Role of Extracellular Vesicles in Skin Barrier Repair: Applications in Atopic Dermatitis and Chronic Wounds

Introduction

The skin barrier maintains homeostasis through structural and biochemical components: lipid lamellae (ceramides, cholesterol, fatty acids), corneocyte envelope assembly, toll-like receptor-mediated antimicrobial peptide production (β-defensins, cathelicidins, lipocalin-2), and commensal microbiota regulation.^1–8 Together, the stratum corneum, viable epidermis, and dermal–epidermal junction forms a physicochemical and immunological shield maintained by continuous keratinocyte differentiation and desquamation.^4,9

Barrier disruption underlies both atopic dermatitis (AD) and chronic wounds.¹⁰ In AD, filaggrin (FLG) loss-of-function mutations, altered lipid composition, and type-2 immune dysregulation weakens epidermal integrity and sustain chronic inflammation via allergen penetration.^11–14 Chronic wounds, including diabetic foot ulcers (DFUs), venous leg ulcers, and pressure ulcers, arise from failure to progress through the normal healing cascade, with persistent inflammation, protease imbalance, cellular senescence, and impaired angiogenesis collectively preventing re-epithelialization.¹⁵

Despite distinct initiating factors, both conditions converge on unresolved inflammation, oxidative stress, ECM disruption, and vascular dysregulation.^16–21 Mechanistically, however, they diverge where AD inflammation is allergen-driven with depleted pro-resolving mediators, while chronic wounds sustain inflammation through biofilm and hypoxia.¹⁶ Angiogenic responses are likewise contrasting, dysregulated and often excessive in AD but impaired and insufficient in chronic wounds, highlighting distinct mechanistic pathways that ultimately converge on a common outcome of barrier dysfunction.^22–24

Current treatments offer symptomatic relief without restoring full barrier function. Emollients, topical corticosteroids, and wound dressings remain standard, but prolonged corticosteroid use carries risks including epidermal atrophy, striae, and systemic complications,^25–28 while dressings lack intrinsic regenerative bioactivity.²⁹

Extracellular vesicles (EVs; 30–1000 nm) have emerged as promising cell-free therapeutics. They transport bioactive cargo including messenger RNAs (mRNAs), microRNAs (miRNAs), proteins, lipids, and metabolites, protecting miRNAs from degradation and enabling targeted gene regulation.^30,31 Compared to cell-based therapies, EVs offer reduced immunogenicity, no replicative capacity, lower tumorigenicity risk, and versatile delivery routes.^32–37 Preclinical studies show MSC-, keratinocyte-, and fibroblast-derived EVs promote inflammation suppression, cell migration and proliferation, angiogenesis, and ECM remodelling.^38–40

Several recent reviews have examined EVs in dermatological diseases, each addressing a distinct scope. Yuan et al surveyed the biological functions and therapeutic potential of EVs broadly across skin conditions, covering wound healing, inflammatory diseases, and skin aging, but without disease-specific mechanistic depth for AD or chronic wounds individually.⁴¹ Eerdekens et al focused specifically on EV-based strategies for wound repair, reviewing cell-free EV approaches across diabetic ulcers and other wound types, but did not address AD or compare delivery translation between preclinical and clinical settings.³¹ Wang et al similarly examined EVs across skin health, disease, and aging with broad scope, without comparative mechanistic analysis between specific conditions.⁴² Fang et al and Lei et al each reviewed EVs in skin health and inflammatory skin diseases respectively, providing useful overviews of EV biology but again without directly contrasting EV mechanisms across AD and chronic wounds or identifying the delivery translational gap as a finding.^43,44 This review identifies a translational gap between preclinical delivery innovations and current clinical trial design.

This review systematically evaluates EV-based therapeutic strategies for skin barrier repair in both AD and chronic wounds, comparing their distinct mechanisms of barrier disruption and summarizing emerging delivery approaches including topical formulations, hydrogels, microneedles, and scaffold-based systems, and proposes a phased clinical roadmap, providing actionable direction for next-generation trial design.

