Mesenchymal stem cells and extracellular vesicles for MAFLD: from biological mechanisms to translational prospects

Abstract

Metabolic dysfunction-associated fatty liver disease (MAFLD) is a prevalent metabolic disorder that can progress from simple steatosis to chronic inflammation and fibrosis. Currently, pharmacological agents for MAFLD remain limited, and lifestyle-based management often fails to halt disease progression, underscoring the need for multitarget therapeutic strategies. Mesenchymal stem cells (MSCs) represent a promising regenerative and immunomodulatory therapy capable of reprogramming the hepatic inflammatory microenvironment, improving metabolic homeostasis, and limiting fibrogenesis. Increasing evidence indicates that MSC efficacy is largely mediated through paracrine mechanisms, with MSC-derived small extracellular vesicles (MSC-EVs) acting as key effectors by delivering functional cargos, including microRNAs (miRNAs) and proteins. MSC-EVs attenuate inflammation via inflammasome inhibition and macrophage polarization, restore insulin sensitivity and lipid handling through AMP-activated protein kinase (AMPK)-associated metabolic signaling, and also limit fibrogenesis by suppressing hepatic stellate cell (HSC) activation as part of their integrated multitarget actions. This review integrates current insights into MAFLD pathogenesis and the therapeutic mechanisms of MSCs and MSC-EVs, and highlights translational priorities for their clinical development.

Introduction

The pathogenesis of MAFLD is best explained by the “multiple-hit” theory, wherein metabolic, genetic, inflammatory, and gut microbiota-related Metabolic dysfunction-associated fatty liver disease (MAFLD), which evolved from the concept of nonalcoholic fatty liver disease (NAFLD), is a systemic disorder characterized by hepatic steatosis concurrent with metabolic dysfunction [1–4]. The diagnosis of MAFLD is established based on imaging or histological evidence of hepatic steatosis in conjunction with at least one of the following criteria: overweight/obesity, type 2 diabetes mellitus, or evidence of metabolic dysfunction. Compared to NAFLD, the MAFLD definition more accurately captures the pivotal role of metabolic abnormalities in disease initiation and progression [5–7].

The nomenclature of fatty liver disease has evolved substantially in recent years. MAFLD was proposed to highlight the pathogenic role of metabolic dysfunction and to overcome the limitations of the exclusion-based NAFLD definition. More recently, an international Delphi consensus recommended MASLD as the updated umbrella term, with MASH denoting its progressive inflammatory subtype [8, 9]. Although MAFLD and MASLD substantially overlap in identifying fatty liver disease associated with metabolic dysfunction, they are not fully interchangeable with respect to their definitions [8, 9]. In this review, we retain the term MAFLD because it remains widely used in the preclinical and translational literature discussed here, while acknowledging the field’s ongoing transition toward MASLD.

The pathogenesis of MAFLD is best explained by the “multiple-hit” theory, wherein metabolic, genetic, inflammatory, and gut microbiota-related factors synergistically drive disease development (Fig. 1). Among these, insulin resistance (IR) and consequent hepatic lipid accumulation are key upstream events. In the context of obesity, unhealthy dietary patterns, and genetic susceptibility, impaired insulin sensitivity promotes enhanced adipose tissue lipolysis, resulting in an increased portal influx of free fatty acids (FFAs) into the liver. Concurrently, an imbalance between upregulated de novo lipogenesis, mediated in part by sterol regulatory element-binding protein 1c (SREBP-1c), acetyl-CoA carboxylase (ACC), and fatty acid synthase (FASN), and impaired fatty acid β-oxidation, involving carnitine palmitoyltransferase 1 (CPT-1) and peroxisome proliferator-activated receptor alpha (PPARα), drives excessive triglyceride and lipid droplet accumulation within hepatocytes [10–12]. Excess saturated fatty acids and lipotoxic species, such as ceramides, induce endoplasmic reticulum (ER) stress and mitochondrial dysfunction, activate the protein kinase RNA-like endoplasmic reticulum kinase–eukaryotic initiation factor 2 alpha–C/EBP homologous protein (PERK–eIF2α–CHOP) signaling pathway, and promote reactive oxygen species (ROS) overproduction, ultimately exacerbating lipid peroxidation and hepatocellular damage [13, 14]. During the transition from simple steatosis to nonalcoholic steatohepatitis (NASH), oxidative stress and inflammation play pivotal roles. Injured hepatocytes release damage-associated molecular patterns (DAMPs), which activate resident Kupffer cells and infiltrating monocytes, thereby amplifying the production of proinflammatory cytokines, including tumor necrosis factor alpha (TNF-α), interleukin-6 (IL-6), and interleukin-1 beta (IL-1β), through toll-like receptor 4/nuclear factor kappa B (TLR4/NF-κB), mitogen-activated protein kinase (MAPK), and related signaling pathways. This persistent inflammatory milieu not only exacerbates hepatocellular injury but also fosters hepatic stellate cell (HSC) activation and collagen synthesis, initiating and driving the progression of liver fibrosis [12, 15, 16].

Fig. 1.

The pathogenesis of MAFLD. Insulin resistance–associated lipid overload triggers lipotoxicity and oxidative stress, resulting in hepatocellular injury and Kupffer cell–mediated inflammation. Sustained inflammatory signaling promotes hepatic stellate cell activation and extracellular matrix deposition, driving fibrosis progression. In parallel, gut dysbiosis and LPS translocation via the gut–liver axis further exacerbate hepatic inflammation and fibrogenesis. Figure designed by Adobe Illustrator CC 2020

Gut dysbiosis represents another critical contributor to MAFLD pathogenesis, characterized by reduced microbial diversity, and an altered Firmicutes-to-Bacteroidetes ratio, and an increased abundance of proinflammatory microbial taxa. Microbiota-derived metabolites, including short-chain fatty acids (SCFAs), bile acids (BAs), and lipopolysaccharide (LPS), modulate hepatic metabolism and inflammatory responses through the gut-liver axis. Impairment of the intestinal barrier, such as reduced expression of ZO-1 and occludin, permits the translocation of LPS and other microbial products to the liver, triggering the TLR4/NF-κB signaling pathway and amplifying inflammatory and fibrotic responses [17, 18].

Collectively, MAFLD is a complex systemic disease centered on insulin resistance, driven by metabolic dysregulation, and culminating in chronic inflammation and fibrosis. Its pathogenesis involves multiple interconnected molecular pathways, including TLR4/NF-κB, c-Jun N-terminal kinase (JNK), AMPK, peroxisome proliferator-activated receptor alpha (PPARα), and transforming growth factor beta/Smad (TGF-β/Smad) signaling, as well as complex multicellular interactions among hepatocytes, Kupffer cells, HSCs, and intestinal epithelial cells, forming a highly integrated pathogenic network [4–7].

To date, pharmacological agents for the broad MAFLD remain limited. Notably, in 2024 the U.S. Food and Drug Administration (FDA) granted accelerated approval to resmetirom (Rezdiffra) for adults with noncirrhotic NASH with moderate to advanced fibrosis (F2-F3). Current therapeutic strategies primarily focus on lifestyle modification and the management of metabolic comorbidities using antidiabetic, lipid-lowering, and weight-reducing agents, such as insulin sensitizers, glucagon-like peptide-1 (GLP-1) receptor agonists, and sodium-glucose cotransporter 2 (SGLT2) inhibitors (Table 1). Although these approaches can improve metabolic parameters and hepatic steatosis in selected patients, their efficacy in halting disease progression, particularly with regard to inflammation and fibrosis, remains limited. Moreover, long-term drug use is often associated with suboptimal adherence and potential adverse effects [27–30]. These limitations highlight the urgent need for novel therapeutic strategies capable of simultaneously targeting metabolic dysregulation, inflammation, and fibrogenesis [21, 23, 27].

Table 1.

