.

AMPK activation acts as a metabolic switch, restoring mitochondrial biogenesis and reducing oxidative stress in MSCs. Metformin reverses senescence-associated dysfunction via AMPK signaling, enhancing MSC migration, immunomodulatory capacity, and therapeutic efficacy in models of metabolic syndrome and diabetes^[85,86]. In aging human adipose MSCs, metformin reduces oxidative stress and senescence while restoring adipogenic capacity and insulin sensitivity through AMPK activation^[87]. These data support metformin (and other AMPK activators such as AICAR) as candidates to counteract obesity-induced MSC alterations. Modulating mTOR signaling may rebalance MSC lineage commitment and restore stemness: the DEP-domain-containing mTOR-interacting protein (DEPTOR) promotes adipogenesis while suppressing osteogenesis, whereas DEPTOR ablation restores bone mass and reduces marrow fat^[88]. Targeted mTOR modulation may counter obesity-induced MSC dysfunction, although direct evidence for functional rescue remains limited and warrants further investigation.

NAD⁺ precursors such as nicotinamide riboside (NR) or nicotinamide mononucleotide (NMN) enhance SIRT1 and sirtuin-3 (SIRT3) activity, improve mitochondrial function, reduce senescence markers, and restore differentiation potential in MSCs^[89,90]. NR also enhances MSC stemness, lowers ROS, and inhibits terminal adipocyte differentiation^[91]. Collectively, these findings provide a strong mechanistic rationale for testing NAD⁺ repletion in obesity-impaired MSCs, although direct rescue studies in obese-donor MSCs remain limited. Several mitochondrial-targeted approaches have shown potential. Pyrroloquinoline quinone (PQQ) rescues cardiolipin content, mitophagy, and mitochondrial donation, improving therapeutic efficacy^[37], whereas silencing miR-181a improves mitochondrial gene expression and reduces oxidative stress in models of mitochondrial dysfunction^[27]. MOTS-c has broader metabolic and mitochondrial effects in experimental systems, collectively supporting MSC stemness, differentiation, and paracrine function^[29]. Yet, direct evidence that MOTS-c restores the function of obesity-impaired MSCs is currently lacking. Genetic modification offers a complementary route: AT-MSCs engineered to modulate regulators such as tumor necrosis factor-stimulated gene-6 (TSG-6) have been proposed to correct inflammatory phenotypes and metabolic complications of obesity, although data remain scarce^[92].

Because senescence is a convergent endpoint of the mechanisms described above, reducing senescent-cell burden represents a conceptually distinct strategy. Removal of senescent cells improves metabolic function in obese mice^[93]. Genetic or pharmacologic senolysis improves glucose handling and insulin sensitivity, restores adipogenesis, and reduces circulating inflammatory mediators and monocyte trafficking, with parallel gains in renal and cardiac functional measures^[93]. The implication is that senescent burden is not merely a marker of the obese niche but a causal contributor to it. Whether senolytic pre-treatment of a prospective donor, or senomodulation of harvested cells ex vivo, can restore the reparative competence of obese-MSCs has not been directly tested and represents a logical next step.

Hypoxic preconditioning enhances MSC cardioprotection by inducing leptin expression, while pro-inflammatory cytokines [tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ)] promote an immunosuppressive MSC phenotype^[94]. Obesity-related adiponectin deficiency impairs MSC survival and function; priming obese-MSCs with globular adiponectin restores glycolysis via the adiponectin receptor-1 (AdipoR1)/p38/hypoxia-inducible factor-1α (HIF-1α) axis, boosts immunoregulatory gene expression, and improves therapeutic efficacy^[95]. These findings support hypoxia-, cytokine-, and adipokine-based metabolic reprogramming strategies. In contrast to autologous obese-MSCs, allogeneic MSCs from lean, healthy donors ameliorate obesity, improve insulin sensitivity, enhance mitochondrial biogenesis and thermogenesis, and reduce inflammation^[96,97]. Lean-MSC-EVs further mitigate diabetes- and obesity-induced muscle atrophy via AMPK/Unc-51-like kinase-1 (ULK1)-mediated autophagy^[96]. These data support using allogeneic MSCs or engineered EVs when autologous cells are compromised.

A glaring weakness of the strategies above is that none has advanced to human testing for the specific goal of restoring the competence of obesity-impaired MSCs. The clinical experience that exists concerns MSC therapy for metabolic disease as an indication, rather than restoration of endogenous MSC function, and was not designed to address the donor-quality problem.

A 2025 systematic review and meta-analysis included 13 randomized trials in type-1 and type-2 diabetes involving 507 patients. At 12 months, MSC therapy was associated with lower glycated hemoglobin (HbA1c) [mean difference -0.72%; 95% confidence interval (CI) -1.11 to -0.33], a reduced daily insulin requirement (-14.50 U/day; 95%CI -19.45 to -9.55), higher fasting C-peptide (+0.24 ng/mL; 95%CI 0.05 to 0.43), and lower postprandial glucose. Fasting glucose, by contrast, did not separate from control, and between-study heterogeneity was high for C-peptide and fasting glucose in particular, so these pooled estimates should be read as encouraging rather than definitive^[98]. A separate synthesis of type-2 diabetes trials reported reduced anti-diabetic drug requirements over 12 months with effective doses ranging from approximately 1 × 10⁶ to 3.7 × 10⁶ cells/kg and no severe adverse events^[99]. Reported routes are predominantly intravenous or intra-pancreatic, and optimal dose, route, and timing remain unsettled.