Pathophysiology of Skin Barrier Dysfunction in Atopic Dermatitis and Chronic Wounds

Atopic Dermatitis

Atopic dermatitis (AD), also known as atopic eczema, is a chronic, inflammatory skin disorder that weakens the skin’s protective barrier, resulting in the skin functioning less effectively.^45,46 According to the World Health Organization (WHO), approximately 220 million people worldwide are affected by this disease, which has become a burden that impairs the quality of life for affected individuals and families.⁴⁵ The clinical manifestations experienced by most individuals include pruritus, erythema, and sleep disturbances.⁴⁵

The pathophysiology of AD involves complex interactions among genetic, immunological, and environmental factors. Genetic predisposition plays a vital role in AD, specifically mutations in the FLG gene, as these defects may initiate the primary barrier dysfunction, which is subsequently worsened by internal or external stressors.^45,47 Mutations in the FLG gene reduce the synthesis of FLG, a crucial structural protein that maintains epidermal integrity, leading to increased transepidermal water loss (TEWL), reduced lipid composition, and increased skin dryness.⁴⁵ However, AD is a multifactorial disease in which immune dysregulation also contributes significantly to barrier dysfunction. In acute lesions, a defective skin barrier, whether arising from genetic factors or external damage, stimulates keratinocytes to produce pro-inflammatory cytokines, which activate antigen-presenting cells (APCs) and subsequently T-helper (Th) cells. Activated Th2 cells release cytokines including interleukin (IL)-4, IL-5, IL-13, and IL-31, which further impair the skin barrier, promote inflammation, and cause pruritus.^45,46,48 In the chronic phase, Th1, Th2, and Th22 cells collectively drive abnormal keratinocyte proliferation and epidermal thickening.⁴⁵

In addition to genetic and immunological factors, alterations in the skin microbiome play a vital role in AD pathophysiology. Studies have shown that Staphylococcus aureus colonises 60–100% of AD patients, compared to only 5–30% of healthy individuals.⁴⁸ S. aureus can damage and penetrate the skin barrier; its enterotoxins trigger decreased FLG levels and increased pro-inflammatory cytokines, ultimately resulting in Th2 overexpression, pruritus, and cutaneous inflammation, which further promotes S. aureus colonisation and perpetuates this cycle.⁴⁸

Chronic Wounds

Chronic wounds including DFUs, venous leg ulcers, and pressure ulcers, fail to progress through normal healing phases due to persistent inflammation, impaired cellular function, extracellular matrix (ECM) remodelling dysfunction, and deficient angiogenesis.⁴⁹ These interconnected pathological features create a hostile microenvironment characterised by elevated pro-inflammatory cytokines (TNF-α, IL-6, IL-1β), excessive matrix metalloproteinase (MMP) activity, oxidative stress, and cellular senescence, collectively preventing effective tissue repair.^50–55

At the cellular level, fibroblast and keratinocyte dysfunction are driven by senescence-associated secretory phenotype (SASP), disrupted growth factor signalling, and activation of NF-κB, p53, and p16INK4a pathways, which impairs re-epithelialisation and ECM synthesis.^56–61 MMP-TIMP imbalance further degrades newly synthesised collagen, with reduced COL1A1 and COL1A2 expression compounding structural repair deficits.⁶² Deficient angiogenesis is driven by VEGF degradation, accumulation of anti-angiogenic factors including thrombospondin-1 (TSP-1), endostatin, and pigment epithelium-derived factor (PEDF), and HIF-1α dysfunction, resulting in sustained hypoxia that further perpetuates inflammation, bacterial growth, and cellular senescence.^63–73

These converging pathological mechanisms highlight the need for a multi-targeted therapeutic approach, providing the rationale for extracellular vesicle (EV)-based interventions capable of simultaneously addressing immune dysregulation, restoring cellular function, rebalancing ECM remodelling, and promoting angiogenesis.