Current therapeutic approaches for MAFLD/NASH and their limitations

Therapeutic class	Representative agents	Primary mechanism(s)	Main clinical benefit(s)	Key limitations	Clinical evidence stage
Antioxidants	Vitamin E	Reduces oxidative stress	Improves steatosis and inflammation in non-diabetic NASH	Limited antifibrotic effect; long-term safety concerns [19, 20]	Phase III RCTs; guideline-restricted recommendation (off-label)
PPAR Agonists	Pioglitazone	Improves insulin sensitivity; modulates lipid metabolism	Improves steatosis and inflammation	Weight gain; limited fibrosis regression [21, 22]	Phase III RCTs; guideline-supported (off-label for NASH)
GLP-1–Based Agents	Liraglutide, Semaglutide	Weight loss; improved insulin sensitivity	Promotes NASH resolution	Minimal direct antifibrotic effect [21, 23, 24]	Phase II–III RCTs; not yet approved for NASH
FXR Agonists	Obeticholic acid, Tropifexor	Regulates bile acid and lipid metabolism	Shows fibrosis improvement signals in selected trials	Pruritus; lipid abnormalities; safety concerns [23, 24]	Phase II–III; inconsistent Phase III outcomes
THR-β Agonists	Resmetirom	Enhances fatty acid oxidation	Reduces liver fat; NASH resolution without fibrosis worsening	Long-term outcomes under evaluation [21, 24]	Phase III positive; FDA-approved (2024)
Gut Microbiota–Targeted	Probiotics, Prebiotics	Modulates gut–liver axis	Variable improvements in metabolic and inflammatory markers	Limited histological evidence [17, 18]	Preclinical–early Phase II; exploratory
Anti-inflammatory / Antifibrotic	Selonsertib, Cenicriviroc	Inhibits inflammatory or fibrotic signaling	Transient biomarker improvement	Failed to show clinical benefit [21, 25, 26]	Phase III trials negative (development discontinued)

MAFLD, metabolic dysfunction–associated fatty liver disease; NASH, nonalcoholic steatohepatitis; RCT, randomized controlled trial; PPAR, peroxisome proliferator-activated receptor; GLP-1, glucagon-like peptide-1; FXR, farnesoid X receptor; THR-β, thyroid hormone receptor beta

Against this backdrop, mesenchymal stem cells (MSCs) have emerged as a promising therapeutic candidate due to their multipotent differentiation capacity, potent immunomodulatory effects, and pro-regenerative properties. Accumulating preclinical evidence suggests that MSCs and their derivatives can ameliorate MAFLD by improving insulin sensitivity, reducing hepatic lipid accumulation, alleviating oxidative stress, suppressing pro-inflammatory cytokine production, and inhibiting HSC activation [25–27].

This review provides an overview of the core pathogenic mechanisms underlying MAFLD and highlights the therapeutic potential of MSC-based interventions. We focus on the multifaceted molecular mechanisms by which MSCs exert their beneficial effects, including immunomodulation, metabolic regulation, oxidative stress alleviation, and, as one component, the suppression of fibrogenesis via pathways such as TGF-β/Smad, as well as their role in regulating intercellular crosstalk within the hepatic microenvironment.

Biological properties and therapeutic potential of MSCs

MSCs are multipotent adult stromal cells that can be isolated from a variety of tissues, including bone marrow, adipose tissue, umbilical cord, and placenta. According to the criteria established by the International Society for Cell and Gene Therapy (ISCT), MSCs are defined by their adherence to plastic under standard culture conditions, expression of characteristic surface markers (CD73, CD90, and CD105), absence of hematopoietic markers, and capacity for tri-lineage differentiation into osteoblast, chondrocytes, and adipocytes. Beyond their differentiation potential, MSCs are primarily valued for their robust paracrine activity and potent immunomodulatory properties [34, 35], which form the mechanistic basis for their therapeutic efficacy in metabolic and inflammatory disorders such as MAFLD [31, 32, 36].

MSCs exert hepatoprotective effects through coordinated modulation of inflammation, oxidative stress, fibrogenesis, and lipid metabolism. Preclinical studies consistently demonstrate that MSC administration attenuates hepatic inflammation, as evidenced by reduced levels of pro-inflammatory cytokines, including TNF-α, IL-1β, and IFN-γ. Given the central role of macrophages in MAFLD progression, MSCs promote a phenotypic shift from pro-inflammatory M1 macrophages toward anti-inflammatory M2 macrophages. This polarization is mediated through the secretion of immunoregulatory cytokines such as IL-4 and IL-10, as well as suppression of the NF-κB signaling, thereby mitigating hepatic inflammation and tissue injury [12, 16, 37].

In parallel, MSCs alleviate oxidative stress and inhibit fibrogenesis via multiple complementary mechanisms. MSC-derived factors, including hepatocyte growth factor (HGF), IL-10, and antioxidant enzymes, enhance endogenous cellular antioxidant defenses and reduce lipid peroxidation. Moreover, MSCs suppress HSC activation by inhibiting the TGF-β/Smad signaling cascade, thereby limiting excessive extracellular matrix deposition. Additionally, by activating the AMPK/PPARα pathway, MSCs enhance fatty acid β-oxidation, reduce intracellular lipid droplet accumulation, and restore hepatocellular metabolic homeostasis [33, 38, 39].

MSCs derived from different tissue sources exhibit distinct biological and functional profiles. Bone marrow–derived MSCs (BM-MSCs) possess strong immunomodulatory and regenerative capacities but are limited by invasive harvesting procedures and relatively modest proliferative capacity [40–42]. Adipose-derived MSCs (ADSCs) offer greater accessibility, high cell yield, and robust paracrine activity; preclinical evidence supports their effectiveness in ameliorating hepatic steatosis and improving insulin sensitivity in MAFLD models. Umbilical cord–derived MSCs (UC-MSCs), by contrast, display low immunogenicity, high proliferative capacity, and pronounced anti-inflammatory and antifibrotic properties, making them a particularly promising candidate for clinical translation in the context of MAFLD [43–46]. To better illustrate these source-dependent differences and inform therapeutic selection, we summarize the key characteristics of BM-MSCs, ADSCs, and UC-MSCs relevant to MAFLD treatment in Table 2. Although these MSC populations share core mesenchymal features, differences in tissue origin may affect their secretory profile, metabolic activity, immunomodulatory potency, and translational feasibility. Therefore, comparative evaluation of MSC sources is important for optimizing cell-based therapeutic strategies for MAFLD.

Table 2.

Comparative characteristics of MSCs from different tissue sources for MAFLD therapy

MSC source	Accessibility	Proliferative capacity	Immunogenicity	Main therapeutic relevance in MAFLD	Translational advantage	Main limitations
Bone marrow-derived MSCs (BM-MSCs)	Low (Invasive aspiration)	Moderate	Moderate	Immunomodulation; tissue repair; apoptosis inhibition	Well characterized; established isolation and culture protocols	Invasive harvest; relatively low yield; limited proliferative capacity
Adipose-derived MSCs (ADSCs)	High (Liposuction waste)	High	Low	Metabolic regulation; improvement of insulin resistance and steatosis	Easily accessible; high yield; strong paracrine activity	Donor metabolic status may affect cell quality; source heterogeneity
Umbilical cord-derived MSCs (UC-MSCs)	Highest (Non-invasive, discarded tissue)	Very high	Very low	Anti-inflammatory and antifibrotic effects; attenuation of metabolic injury	Scalable; suitable for allogeneic off-the-shelf use	Standardization of collection, expansion, and manufacturing remains needed

MSCs, mesenchymal stromal/stem cells; UC-MSCs, umbilical cord–derived MSCs; BM-MSCs, bone marrow–derived MSCs; ADSCs, adipose-derived stromal/stem cells

The therapeutic efficacy of MSC-based therapy is critically dependent on cell survival and homing efficiency. Following systemic infusion, a substantial proportion of administered MSCs become entrapped in the pulmonary vasculature, resulting in limited engraftment within the liver. To address these challenges, several strategies have been developed, including hypoxic preconditioning, pharmacological pretreatment, inflammatory cytokine priming, and genetic or surface engineering. These approaches have been shown to significantly improve MSC survival, migratory capacity, and immunomodulatory function, thereby augmenting therapeutic outcomes in experimental models of MAFLD [47–49].

Collectively, owing to their pleiotropic immunomodulatory, antifibrotic, and metabolic regulatory properties, MSC-based therapies represent a promising experimental strategy for the treatment of MAFLD and related metabolic liver diseases. Accumulating preclinical data and emerging early-phase clinical trials indicate that MSC administration can effectively ameliorate hepatic steatosis, inflammation, and fibrogenesis. Importantly, growing evidence suggests that MSC-derived products, particularly extracellular vesicles (MSC-EVs), recapitulate many of the therapeutic benefits of their parent cells while offering potential advantages in terms of safety, stability, and manufacturing consistency [33, 50–52].