Two caveats are essential for interpreting these results. First, both autologous and allogeneic sources have been used across trials, and their relative efficacy remains unresolved; at least one pooled analysis has reported more favorable outcomes with autologous than allogeneic MSCs, whereas subgroup analyses elsewhere have favored other cell types entirely, and certainty of evidence has been graded low to very low in guideline-oriented reviews^[98,99]. Second, and more pertinently, these trials were not designed to test whether donor adiposity degrades cell potency: donor BMI and metabolic status are rarely reported, participants are not stratified by donor obesity, and potency assays that might detect the relevant deficits are not routinely applied^[84]. The clinical literature therefore neither confirms nor refutes the translational relevance of obesity-induced MSC dysfunction.

Cell-free EV therapeutics are at an earlier clinical stage still. A systematic appraisal of the registered EV trial landscape identified 471 studies spanning more than 200 conditions, the vast majority being diagnostic; among the smaller therapeutic subset, MSC-derived EVs predominated and respiratory indications were the most common^[100]. Narrowing to MSC-EV therapeutics registered over the decade to 2024 yields 66 trials, across which neither potency criteria nor dose selection follow any shared framework^[101]. No registered trial has evaluated engineered or lean-donor EVs specifically to address obesity-induced MSC or EV dysfunction. Likewise, MSC-directed conditioning with AMPK activators, NAD⁺ precursors, PQQ, anti-mitomiR-181a, adiponectin, or senolytics has not been clinically tested for this purpose. Several of these agents have independent human safety data, which may lower the barrier to future MSC-focused trials, but efficacy for MSC rescue in humans is unproven. Bridging this gap will require standardized MSC and EV characterization, validated potency assays, uniform definitions of the obese phenotype, routine reporting of donor metabolic status, and trial designs that assess restored reparative function rather than surrogate metabolic endpoints alone.

The conclusions that can be drawn from current evidence remain tentative due to several methodological and empirical limitations within the existing body of evidence. First, reversibility remains unsettled: existing evidence for a durable “obesogenic memory” in adipose tissue and MSCs after weight loss^[15,57-59] conflicts with recent single-nucleus evidence that senescence in adipose precursor and vascular cells is substantially reversed by weight loss^[60]. Reconciling these findings will require studies specifying the phenotype, cell compartment, and timescale being assessed. Second, much of the mechanistic literature is cross-sectional and associative, derived from small human cohorts or from animal models whose metabolic phenotype does not necessarily map onto human obesity, limiting causal and translational inference. Third, heterogeneity in tissue source, species, obesity definition, and MSC/EV characterization constrains cross-study comparison and formal meta-analysis, and inter-donor variability persists even under harmonized manufacturing^[9]. Fourth, most proposed rescue strategies have been validated only on isolated readouts, such as a single mitochondrial or senescence marker, rather than on integrated in vivo restoration of reparative function, and standardized potency assays are lacking. Fifth, the epitranscriptomic layer discussed above is inferred largely from adipose tissue and from osteoporosis and aging models; direct m⁶A profiling of MSCs from donors with obesity has not been reported. Finally, the near-absence of clinical data for MSC-directed conditioning in obesity means that safety, durability, and efficacy in humans are untested, and existing trials do not report the donor characteristics that would make them informative here.

Addressing these gaps will require prospective, adequately powered studies with uniform definitions of obesity and standardized MSC (ISCT) and EV (MISEV2023) characterization. It will also require longitudinal designs that test reversibility before and after weight loss or pharmacologic conditioning, with pre-specified phenotypes; head-to-head comparison of autologous rescue vs. allogeneic healthy-donor approaches; routine reporting of donor metabolic status; and functional, in vivo endpoints of tissue repair. Defining validated biomarkers of “MSC competence” would additionally allow patient selection and monitoring in future trials^[6,62].

Obesity profoundly alters MSC biology and can compromise their reparative potential. The obese microenvironment, characterized by chronic metabolic stress, lipotoxicity, oxidative stress, and inflammation, induces mitochondrial dysfunction, epigenetic remodeling, and dysregulation of key metabolic pathways including AMPK-mTOR signaling and NAD⁺ homeostasis. These changes impair MSC bioenergetics, reduce stemness, promote senescence, and shift MSCs toward a pro-inflammatory phenotype. In parallel, obesity alters the molecular cargo and biological activity of MSC-derived EVs, further limiting their capacity to mediate tissue repair.

Consequently, obese-MSCs exhibit reduced angiogenic, immunomodulatory, and regenerative capacity and demonstrate diminished therapeutic efficacy in several preclinical disease models. These findings raise important considerations for the use of autologous MSC therapies in individuals with obesity. Future research should define the reversibility of obesity-induced MSC dysfunction, identify biomarkers of stem cell competence, and optimize therapeutic strategies tailored to metabolically compromised patients. Emerging approaches, including metabolic reprogramming, mitochondrial protection, NAD⁺ replenishment, and use of allogeneic MSCs or engineered EVs, may help restore stem cell function and improve regenerative outcomes, but translation will depend on standardized characterization and rigorous clinical evaluation.