Mechanistic Differences and Shared Outcomes

Despite their distinct aetiologies, AD and chronic wounds share a converging pathological endpoint: disruption of the skin barrier leading to impaired homeostasis, susceptibility to infection, and diminished quality of life. In AD, barrier failure is primarily inside-out, initiated by genetic defects, notably FLG mutations, amplified by Th2-dominant immune dysregulation, and sustained by dysbiotic microbiota-driven inflammation.^45–48 Conversely, in chronic wounds, barrier destruction proceeds outside-in, driven by persistent inflammatory signalling, cellular senescence, MMP-mediated ECM degradation, and deficient neovascularisation.^49–73 However, both conditions are ultimately characterised by pro-inflammatory cytokine excess, defective keratinocyte and fibroblast function, disrupted lipid and structural protein composition of the epidermis, and impaired barrier integrity, representing a shared molecular landscape that positions EV-based therapy as a unifying, multi-targeted therapeutic strategy. Table 1 provides a structured comparison of the key pathophysiological features of AD and chronic wounds, highlighting their mechanistic differences alongside their common outcomes, and illustrating how each feature informs the rationale for EV-based intervention.

Table 1 Comparative Pathophysiology of AD and Chronic Wounds: Mechanistic Differences and Shared Endpoints Relevant to EV Therapy

Role of Extracellular Vesicles in Skin Barrier Repair

Definition and Types of Extracellular Vesicles

Extracellular vesicles (EVs) are lipid bilayer-delimited particles released by all cell types, facilitating intercellular communication through the transfer of bioactive molecules.¹¹⁰ These naturally released cell-derived vesicles contain proteins, lipids, and nucleic acids that cannot replicate, distinguishing them from cellular organisms.¹¹¹ EVs are classified primarily into two subtypes based on their biogenesis mechanism and size range.¹¹²

Exosomes represent the smaller subset of EVs, typically ranging from 30 to 150 nm in diameter.^111,113 These vesicles originate from the endosomal pathway, specifically through the formation of multivesicular bodies (MVBs) and subsequent fusion with the plasma membrane.¹¹⁴ Exosomes are characterized by specific markers including tetraspanins (CD9, CD63, CD81), TSG101, and Alix, which are involved in their biogenesis through the endosomal sorting complex required for transport (ESCRT)-dependent and independent pathways.^115,116 The ESCRT-independent pathway involves sphingomyelinase hydrolysis and ceramide formation, which stimulates the spontaneous negative curvature of the MVB membrane to form intraluminal vesicles.^117,118

Microvesicles, also known as ectosomes or microparticles are larger EVs with diameters ranging from 100 to 1000 nm.^111,119 Unlike exosomes, microvesicles are formed through direct outward budding of the plasma membrane.¹²⁰ Both EV subtypes carry complex cargo including miRNAs, mRNAs, proteins, and lipids, such as ceramide, sphingomyelin, phosphatidylserine, and cholesterol, as well as growth factors such as vascular endothelial growth factor (VEGF), transforming growth factor-beta (TGF-β), and epidermal growth factor (EGF).^121–123 This diverse molecular cargo enables EVs to modulate multiple cellular processes crucial for tissue repair and regeneration.⁷⁴ A schematic overview of EV structure, including membrane composition, surface markers, and bioactive cargo, is presented in Figure 1.

Figure 1 Structural and functional overview of EVs. This schematic illustrates the structural organisation of EVs, including the phospholipid bilayer and encapsulated bioactive cargo such as proteins, lipids, and nucleic acids involved in intercellular communication and regulation of target cell function.

Cellular Sources of Therapeutic EVs

Mesenchymal stromal cell (MSC)-derived EVs have emerged as one of the most extensively studied therapeutic EV sources for skin barrier repair. MSCs can be isolated from various tissues including bone marrow, adipose tissue, umbilical cord blood, and Wharton’s jelly.^124,125 These cells exhibit anti-inflammatory, immunomodulatory, and pro-regenerative properties that are largely mediated through their paracrine secretion of EVs rather than direct differentiation into skin cells.^126,127 MSC-EVs offer significant advantages over parent cell therapy, including lower risk of tumorigenicity and immunogenicity, greater amenability to cryopreservation with minimal loss of bioactivity, and improved quality control for storage and administration.^34,128