Clinical studies of MSC-based cell therapy in MAFLD

Given the multifaceted pathogenesis of MAFLD, which involves dysregulation across multiple signaling pathways and pathological processes, including metabolic imbalance, chronic inflammation, oxidative stress, and fibrogenesis, conventional pharmacological therapies often fall short of simultaneously targeting these interconnected mechanisms, leading to suboptimal clinical outcomes. As a result, an increasing number of researchers have shifted focus toward systemic therapeutic strategies centered on MSCs or MSC-derived products. A substantial body of preclinical evidence has shown that MSC administration significantly improves liver function, reduces lipid accumulation, attenuates inflammatory responses, and ameliorates fibrosis in animal models of MAFLD [31–33, 36].

To assess the current status of clinical translation, a comprehensive search of ClinicalTrials.gov was conducted using disease-related terms including NAFLD, MAFLD, MASLD, NASH, MASH, cirrhosis, liver failure, decompensated cirrhosis, and HBV-related liver cirrhosis, in combination with MSC-based interventions (as of November 20, 2025). This search yielded a total of 120 registered MSC-related clinical trials involving liver diseases.

After exclusion of ongoing or protocol-based trials without available efficacy data, 27 studies with published preliminary or final clinical results were included and summarized (Table 3). The majority of these studies are early-phase I/II trials, primarily designed to assess safety and preliminary efficacy. Among these, human umbilical cord–derived MSCs (hUC-MSCs) and bone marrow–derived MSCs (BM-MSCs) represent the most frequently utilized cell sources. A smaller number of trials have explored alternative approaches, such as autologous MSCs or combined infusion of MSCs with hematopoietic stem cells [36].

Table 3.

Clinical studies of MSC-based therapy for liver diseases

NCT Number	Cell source	Route	Sample size	Endpoints	Key findings	Disease type	Phase
NCT05442437	hUC-MSCs	IV	24	Safety; liver function	Transient liver function improved	HBV decompensated cirrhosis	I
NCT04522869	hUC-MSCs	Hepatic artery	32	Safety; cholestatic enzymes	Alb, TBIL, and ALP improved	Biliary atresia	I/II
NCT04243681	Autologous MSC/HSC	Hepatic artery	5	Safety	No serious adverse events observed	Decompensated cirrhosis	IV
NCT02957552	BM-MSCs	PV/IV	7	Safety; feasibility	Well tolerated; immune changes observed	Pediatric liver transplantation	I
NCT02260375	Allogeneic BM-MSCs	IV	20	Immune modulation	Treg and NK cell changes observed	Liver transplantation	I
NCT02223897	hUC-MSCs	IV	66	Cholestatic enzymes	Transient TBIL, GGT, ALP reduction	Biliary diseases	II/III
NCT01875081	BM-MSCs	Hepatic artery	72	Fibrosis; Child–Pugh	Fibrosis and Child–Pugh score improved	Alcoholic cirrhosis	II
NCT01741090	BM-MSCs	Hepatic artery	11	Liver histology	Histological improvement reported	Alcoholic cirrhosis	II
NCT01724398	hUC-MSCs	IV	120	Liver function	Liver function improved	Liver failure	I/II
NCT01662973	hUC-MSCs	IV	7	Cholestatic enzymes	ALP and GGT reduced	Biliary cirrhosis	I/II
NCT01499459	BM-MSCs	IV	25	Liver function; MELD	Albumin and MELD improved	Cirrhosis	NA
NCT01429038	BM-MSCs	IV	10	Safety	No unexpected safety concerns	Liver transplantation	I/II
NCT01322906	BM-MSCs	IV	110	Liver function; infections	Liver function improved; infections reduced	HBV liver failure	II
NCT01220492	hUC-MSCs	IV	219	Overall survival	Overall survival improved	HBV decompensated cirrhosis	I/II
NCT01218464	hUC-MSCs	IV	43	Survival; liver function	Survival and liver function improved	Liver failure	I/II
NCT00956891	BM-MSCs	Hepatic artery	527	Liver function	Short-term improvement observed	HBV liver failure	Observational
NCT00420134	BM-MSCs	PV/IV	8	Safety; MELD	MELD score improved	Cirrhosis	I/II
NCT01729221	Autologous BM-MSCs	IV	40	Liver function	Albumin, bilirubin, ALT, prothrombin concentration and INR improved; well tolerated	HCV-related cirrhosis	Observational
NCT01560845	BM-MSCs	Hepatic artery	56	MELD; immune balance	Liver function and Treg/Th17 balance improved	Cirrhosis	II/III
NCT00476060	Autologous BM-MSCs	IV	4	Safety; biodistribution	Peripheral infusion safe; hepatic accumulation observed	Decompensated cirrhosis	I
NCT05227846	Allogeneic MSCs	IV	15	Safety; dose escalation	Well tolerated across dose levels	Decompensated cirrhosis	I
NCT05984303	Allogeneic MSCs	IV	9	Safety; multiple-dose feasibility	Multiple-dose administration well tolerated	Decompensated cirrhosis	I
NCT02706132	hUC-MSCs	IV	43	Survival; liver function	Short-term survival and liver function improved	Acute-on-chronic liver failure	I/II
NCT01661842	hUC-MSCs	IV	60	Safety; liver function; Child–Pugh score	Ascites reduced; liver function improved	HBV-related decompensated cirrhosis	I/II
NCT05507762	hUC-MSCs	IV	20	Safety; liver fibrosis; Child–Pugh score	Well tolerated; no significant difference in fibrosis score between groups	HBV-related compensated cirrhosis	I/II
NCT01342250	hUC-MSCs	IV	20	Safety; MELD; quality of life	Well tolerated; MELD and MELD-Na scores decreased	Decompensated cirrhosis	I
NCT01062750	Autologous ADRCs (adipose-derived MSC-enriched stromal cells)	Hepatic artery	4	Safety; liver function	No serious adverse events observed	viral and NASH-related cirrhosis	I

MSCs, mesenchymal stromal/stem cells; UC-MSCs, umbilical cord–derived MSCs; BM-MSCs, bone marrow–derived MSCs; IV, intravenous; PV, portal vein; Alb, albumin; TBIL, total bilirubin; GGT, gamma-glutamyl transferase; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; MELD, Model for End-Stage Liver Disease; Child–Pugh score, a clinical scoring system assessing cirrhosis severity based on bilirubin, albumin, INR, ascites, and encephalopathy

However, clinical evidence specifically targeting MAFLD/NAFLD remains limited. To date, although several clinical studies have included patients with NAFLD or metabolic liver disease, well-designed clinical trials primarily targeting MAFLD/NASH populations and reporting robust efficacy outcomes remain insufficient. Nevertheless, several exploratory studies have begun to provide preliminary human data. A post hoc analysis of a randomized, double-blind, placebo-controlled trial demonstrated that intravenous infusion of human umbilical cord–derived MSCs (UC-MSCs) improved hepatic steatosis assessed by ultrasonography, reduced serum alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT), and triglyceride levels, and enhanced insulin sensitivity in patients with type 2 diabetes mellitus complicated by NAFLD [53]. Notably, fibrosis-related indices such as the fibrosis-4 index (FIB-4) were not significantly altered, likely reflecting the early-stage disease profile of the enrolled population.

In addition, a phase I clinical trial using freshly isolated autologous adipose tissue–derived regenerative cells (ADRCs) administered via the hepatic artery reported improvements in serum albumin and prothrombin activity in patients with NASH- or fatty liver disease–related cirrhosis, with no treatment-related adverse events observed [54]. Although limited by small sample size and the absence of a control group, this study provides early clinical evidence supporting the feasibility of MSC-based therapy in advanced metabolic liver disease.