MSC-EVs contain abundant growth factors and cytokines, including TGF-β, EGF, VEGF, and hepatocyte growth factor (HGF), which promote cell proliferation, migration, and angiogenesis.^89,90 Additionally, MSC-EVs carry specific miRNAs such as miR-21-3p, miR-125a, and miR-126-3p that regulate wound healing pathways.¹²⁹ Preconditioning strategies, particularly priming MSCs with interferon-gamma (IFN-γ), have been shown to enhance the immunomodulatory capabilities of MSC-derived EVs by upregulating indoleamine 2,3-dioxygenase (IDO) and other immunosuppressive factors.^130,131

While less extensively studied than MSC-EVs, epidermal stem cell-derived EVs show promise in promoting keratinocyte proliferation and differentiation.¹³² Keratinocyte-derived EVs play important roles in skin homeostasis by regulating immune responses and wound healing processes.¹³³ These EVs can modulate intercellular communication within the epidermis and influence the behaviour of fibroblasts and immune cells.¹³⁴

Commensal skin microbiota, particularly Staphylococcus epidermidis, produces EVs that can modulate immune responses and restore microbial balance.⁸⁸ S. epidermidis-derived EVs have been shown to exert significant protective effects by markedly reducing the expression of key pro-inflammatory mediators, including VEGF-A, IL-6, keratinocyte chemoattractant (KC), IL-23, IL-17F, IL-36γ, and IL-36 receptor (IL-36R), while simultaneously upregulating the IL-36 receptor antagonist (IL-36Ra).⁸⁸ Notably, the study also reported that different strains of S. epidermidis exhibit varying capacities to modulate skin inflammation. In contrast, Hong et al demonstrated that treatment of dermal fibroblasts with Staphylococcus aureus–derived EVs significantly increased the production of IL-6, thymic stromal lymphopoietin (TSLP), microphage inflammatory protein (MIP)-1α, and eotaxin.¹³⁵ Furthermore, topical application of S. aureus-derived EVs to tape-stripped murine skin induced epidermal thickening, dermal infiltration of mast cells and eosinophils, and elevated levels of IL-4, IL-5, IFN-γ, and IL-17, thereby recapitulating key features of AD-like inflammation.¹³⁵

Functional Mechanisms of EV-Mediated Skin Barrier Repair

The therapeutic effects of EVs on skin barrier repair can be organised into three interconnected functional domains, including immune modulation, epidermal barrier restoration, and dermal remodelling with angiogenesis. Although these mechanisms operate concurrently, their relative contributions differ between AD and chronic wounds, reflecting the distinct pathophysiological priorities of each condition. The context-dependent roles of EVs in AD and chronic wounds are summarised in Figure 2. The key signalling pathways mediating each functional domain are illustrated in Figure 3.

Figure 2 Comparative framework of EV mechanisms in AD and chronic wounds. This figure presents a comparative framework of EV-mediated mechanisms across these conditions, highlighting key functional domains including immune modulation, epidermal barrier restoration, and dermal remodelling and angiogenesis. The relative contribution of each mechanism differs between disease contexts, with immune modulation predominating in AD and angiogenesis and tissue regeneration playing a central role in chronic wound healing.

Figure 3 Signalling pathways underlying EV-mediated functional effects. This figure illustrates key signalling pathways involved in EV-mediated skin barrier repair, including AKT/ERK, NF-κB, and HIF-1α/VEGF pathways. Abbreviations: EV, extracellular vesicle; MSC, mesenchymal stromal cell; CerS3, ceramide synthase 3; Sphk1, sphingosine kinase 1; EGF, epidermal growth factor; FGF, fibroblast growth factor; TGF, transforming growth factor; PDGF, platelet-derived growth factor; eNOS, endothelial nitric oxide synthase; TSLP, thymic stromal lymphopoietin; TNF, tumour necrosis factor; TEWL, transepidermal water loss.