Notably, current MSC-based clinical studies are predominantly focused on end-stage liver diseases, including cirrhosis, liver failure, alcoholic liver disease, and post-transplantation complications [55]. In contrast, large-scale clinical trials targeting MAFLD or nonalcoholic steatohepatitis (NASH) remain scarce. Available evidence indicates that MSC therapy is generally well tolerated across various liver disease settings and is associated with varying degrees of improvement in key clinical parameters, including serum albumin, bilirubin levels, prothrombin activity, as well as Model for End-Stage Liver Disease (MELD) score and Child–Pugh score. Beneficial effects on inflammatory markers and antifibrotic outcomes have also been reported in a subset of studies [36, 54, 56, 57]. Overall, current data support a generally favorable short-term safety profile and provide preliminary evidence of clinical benefit for MSC-based therapy, as well as the fact that most published studies have been conducted in advanced rather than early-stage metabolic liver disease. However, the overall quality of this evidence is limited by small sample sizes and considerable heterogeneity in study designs, cell sources, dosing regimens and outcome assessment criteria [36].

A critical barrier to clinical translation is that most existing MSC-based trials in liver disease have been designed for acute or end-stage settings, such as decompensated cirrhosis or liver failure, rather than for early-stage MAFLD/MASH [36]. This distinction is important because MAFLD/MASH is characterized by chronic, low-grade metabolic inflammation, slower disease progression, and substantial inter-patient heterogeneity, all of which demand a different therapeutic and trial-design framework [58, 59]. Current evidence does not suggest that safety concerns alone are the main factor limiting early-stage MASH trials; rather, the challenge lies in the lack of a well-defined translational strategy for this disease stage [58]. Although MSC therapy has generally shown an acceptable short-term safety profile in advanced liver disease studies, the long-term safety of repeated or chronic dosing regimens, which may be required for sustained benefit in MAFLD/MASH, remains insufficiently characterized [36, 58]. Another major challenge is endpoint selection. Traditional histological endpoints remain invasive, costly, and difficult to apply repeatedly in long-term studies, thereby limiting patient recruitment, trial scalability, and feasibility [59, 60]. At the same time, non-invasive biomarkers are increasingly being used for risk stratification and trial enrichment, but their integration as validated efficacy endpoints is still evolving [59, 60]. More fundamentally, the transition from treating acute or end-stage liver injury, where the goal is rapid functional rescue, to managing chronic metabolic dysfunction requires re-evaluation of key therapeutic parameters, including dose, treatment frequency, delivery route, and clinically meaningful outcome measures [58]. Therefore, future MSC trials in MAFLD/MASH should move beyond simple extrapolation from end-stage liver disease studies and instead adopt disease-stage-appropriate designs incorporating non-invasive biomarkers, longitudinal metabolic and inflammatory readouts, quality-of-life measures, and clinically meaningful indicators of fibrosis progression and treatment response [58, 59].

These clinical observations indirectly corroborate the multifaceted therapeutic potential of MSCs, particularly their ability to modulate inflammation, promote hepatic tissue repair, and restore liver function, thereby offering a compelling theoretical rationale for their application in MAFLD. Nevertheless, given that the majority of existing trials have focused on patients with decompensated or end-stage liver disease, clinical studies specifically targeting MAFLD/NASH populations remain scarce. This gap underscores an important avenue for future translational research [36, 61].

Preclinical studies of MSC-based cell therapy in MAFLD

Preclinical evidence demonstrates that MSC-based interventions exert broad hepatoprotective effects in experimental models of MAFLD/NASH, thereby providing a robust foundation for both mechanistic exploration and translational development (Table 4).

Table 4.

Preclinical studies of MSC-based therapy in MAFLD/NASH

Cell source	Route of administration	Dose and regimen	Disease model	Animal	Mechanisms	Key outcomes	Ref.
hUC-MSCs	Tail vein injection	1 × 10⁶ cells, once weekly for 6 weeks	Leptin receptor–deficient / T2DM MAFLD	Mouse	Activation of the HNF4α–CES2 signaling pathway and regulation of lipid metabolism–related gene expression	Attenuation of hepatic steatosis and improvement of liver injury	[62]
BM-MSCs	Intravenous injection	1 × 10⁶ cells, two infusions	MAFLD	Mouse	Suppression of CD4⁺ T-cell proliferation	Reduction of hepatic fibrosis and inflammation	[63]
Human SHED-MSC–conditioned medium	Tail vein injection	0.5 mL SHED-CM, once weekly for 3 weeks	NASH	Mouse	Protection of intestinal barrier integrity and inhibition of TLR4 signaling via the gut–liver axis	Attenuation of hepatic fibrosis and inflammation	[64]
ADSCs	Intrasplenic injection	1 × 10⁵ cells, twice at 2-week intervals	NASH	Mouse	Activation of the Notch signaling pathway and upregulation of transcription factor HES1	Promotion of hepatocyte proliferation, reduced apoptosis, and improvement of fibrosis and inflammation	[43]
hUC-MSC–conditioned medium	Tail vein injection	200 µL MSC-CM every 3 days for 2 months	T2DM NAFLD	Mouse	Activation of SIRT1 signaling and downregulation of COX-2, ICAM-1, CSF-1, and TGF-β	Improved body weight, glucose metabolism, insulin sensitivity, mitochondrial function, hepatic steatosis, lipid accumulation, apoptosis, and inflammation	[48]
hUC-MSCs	Tail vein injection	5 × 10⁵ cells, three infusions at 3-week intervals	NAFLD	Mouse	HGF-mediated downregulation of hepatic RNF186 expression	Improvement of insulin resistance and glucose metabolism	[65]
BM-MSCs	Tail vein injection	2 × 10⁶ cells, two infusions at 2-week intervals	NAFLD	Rat	Enhancement of SERCA activity and restoration of Ca²⁺ homeostasis, leading to attenuation of endoplasmic reticulum stress	Reduced inflammation and hepatocyte pyroptosis with improved SERCA activity	[14]
Mitochondria-preconditioned hUC-MSCs	Tail vein injection	1 × 10⁶ cells, once weekly for 4 weeks	T2DM NAFLD	Mouse	Mitochondria-induced autophagy facilitating mitochondrial transfer from MSCs to hepatocytes	Greater reductions in blood glucose, liver enzymes, and triglycerides, with improved liver histology compared with unconditioned MSCs	[49]
hUC-MSCs	Tail vein injection	1 × 10⁶ cells, single dose	NASH	Mouse	Downregulation of ACOT family members (Acot1, Acot2, Acot3) and CYP4a family members (Cyp4a10, Cyp4a14), modulating fatty acid metabolism and PPAR signaling	Improvements in obesity, glucose tolerance, insulin sensitivity, hepatic lipid deposition, serum ALT/AST, inflammatory cytokines, and fibrosis-related gene expression	[45]
LPS-preconditioned ADSCs	Intraperitoneal injection	1.5 × 10⁶ cells, once weekly for 6 weeks	NAFLD	Rat	Reduction of inflammatory gene expression (IL-1β, IL-6, TNF-α, TGF-β) and reactive oxygen species	Improved obesity, ALT, AST, LDL-C, HDL-C, inflammatory gene expression, and oxidative stress	[65]
hUC-MSCs	Tail vein injection	1 × 10⁶ cells, two infusions at 4-week intervals	T2DM/NAFLD	Rat	Inhibition of the TLR4/NF-κB inflammatory pathway and attenuation of oxidative stress	Improvements in lipid metabolism, insulin resistance, inflammation, oxidative stress, and liver injury	[67]
Atg5-deficient ADSCs	Tail vein injection	1 × 10⁶ cells, three infusions at 2-week intervals	NAFLD	Mouse	Modulation of AMPKα/mTOR/S6K/SREBP1 signaling	Reduced lipid accumulation with improvement of liver function and inflammation	[68]
hUC-MSCs	Tail vein injection	1 × 10⁶ cells, single dose	MAFLD	Rat	Correction of diabetes-associated hepatic metabolic dysregulation and delay of MAFLD progression	Improved liver function, glucose and lipid metabolism, and hepatic injury	[46]
BM-MSCs	Tail vein injection	1 × 10⁶ cells, three infusions at 2-week intervals	NASH	Rat	Regulation of mitochondrial function, hypoxia, inflammation, and angiogenesis	Cardioprotective effects and improved mitochondrial function	[69]
ADSCs	Tail vein injection	1 × 10⁶ cells, once weekly for 3 weeks	NASH	Mouse	Suppression of endoplasmic reticulum stress and restoration of autophagic flux	Reduced ER stress, apoptosis, and improved autophagy	[44]
ADSCs	Tail vein injection	1 × 10⁶ cells, twice weekly for 4 weeks	Mc4r-KO NASH	Mouse	Mechanism not fully elucidated; phenotypic outcomes reported	Reduced ALT levels, inflammation, and fibrosis	[70]
hBM-MSCs	Intrasplenic injection	0.9–1 × 10⁶ cells, single dose	NASH	Mouse	Mitochondrial transfer–mediated metabolic rescue	Reduced lipid accumulation, improved lipid metabolism, inflammation, and fibrosis	[39]
u-ADSCs	Intrasplenic injection	1 × 10⁶ or 7.5 × 10⁵ cells, two infusions at 2-week intervals	NASH	Mouse	Mechanism not fully elucidated; phenotypic outcomes reported	Improved liver function with reduced inflammation and fibrosis	[71]