Immune Modulation

EVs derived from mesenchymal stromal cells (MSC-EVs) exert potent immunomodulatory effects through multiple, coordinated signalling mechanisms. At the level of adaptive immunity, MSC-EVs regulate T-helper cell differentiation, specifically inhibiting pro-inflammatory Th2 and Th17 responses while concurrently upregulating the anti-inflammatory mediators, including interleukin (IL)-10 and TGF-β, thereby restoring immune homeostasis.^82–85 At the innate immune level, MSC-EVs modulate macrophage polarisation from the pro-inflammatory M1 to the anti-inflammatory M2 phenotype, and suppress activation of mast cells and neutrophils.^81,86 Neural stem cell-derived EVs similarly reduce pro-inflammatory cytokine expression and inhibit NF-κB phosphorylation in keratinocytes and macrophages, demonstrating that immunomodulatory capacity is not exclusive to MSC-derived populations.⁸⁷

In AD, immune dysregulation is the primary driver of barrier failure, and EV-mediated immune modulation is correspondingly the central therapeutic mechanism. MSC-EVs have been shown to significantly reduce circulating levels of IL-4, IL-6, IL-13, IL-31, TNF-α, and TSLP, key cytokines that collectively sustain the Th2-skewed inflammatory milieu characteristic of AD.^80–82 Notably, interferon-gamma (IFN-γ)-primed MSC-EVs achieved therapeutic outcomes comparable to, or exceeding, those of the JAK inhibitor baricitinib and the corticosteroid clobetasol in AD mouse models, including reductions in skin thickness and mast cell infiltration, without the systemic side effects associated with corticosteroid use.^82,130

In chronic wounds, persistent inflammation impedes tissue repair rather than initiating barrier disruption, and immune modulation therefore functions in a supportive rather than primary role. MSC-EVs reduce levels of TNF-α and IL-6 in the wound microenvironment and promote macrophage polarisation toward the M2 phenotype, creating pro-regenerative conditions that enable subsequent cellular repair processes.^81,86 This shift from a pro-inflammatory to a reparative immune environment is a prerequisite for effective re-epithelialisation, ECM remodelling, and neovascularisation.

Epidermal Barrier Restoration

Restoration of the epidermal barrier requires the coordinated proliferation, migration, and functional differentiation of keratinocytes. MSC-EVs deliver a repertoire of growth factors, including EGF, VEGF, TGF-β, fibroblast growth factor-2 (FGF-2), and platelet-derived growth factor (PDGF) that activate proliferation-related genes (cyclin D1, cyclin D2, cyclin A1, and cyclin A2) through the protein kinase B (AKT) and extracellular signal-regulated kinase (ERK) signalling pathways.^75,91,92 The miRNA cargo of EVs further fine-tunes these cellular responses; miR-135a, for example, promotes keratinocyte and fibroblast migration by directly inhibiting large tumour suppressor kinase 2 (LATS2) expression.⁹³

A critical component of epidermal barrier competence is the lipid matrix of the stratum corneum, particularly its ceramide composition. Adipose tissue-derived MSC-EVs promote de novo ceramide synthesis through enhanced expression of serine palmitoyltransferase, HMG-CoA reductase, and ceramide synthase 3 (CerS3).^77–79 High-throughput RNA sequencing has confirmed that MSC-EVs restore gene expression programmes governing keratinocyte differentiation, lipid metabolism, and sphingolipid biosynthesis.⁷⁹ At the metabolic level, MSC-EVs elevate sphingosine and sphingosine-1-phosphate (S1P) concentrations by activating sphingosine kinase 1 (Sphk1) and simultaneously reducing S1P lyase (S1PL) activity, thereby shifting the lipid balance toward barrier-forming ceramide species.⁷⁹ MSC-EVs additionally promote formation of lamellar layers at the stratum granulosum–stratum corneum interface, further consolidating structural barrier integrity.^78,79

In AD, the epidermal barrier is intrinsically compromised due to FLG deficiency and impaired lipid synthesis. EV-based interventions targeting ceramide biosynthesis and keratinocyte differentiation therefore address the underlying structural deficit directly. Preclinical evidence in AD models demonstrates that MSC-EV treatment reduces TEWL and restores expression of barrier-associated proteins, confirming functional and not merely histological recovery of the skin barrier.^{79,82,130,136,137}

In chronic wounds, re-epithelialisation is impaired by cellular senescence and a hostile wound microenvironment, making keratinocyte activation the key mechanistic target. MSC-EVs accelerate wound closure by stimulating keratinocyte proliferation and migration through AKT/ERK signalling, while simultaneously restoring the epidermal lipid barrier in healed tissue.^75,138,139 Restoration of barrier-associated proteins including keratin-1 (KRT1) and aquaporin-3 (AQP3) following EV treatment further supports functional epidermal competence in the repaired wound.¹⁴⁰

Dermal Remodelling and Angiogenesis

Dermal remodelling is driven by fibroblast activation and collagen matrix reconstitution. MSC-EVs stimulate fibroblast proliferation, migration, and functional differentiation through delivery of TGF-β, PDGF, and FGF-2, and through activation of AKT and ERK pathways.^75,92 Downstream, MSC-EVs promote increased expression of collagen type I and III, fibronectin, and elastin.^75,76 The miRNA cargo of EVs contributes to ECM stabilisation, where miR-29a-3p enhances the angiogenic capacity of endothelial cells through the Wnt/β-catenin signalling pathway, and miR-486-5p mediates wound healing by promoting angiogenesis.^94,95

Neovascularisation is coordinated through several convergent signalling axes. MSC-EVs activate the HIF-1α/VEGF pathway to promote endothelial cell survival and sprouting under hypoxic conditions, the AKT/endothelial nitric oxide synthase (eNOS) pathway to enhance endothelial migration and tube formation, and the Wnt/β-catenin pathway to sustain proliferative angiogenic signalling.^95,101,102 Immunohistochemical evidence demonstrates significantly elevated tissue levels of VEGF, angiopoietin-1, angiopoietin-2, and the vascular marker CD31 following MSC-EV administration.^98,99 Specific miRNAs, including miR-612 and miR-126, further promote neovascularisation by regulating target genes involved in vessel formation.^103,104

In AD, dermal remodelling and angiogenesis play a secondary role relative to immune modulation and epidermal barrier restoration. The disease is not primarily characterised by deficient vasculature. However, MSC-EV-mediated support of fibroblast function and tissue homeostasis may contribute to overall dermal health and reduce the structural consequences of chronic inflammation, such as lichenification and epidermal thickening in long-standing disease.

In chronic wounds, impaired angiogenesis is a defining pathological feature, which sustained hypoxia resulting from deficient neovascularisation perpetuates inflammation, promotes bacterial colonisation, and prevents effective tissue repair. EV-based restoration of angiogenesis is therefore a primary therapeutic objective in this context. MSC-EVs have been shown to effectively restore impaired neovascularisation in diabetic wound models, with enhanced expression of VEGF, CD31, and the proliferation marker Ki67 confirming both vascular and cellular recovery.^98,105,106 Endothelial progenitor cell-derived EVs similarly enhance wound healing through activation of the advanced glycation end product–receptor for advanced glycation end products (AGE–RAGE) signalling pathway.⁹⁸

Preclinical and Clinical Evidence of EV Therapy for Skin Barrier Repair

While Functional Mechanisms of EV-Mediated Skin Barrier Repair organized evidence by functional mechanisms to explain how EVs mediate skin barrier repair, the following section surveys the same body of preclinical work by disease model and EV source, establishing the breadth and consistency of efficacy evidence that underpins the clinical translation discussion in Clinical Translation: Current State and Critical Gaps.

Preclinical Studies

Across these studies, EV-based interventions have consistently demonstrated the capacity to modulate all three functional domains outlined above, including immune regulation, epidermal barrier restoration, and dermal remodelling with angiogenesis, with the relative contribution of each domain varying predictably according to the underlying disease pathophysiology.