HFD, high-fat diet; MCD, methionine- and choline-deficient diet; T2DM, type 2 diabetes mellitus; ADSCs, adipose-derived stromal/stem cells; SHED, stem cells from human exfoliated deciduous teeth; HSCs, hepatic stellate cells; α-SMA, alpha-smooth muscle actin; ROS, reactive oxygen species; NF-κB, nuclear factor kappa B; TNF-α, tumor necrosis factor alpha; IL, interleukin. Other gene and protein abbreviations follow standard nomenclature

Steatosis and metabolic improvement

Across preclinical MAFLD/NASH models, MSC-based interventions consistently attenuate hepatic lipid accumulation and improve metabolic phenotypes through coordinated remodeling of lipid-handling networks. These effects are characterized by activation of metabolic regulators, such as the hepatocyte nuclear factor 4 alpha–carboxylesterase 2 (HNF4α–CES2) axis and sirtuin 1 (SIRT1) signaling, modulation of peroxisome proliferator-activated receptor (PPAR)-centered fatty acid metabolic programs, and suppression of key lipogenic checkpoints, including AMP-activated protein kinase alpha/mechanistic target of rapamycin/ribosomal protein S6 kinase/sterol regulatory element-binding protein 1 (AMPKα/mTOR/S6K/SREBP1), as well as lipid-handling gene families such as acyl-CoA thioesterase (ACOT) and cytochrome P450 4 A (CYP4A) members [45, 48, 68]. In leptin receptor–deficient, diabetes-associated MAFLD, hUC-MSC treatment ameliorated hepatic steatosis and liver injury, which was linked to activation of HNF4α–CES2 and reprogramming of lipid metabolism–related gene expression [62]. Likewise, in NAFLD/NASH settings, several studies reported parallel improvements in obesity-related phenotypes, glucose tolerance, insulin sensitivity, and hepatic lipid deposition following MSCs treatment [45, 46, 67]. Consistent with a paracrine mode of action, MSC-derived conditioned medium also produced robust metabolic benefits in type 2 diabetes mellitus (T2DM)-associated NAFLD, accompanied by SIRT1 activation and reduced inflammatory/fibrogenic mediators [48]. Moreover, functional tuning of MSC states may further strengthen anti-steatotic efficacy; for example, Atg5-deficient ADSCs reduced lipid accumulation via AMPKα/mTOR/S6K/SREBP1 signaling, supporting the feasibility of optimizing MSC biology to enhance metabolic outcomes [68].

Anti-inflammatory and immunomodulatory effects

Beyond metabolic regulation, MSC therapy exerts prominent anti-inflammatory and immunomodulatory activity in MAFLD/NASH models, typically involving suppression of innate inflammatory cascades (e.g., TLR4/NF-κB), reduction of pro-inflammatory cytokine output, and reprogramming of immune-cell phenotypes toward a more inflammation-resolving state [63, 66, 67]. In diabetes-associated NAFLD/MAFLD, hUC-MSCs improved metabolic and hepatic injury parameters while inhibiting the TLR4/NF-κB pathway, accompanied by attenuation of oxidative stress [67]. This anti-inflammatory activity may be reinforced by macrophage reprogramming toward an M2-like, inflammation-resolving phenotype, thereby reducing downstream cytokine production and hepatic inflammatory signaling [12, 16, 63, 66, 67]. In parallel, BM-MSC–mediated suppression of CD4⁺ T-cell proliferation was associated with reductions in hepatic inflammation and fibrosis, implicating adaptive immune regulation in limiting disease progression [63]. Preconditioning strategies further support anti-inflammatory potency; LPS-preconditioned ADSCs reduced inflammatory gene expression (including IL-1β, IL-6, TNF-α, and TGF-β) and improved metabolic and biochemical indices in NAFLD rats, alongside reduced oxidative stress [66]. Notably, immune modulation may also occur via extrahepatic inputs, particularly through restoration of the gut-liver axis [64, 72]. For example, conditioned medium derived from stem cells from human exfoliated deciduous teeth (SHED-MSC-conditioned medium) preserved intestinal barrier integrity by upregulating tight junction proteins, such as zonula occludens-1 (ZO-1) and occludin, and by inhibiting toll-like receptor 4/nuclear factor kappa B (TLR4/NF-κB) signaling, thereby reducing bacterial translocation and subsequent hepatic inflammation and fibrosis in NASH models [64]. Consistent with this concept, adipose-derived MSCs (ADSCs) have also been reported to improve intestinal homeostasis and reduce gut-derived inflammatory signaling in experimental settings [36]. In addition, MSC therapy may partially reverse diet-induced dysbiosis, including alterations in the Firmicutes/Bacteroidetes ratio and depletion of beneficial bacterial taxa such as Lactobacillus, changes that are associated with improved metabolic and hepatic outcomes [73]. Although systemic administration remains the predominant route, oral delivery of MSC-derived extracellular vesicles (MSC-EVs) is conceptually attractive as a gut-targeted strategy. However, this possibility remains largely exploratory, and its therapeutic feasibility may be influenced by gastrointestinal stability, epithelial uptake, and bioavailability [74, 75]. Collectively, these findings support the gut-liver axis as an important therapeutic target for MSC-based interventions in MAFLD.

Oxidative stress and mitochondrial homeostasis

Oxidative stress and mitochondrial dysfunction serve as major amplifiers of metabolic and inflammatory injury in MAFLD, and MSC therapy appears to engage this axis by restoring redox balance, improving mitochondrial homeostasis, and—under certain conditions—facilitating intercellular mitochondrial transfer as a form of metabolic rescue [39, 49, 67]. In T2DM-associated NAFLD, hUC-MSC treatment improved lipid metabolism and insulin resistance while reducing oxidative stress, consistent with concomitant suppression of inflammatory signaling [67]. The relevance of mitochondrial quality control is further supported by preconditioning strategies: mitochondria-preconditioned hUC-MSCs produced greater improvements in glycemic control, liver enzymes, triglycerides, and histological injury than unconditioned MSCs, in association with autophagy-facilitated mitochondrial transfer from MSCs to hepatocytes [49]. Likewise, mitochondrial transfer–mediated metabolic rescue by hBM-MSCs in NASH was linked to reduced lipid accumulation and improvements in lipid metabolism, inflammation, and fibrosis [39]. Together, these findings suggest that improving redox balance and mitochondrial function may contribute to the therapeutic effects of MSCs in MAFLD/NASH models by dampening metabolic stress and inflammatory injury.

Hepatoprotection and functional recovery

At the tissue level, MSC-based interventions often converge on hepatoprotection and functional recovery, integrating cytoprotective stress adaptation with pro-regenerative signaling—manifested as improvements in liver injury indices and histological features of hepatic damage [45, 49, 67, 70]. Mechanistically, direct cytoprotection has been linked to restoration of intracellular homeostasis; for example, BM-MSCs enhanced sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) activity, restored Ca²⁺ handling, and attenuated ER stress in NAFLD rats, coinciding with reduced inflammation and hepatocyte pyroptosis [14]. Regeneration-associated actions have also been described: ADSCs promoted hepatocyte proliferation and reduced apoptosis via activation of the Notch signaling pathway and upregulation of Hes family bHLH transcription factor 1 (HES1), with parallel improvements in fibrosis and inflammation [43]. In addition, hUC-MSCs improved insulin resistance and glucose metabolism via HGF-mediated downregulation of hepatic ring finger protein 186 (RNF186), consistent with a trophic-factor contribution to functional recovery in metabolic liver injury [65]. Collectively, these studies suggest that MSC therapy not only protects hepatocytes but also activates repair and regenerative programs, supporting functional recovery beyond steatosis reduction alone; because persistent injury and inflammation drive fibrogenesis, such hepatoprotective effects may also help restrain downstream fibrotic remodeling.

Antifibrotic effects

Among their multiple therapeutic actions, also exhibit antifibrotic potential. Mechanistically, they may attenuate fibrogenesis by suppressing TGF-β/Smad signaling and reducing expression of key fibrogenic genes (e.g., Col1a1 and Col3a1), thereby inhibiting HSC activation and collagen deposition [43–45]. Notably, these antifibrotic readouts are frequently observed in parallel with improvements in upstream immunometabolic injury, suggesting that reduced inflammatory drive and hepatocellular stress contribute substantially to lowering fibrogenic output. For example, reduced fibrosis has been associated with immune modulation (e.g., suppression of CD4⁺ T-cell proliferation) and gut–liver axis protection, including preservation of intestinal barrier integrity with inhibition of TLR4 signaling [63, 64]. These changes are expected to attenuate the influx of pro-fibrogenic stimuli into the liver. Current data suggest that MSCs predominantly suppress HSC activation indirectly [33, 74, 77] via anti-inflammatory and metabolic improvements, rather than through direct targeting or reversal of advanced fibrotic lesions [12, 15, 16, 43–45]. This notion is further supported by multiple studies showing parallel improvements in histological fibrosis, fibrogenic gene expression, and markers of hepatic injury and inflammation [45, 70, 71], indicating that MSC-mediated fibrosis attenuation is largely secondary to upstream injury resolution rather than true scar regression. Furthermore, suppression of ER stress and restoration of autophagic flux—reported in NASH models—may further mitigate hepatocyte stress and apoptosis-associated profibrotic signaling, thereby limiting downstream HSC activation and extracellular matrix (ECM) remodeling [44]. Collectively, current data support the antifibrotic potential of MSC-based strategies in metabolic liver disease. However, the relative contribution of direct HSC/ECM-targeting actions versus secondary effects driven by immunometabolic remodeling likely varies by model and fibrosis stage, particularly in settings of established fibrosis [12, 15, 16].

Despite these promising findings, the current research still has several limitations. Most preclinical studies rely on conventional NAFLD/NASH rodent models (e.g., high-fat diet–fed or ob/ob mice), which inadequately recapitulate the metabolic complexity, phenotypic heterogeneity, and comorbidities characteristic of human metabolic dysfunction–associated fatty liver disease (MAFLD) [2, 5, 78]. Furthermore, although reductions in collagen accumulation and fibrotic biomarkers are frequently reported, direct evidence for a primary, lesion-resolving antifibrotic effect of MSCs—especially in the context of established fibrosis—is still lacking [33, 76, 77]. Whether MSCs can directly modulate HSCs at the cellular and molecular levels to reduce true fibrosis regression remains an open question requiring further mechanistic investigation.

Preclinical studies of MSC-derived cell-free therapies in MAFLD

Despite substantial preclinical evidence demonstrating that MSC-derived products—particularly MSC-derived extracellular vesicles (MSC-EVs)—exert potent anti-inflammatory, antioxidative, metabolic regulatory, and antifibrotic effects in animal models of MAFLD/NASH (Table 5), their clinical application in liver diseases remains in its infancy. To date, no clinical trials specifically investigating MSC-EVs for the treatment of MAFLD or NASH have been formally registered or initiated [33, 50–52].

Table 5.

Therapeutic mechanisms of MSC-derived extracellular vesicles in MAFLD/NASH

MSC source	Administration route	Dose and regimen	Disease model	Animal	Mechanisms	Major outcomes	Ref.
hUC-MSC–derived exosomes	Tail vein	100 µg, every 3 days for 6 weeks	NASH	Mouse	Enhancement of Nrf2 phosphorylation and upregulation of NQO-1 expression	Improved hepatic inflammation, lipid metabolism, and oxidative stress	[70]
Diabetic microenvironment–preconditioned hUC-MSC exosomes	Tail vein	100 µg, three times per week for 6 weeks	NAFLD	Mouse	Inhibition of the NLRP3/caspase-1/GSDMD pathway and upregulation of PRDX-1	Improved lipid metabolism, inflammation, fibrosis, and hepatocyte pyroptosis	[71]
hUC-MSC–derived exosomes	Tail vein	120 µg, once weekly for 15 weeks	NAFLD	Mouse	Hepatoprotective effects mediated by exosomal miR-24-3p targeting Keap-1 and activation of Nrf2 signaling	Improved liver function, reduced lipid deposition, ROS generation, and inflammation	[76]
hUC-MSC–derived exosomes	Tail vein	20 mg/kg, twice weekly for 6 weeks	NASH	Mouse	Modulation of anti-inflammatory macrophage polarization and hepatocellular PPARα signaling	Reduced body weight, liver injury, and hepatic inflammation	[77]
hUC-MSC–derived exosomes	Tail vein	100 µg, once weekly for 2 months	NAFLD	Rat	Exosomal miR-627-5p suppresses FTO expression, improving glucose and lipid metabolism	Improved glucose and lipid metabolism, liver injury, and hepatic lipid accumulation	[78]
BM-MSC–derived exosomes	Tail vein	15, 30, or 120 µg/kg, twice weekly for 6 weeks	NASH	Rat	Upregulation of miR-96-5p and suppression of caspase-2	Improved lipid metabolism, reduced apoptosis, and enhanced mitophagy	[79]
hUC-MSC–derived exosomes	Tail vein	100 µg/250 µL, single injection (1 h after LPS/D-GalN)	Acute liver failure (ALF)	Mouse	Suppression of NLRP3 inflammasome activation and downstream cytokines (IL-1β, IL-6) in LPS-stimulated macrophages	Reduced liver injury, ALT/AST levels, and inflammatory signaling	[80]
Curcumin-preconditioned hUC-MSC exosomes	Tail vein	15 µg/kg, single dose	NASH	Mouse	Regulation of inflammation- and oxidative stress–related gene expression	Improved obesity, liver function, lipid deposition, fibrosis, inflammation, and oxidative stress	[81]
hUC-MSC–derived EVs	Tail vein	2 mg/kg, single dose	T2DM/NAFLD	Mouse/Rat	miR-31-5p–mediated inhibition of PDGFB, with systemic vascular and neurovascular protective effects	Improved hepatic inflammation, fibrosis, steatosis, and neurovascular function	[82]
hUC-MSC–derived EVs	Tail vein	10 mg/kg, intravenous administration (outcomes assessed over 4 weeks)	NAFLD	Mouse	Efficient hepatocyte uptake of sEV-RNF31, upregulation of RNF31, promotion of mitophagy, and reduction of lipid deposition and apoptosis	Reduced body weight, ALT, AST, TG, and TC levels	[83]
hUC-MSC–derived exosomes	Tail vein	100 µg, every 3 days for 6 weeks	NASH	Mouse	Activation of the Nrf2/NQO-1 antioxidant pathway and attenuation of oxidative stress and inflammation	Improved ALT, TG, body weight, NAS score, liver TG, TNF-α, and IL-6 levels	[84]
hUC-MSC–derived EVs	Intraperitoneal	50 µg, three times per week for 4 weeks	NASH	Mouse	Regulation of AMPK signaling, maintenance of metabolic homeostasis, and inhibition of adipose lipolysis and hepatic steatosis	Reduced ALT, AST, and TG levels	[85]
BM-MSC–derived EVs	Tail vein	120 µg/kg, once every 2 days for 6 weeks	NASH	Rat	Modulation of mitochondrial dysfunction, hypoxia, inflammation, and angiogenesis	Cardioprotective effects and improved mitochondrial function	[57]
ADSC-derived exosomes	Tail vein	100 µg, once weekly for 3 weeks	NASH	Mouse	Suppression of endoplasmic reticulum stress and restoration of autophagic flux	Improved ER stress, apoptosis, and autophagy	[37]
ADSC-derived EVs	Tail vein	100 µg, twice weekly for 7 weeks	NAFLD	Mouse	Delivery of miR-223-3p to suppress E2F1 expression	Improved ALT, AST, hepatic lipid accumulation, and fibrosis	[86]
hUC-MSC–derived exosomes	Tail vein	10 mg/kg, once weekly for 4 weeks	NAFLD	Mouse	Activation of AMPK signaling, inhibition of SREBP-1c–mediated lipogenesis, and promotion of PPARα-driven fatty acid oxidation	Improved lipid metabolism, hepatic steatosis, and liver function	[87]
hUC-MSC–derived EVs	Tail vein	2 mg/kg every 5 days for 1 month	NASH	Mouse	Regulation of the AMPK/mTOR pathway and upregulation of EI24 to enhance lipophagy and autophagy	Reduced body weight, lipid dysregulation, steatosis, and inflammation	[88]
BM-MSC–derived exosomes	Intraperitoneal	50 µg, once weekly for 8 weeks	NAFLD	Rat	Regulation of PPAR signaling and genes involved in lipogenesis and fatty acid oxidation	Improved body weight, serum lipids, liver function, steatosis, antioxidant capacity, and inflammation	[76]
Human ADSC-derived EVs	Tail vein	1.0, 2.5, or 5.0 µg, twice weekly for 4 weeks	Mc4r-KO NASH	Mouse	Not specified	Improved ALT levels, hepatic inflammation, and fibrosis	[61]

EVs, extracellular vesicles; sEVs, small extracellular vesicles; MSC-EVs, mesenchymal stromal cell–derived extracellular vesicles; miRNA, microRNA; NLRP3, NOD-, LRR- and pyrin domain-containing protein 3; AMPK, AMP-activated protein kinase; mTOR, mechanistic target of rapamycin; TGF-β, transforming growth factor beta; Keap-1, Kelch-like ECH-associated protein 1; Nrf2, nuclear factor erythroid 2–related factor 2; PDGFB, platelet-derived growth factor subunit B. Other molecular targets follow standard gene/protein nomenclature

Accumulating evidence from preclinical studies has established that the therapeutic effects of MSCs are primarily mediated through paracrine signaling, with small extracellular vesicles (sEVs), including exosomes, recognized as key functional effectors [72]. MSC-EVs are membrane-bound vesicles released by MSCs and include small extracellular vesicles (sEVs), such as exosomes, which typically range from 50 to 200 nm in diameter [72]. These vesicles carry diverse bioactive cargos, including microRNAs, messenger RNAs, proteins, lipids, and other signaling molecules [73]. These vesicles can be internalized by hepatocytes, immune cells, and other metabolically relevant cell types via endocytosis or receptor-mediated uptake, thereby triggering multilayered cytoprotective and immunomodulatory responses. Increasing evidence suggests that MSC-EVs reproduce many of the therapeutic effects of parental MSCs while offering potential advantages in safety, stability, storage, and manufacturing consistency [50–52]. In both mouse and rat models of MAFLD/NASH, MSC-EVs, regardless of cellular sources or production method, have consistently demonstrated robust hepatoprotective effects, largely through the modulation of multiple disease-relevant signaling pathways. Importantly, MSC-EVs derived not only from classical sources (e.g., bone marrow, adipose tissues, umbilical cord) but also from non-conventional tissues such as dental pulp and menstrual blood exhibit comparable therapeutic efficacy [42]. This functional convergence across diverse tissue origins underscores the conserved nature of MSCs paracrine mechanisms and supports the broader applicability of MSC-EV-based strategies for MAFLD [50–52]. The detailed therapeutic mechanisms of MSC-EVs for MAFLD are shown in Fig. 2.

Fig. 2.

Mechanistic actions of MSC-derived EVs in MAFLD. EVs derived from bone marrow MSCs, adipose-derived stem cells, and umbilical cord MSCs exert hepatoprotective effects through coordinated anti-inflammatory, metabolic, and antifibrotic actions. MSC-EVs suppress hepatic inflammation by inhibiting inflammasome activation, reducing pro-inflammatory cytokines, and promoting macrophage polarization toward an M2 phenotype. In parallel, MSC-EVs improve metabolic homeostasis by activating antioxidant and AMPK-associated pathways, inhibiting lipogenesis, and enhancing fatty acid oxidation. Moreover, MSC-EVs attenuate fibrogenesis by downregulating TGF-β signaling and limiting hepatic stellate cell activation and extracellular matrix deposition. Figure designed by Adobe Illustrator CC 2020

Anti-inflammatory and immunomodulatory actions

Regarding anti-inflammatory and immunomodulatory effects, MSC-EVs have been shown to suppress the activation of the NLRP3 inflammasome, downregulate pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-α, and promote macrophage polarization toward an anti-inflammatory M2 phenotype. Collectively, these actions lead to a significant attenuation of hepatic inflammation. For instance, EVs derived from human umbilical cord MSCs (hUC-MSC-EVs) preconditioned in a diabetic microenvironment significantly reduced hepatocyte pyroptosis by inhibiting the NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3), caspase-1, and gasdermin D (GSDMD) axis [80]. Bone marrow MSC–derived EVs (BM-MSC-EVs) have been reported to suppress caspase-2 expression via delivery of miR-96-5p, thereby reducing hepatocyte apoptosis and enhancing mitophagy [84]. In addition, hUC-MSC-EVs can activate the nuclear factor erythroid 2-related factor 2/NAD(P)H quinone dehydrogenase 1 (Nrf2/NQO-1) antioxidant pathway, resulting in reduced oxidative stress and mitigation of lipotoxic injury [76, 79, 81].

Metabolic regulation

In terms of metabolic regulation, MSC-EVs exert beneficial effects on hepatic lipid accumulation and insulin resistance through multiple signaling pathways. For example, hUC-MSC-EVs have been shown to exert antioxidative and hepatoprotective effects by targeting Kelch-like ECH-associated protein 1 (Keap-1) [81], while exosomal miR-627-5p alleviates hepatic injury and improves glucose and lipid metabolism through suppression of fat mass and obesity-associated protein (FTO) expression [83]. Other studies have demonstrated that hUC-MSC-EVs upregulate ring finger protein 31 (RNF31), thereby promoting mitophagy and reducing lipid deposition [88]. Moreover, MSC-EVs can activate key metabolic signaling cascades such as AMPK/mTOR and AMPK/PPARα, suppressing SREBP-1c–mediated lipogenesis while enhancing fatty acid oxidation [89–91]. Collectively, these effects translate into improvements in serum ALT and AST levels, hepatic triglyceride and cholesterol content, and body weight [81–83, 88–93].

Antifibrotic potential

Notably, MSC-EVs also exhibit promising antifibrotic potential. Experimental studies have shown that MSC-EVs can downregulate TGF-β–related signaling pathways and inhibit expression of alpha-smooth muscle actin (α-SMA) and collagen, thereby attenuating hepatic fibrogenesis [82, 86, 93]. Adipose-derived MSC EVs (ADSC-EVs) have been reported to alleviate liver fibrosis through delivery of miR-223-3p, which suppresses E2F transcription factor 1 (E2F1) signaling [93]. In addition, EVs derived from MSCs preconditioned with curcumin further ameliorate fibrotic progression by modulating genes involved in inflammation and oxidative stress [86].

Translational advantages and challenges

Compared with live MSCs, MSC-EVs offer several translational advantages, including lower immunogenicity, absence of tumorigenic risk, greater stability, and improved feasibility for storage, transport, and large-scale production [73], making them an attractive cell-free therapy for future clinical applications [50–52]. Furthermore, engineering strategies that modify EV cargo—such as enrichment of specific miRNAs or loading of therapeutic agents—may further enhance targeting efficiency and functional specificity [51, 52]. However, the clinical translation of MSC-EVs remains challenging. The biological activity of MSC-EVs is strongly influenced by cell source, culture conditions, preconditioning strategies, and purification methods [94], and the heterogeneity observed across studies highlights ongoing challenges in standardization and manufacturing. Moreover, the precise molecular mechanisms underlying MSC-EV-mediated actions remain incompletely understood, and key translational parameters, including pharmacokinetics, biodistribution, target-organ retention, and long-term safety, have not yet been adequately characterized [51, 52, 61]. Although MSC-EVs reproduce many of the anti-inflammatory, metabolic, and antifibrotic effects of parental MSCs in preclinical models, it remains uncertain whether they can fully recapitulate the dynamic and context-dependent functions of live cells in complex human disease. Therefore, future progress will require standardized manufacturing workflows, rigorous functional characterization, improved targeting strategies, and well-designed translational studies to define the appropriate clinical setting, dosing regimen, and therapeutic endpoints for MSC-EV-based interventions in MAFLD.

Summary and perspectives

MSCs and their derivatives have emerged as promising therapeutic candidates for MAFLD due to their ability to modulate inflammation, oxidative stress, fibrogenesis, and metabolic homeostasis through coordinated, multitarget mechanisms. Accumulating preclinical evidence indicates that the therapeutic effects of MSCs are predominantly mediated through paracrine mechanisms, with MSC-derived extracellular vesicles emerging as principal mediators and a promising cell-free therapeutic modality [31–33, 50–52]. Compared with live MSCs, MSC-EVs exhibit several translational advantages, including lower immunogenicity, improved safety, and greater feasibility for standardization and large-scale production [50–52, 61, 94].

However, the clinical translation of MSC-based therapies remains constrained by limited high-quality clinical evidence [36, 95], insufficient disease-relevant models that fully capture the heterogeneity of human MAFLD [78], and substantial variability in cell and EV preparation protocols [36, 61]. MAFLD is increasingly viewed as a multisystem disorder shaped by metabolic crosstalk along the gut–adipose–liver axis, in which chronic nutrient excess is coupled to hepatic injury through mitochondrial dysfunction and enhanced de novo lipogenesis [4]. In contrast, commonly used rodent models and liver-centric readouts may underrepresent comorbidity burden and inter-organ signaling, and can overlook major extrahepatic determinants—such as adipose tissue insulin resistance—that influence fibrosis risk in metabolic populations [4].

Progress will require human-relevant models that reflect patient heterogeneity and inter-organ signaling, alongside standardized EVs manufacturing and characterization with clear dosing metrics and release criteria [36, 61]. Engineering strategies such as metabolic preconditioning or surface functionalization may further improve hepatic targeting, retention, and cargo delivery within the inflamed, steatotic liver microenvironment [4, 96]. Early-phase trials should not rely on histological staging alone, but complement it with biomarkers that track dynamic changes in metabolism, immune activity, and fibrogenic activity over time [4]. Such designs will help identify meaningful responders and define where MSC-based therapeutics can be developed as controllable, precision therapies for MAFLD.

Abbreviations

ACC

Acetyl-CoA carboxylase

ACOT

Acyl-CoA thioesterase

ADRCs

Adipose-derived regenerative cells

ADSCs

Adipose-derived stem cells

ADSC-EVs

Adipose-derived stem cell extracellular vesicles

ALF

Acute liver failure

ALP

Alkaline phosphatase

ALT

Alanine aminotransferase

AMPK

AMP-activated protein kinase

AST

Aspartate aminotransferase

BAs

Bile acids

BM-MSCs

Bone marrow–derived mesenchymal stem cells

BM-MSC-EVs

Bone marrow MSC–derived extracellular vesicles

CD

Cluster of differentiation

CES2

Carboxylesterase 2

CHOP

C/EBP homologous protein

COX-2

Cyclooxygenase-2

CPT-1

Carnitine palmitoyltransferase 1

CSF-1

Colony-stimulating factor 1

CYP4A

Cytochrome P450 4 A

DAMPs

Damage-associated molecular patterns

D-GalN

D-galactosamine

E2F1

E2F transcription factor 1

ECM

Extracellular matrix

EI24

Etoposide-induced 2.4

eIF2α

Eukaryotic initiation factor 2 alpha

ER

Endoplasmic reticulum

EVs

Extracellular vesicles

FASN

Fatty acid synthase

FIB-4

Fibrosis-4 index

FFAs

Free fatty acids

FTO

Fat mass and obesity-associated protein

FXR

Farnesoid X receptor

GGT

Gamma-glutamyl transferase

GIP

Glucose-dependent insulinotropic polypeptide

GLP-1

Glucagon-like peptide-1

GSDMD

Gasdermin D

HBV

Hepatitis B virus

HCV

Hepatitis C virus

HDL-C

High-density lipoprotein cholesterol

HES1

Hes family bHLH transcription factor 1

HGF

Hepatocyte growth factor

HNF4α

Hepatocyte nuclear factor 4 alpha

hUC-MSCs

Human umbilical cord–derived mesenchymal stem cells

hUC-MSC-EVs

Human umbilical cord MSC–derived extracellular vesicles

ICAM-1

Intercellular adhesion molecule 1

IFN-γ

Interferon gamma

IL

Interleukin

INR

International normalized ratio

IR

Insulin resistance

ISCT

International society for cell and gene therapy

JNK

c-Jun N-terminal kinase

Keap-1

Kelch-like ECH-associated protein 1

KO

Knockout

LDL-C

Low-density lipoprotein cholesterol

LOXL2

Lysyl oxidase-like 2

LPS

Lipopolysaccharide

MAFLD

Metabolic dysfunction–associated fatty liver disease

MAPK

Mitogen-activated protein kinase

MASLD

Metabolic dysfunction–associated steatotic liver disease

MASH

Metabolic dysfunction–associated steatohepatitis

Mc4r

Melanocortin 4 receptor

Mc4r-KO

Mc4r knockout

MCD

Methionine- and choline-deficient

MELD

Model for End-Stage Liver Disease

MELD-Na

Model for End-Stage Liver Disease–Sodium

miRNA

microRNA

mRNA

messenger RNA

MSCs

Mesenchymal stem/stromal cells

MSC-CM

Mesenchymal stem cell–conditioned medium

MSC-EVs

Mesenchymal stem cell–derived extracellular vesicles

MSC/HSC

Mesenchymal stem cell/hematopoietic stem cell

mTOR

Mechanistic target of rapamycin

NAFLD

Nonalcoholic fatty liver disease

NAS

NAFLD activity score

NASH

Nonalcoholic steatohepatitis

NF-κB

Nuclear factor kappa B

NK

Natural killer

NLRP3

NOD-, LRR- and pyrin domain-containing protein 3

NQO-1

NAD(P)H quinone dehydrogenase 1

Nrf2

Nuclear factor erythroid 2–related factor 2

OCA

Obeticholic acid

PDGFB

Platelet-derived growth factor subunit B

PERK

Protein kinase RNA-like endoplasmic reticulum kinase

PPAR

Peroxisome proliferator-activated receptor

PPARα

Peroxisome proliferator-activated receptor alpha

PPARγ

Peroxisome proliferator-activated receptor gamma

PRDX-1

Peroxiredoxin 1

PV/IV

Portal vein/Intravenous

RCTs

Randomized controlled trials

RNF31

Ring finger protein 31

RNF186

Ring finger protein 186

ROS

Reactive oxygen species

SCFAs

Short-chain fatty acids

SERCA

Sarco/endoplasmic reticulum Ca²⁺-ATPase

sEVs

Small extracellular vesicles

SGLT2

Sodium-glucose cotransporter 2

SHED

Stem cells from human exfoliated deciduous teeth

SHED-CM

SHED-MSC–conditioned medium

SIRT1

Sirtuin 1

Smad

Smad proteins

SREBP

Sterol regulatory element-binding protein

SREBP-1c

Sterol regulatory element-binding protein 1c

STING

Stimulator of interferon genes

T2DM

Type 2 diabetes mellitus

TBIL

Total bilirubin

TC

Total cholesterol

TFEB

Transcription factor EB

TG

Triglycerides

TGF-β

Transforming growth factor beta

THR-β

Thyroid hormone receptor beta

TLR4

Toll-like receptor 4

TNF-α

Tumor necrosis factor alpha

Treg

Regulatory T cell

UC-MSCs

Umbilical cord–derived mesenchymal stem cells

ZO-1

Zonula occludens-1

α-SMA

Alpha-smooth muscle actin

Author contributions

ZH and YP searched literatures, collected data. ZH and NL and wrote the initial draft of the manuscript. NL and FP designed tables and revised the manuscript. All the authors reviewed and approved the final manuscript.

Funding

This study was supported by the Natural Science Foundation of Fujian (grant no. 2024J011219, 2024J011163), Fujian Province Joint Funds for the Innovation of Science and Technology (grant no. 2025Y9481), Fujian Young Healthcare Talent Development Program (grant no. 2025GGA083). Fuzhou Sci-Tech Talent Cultivation Project (grant no. 2025R012).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.