In 2,4-dinitrochlorobenzene (DNCB)-induced AD models, IFN-γ-primed induced pluripotent stem cell-derived MSC-EVs (IFN-γ-iMSC-EVs) significantly reduced dermatitis scores, skin thickness, and epidermal hyperplasia, with histological analysis confirming decreased infiltration of mast cells and inflammatory cells alongside suppression of Th2 immune responses evidenced by decreased IL-4, IL-5, IL-13, and IL-31 signalling.^82,130 Critically, therapeutic outcomes were comparable to or exceeded those of baricitinib and clobetasol without the severe weight loss observed with corticosteroid treatment.⁸² Adipose-derived MSC-EVs demonstrated similar efficacy across house dust mite antigen-induced and ovalbumin-induced AD models, producing reduced clinical scores, significant decreases in IgE production, diminished eosinophil and mast cell counts, and lowered expression of inflammatory cytokines through regulation of Th-cell responses including inhibition of Th17 cell proliferation.^136,137,141 Beyond visible inflammation, these EV-based interventions have been linked to functional barrier endpoints including reduced TEWL and improved expression of barrier-associated proteins, supporting the concept that anti-inflammatory effects and barrier repair are mechanistically coupled in AD pathophysiology. Neural stem cell-derived EVs similarly reduced pro-inflammatory cytokine expression, mast cell infiltration, and improved skin barrier integrity in DNCB-induced AD models, with proteomic analysis identifying multiple bioactive factors contributing to these effects.⁸⁷ Notably, S. epidermidis-derived EVs demonstrated therapeutic potential in calcipotriene-induced AD models, significantly reducing pro-inflammatory gene expression, increasing keratinocyte proliferation and migration, enhancing expression of antimicrobial peptides, and restoring skin homeostasis, highlighting the potential of commensal microbiota-derived EVs as a therapeutic strategy for restoring microbial balance in AD.⁸⁸

In diabetic wound healing models, adipose-derived MSC-EVs from various tissue sources demonstrated remarkable efficacy, enhancing wound closure rates, increasing collagen deposition, promoting re-epithelialisation, and stimulating angiogenesis in streptozotocin-induced diabetic mice and rats.^138,139 These EVs accelerated fibroblast migration and proliferation while reducing oxidative stress and inflammation in the wound microenvironment.^142,143 Mechanistic investigations confirmed that EV cargo composition and conditioning strategies substantially influence therapeutic potency. Hypoxia-preconditioned ADSC-derived exosomes showed distinct miRNA profiles enriched for fibroblast proliferation, migration, and immune regulation, with in vivo results demonstrating reduced IL-6 and increased VEGF expression, enhanced collagen I/III deposition, and increased angiogenesis markers including CD31, consistent with coordinated ECM remodelling and vascular repair.¹⁴⁴ Pluronic F (PF)-127/hADSC-exosome complexes improved wound closure, re-epithelialisation, and myofibroblast activation, while upregulating barrier-associated proteins including keratin-1 (KRT1) and aquaporin-3 (AQP3) and downregulating inflammatory mediators.¹⁴⁰ Mechanistic work further demonstrated that ADSC-exosomes mitigate endothelial dysfunction under hyperglycaemic stress by restoring mitochondrial redox regulation via the SIRT3/SOD2 axis,¹⁰⁷ while bone marrow MSC-derived exosomes promoted endothelial progenitor cell tube formation with amplified benefit when combined with pharmacologic Nrf2 activation.¹⁰⁸

Emerging evidence also supports EV regulation of epidermal cell programmes central to barrier restoration. ADSC-exosomes restored autophagy flux in epidermal cells under high-glucose conditions, improving proliferation and migration in vitro and enhancing epithelialisation in vivo, with mechanistic analyses implicating an autophagy–NAMPT–NAD axis as a potential contributor to stress resilience.¹⁴⁵ Keratinocyte-derived exosomes were shown to deliver miRNA cargo that shapes macrophage phenotype and inflammation resolution, thereby influencing recovery of functional barrier proteins including loricrin, FLG, ZO-1, and occluding. The inhibition of miRNA packaging in these exosomes increased persistence of pro-inflammatory macrophage states and impaired barrier integrity including increased TEWL, despite minimal effects on gross wound closure, underscoring that functional repair depends on EV-mediated immune–epithelial communication rather than re-epithelialisation alone.¹⁴⁶ Macrophage–keratinocyte exosomal crosstalk has further been linked to metabolic competence at the wound edge through TOMM70 transfer supporting keratinocyte mitochondrial bioenergetics and re-epithelialisation capacity.¹⁴⁷ Fibroblast-derived exosomes directly enhanced epidermal stem cell proliferation, migration, and differentiation in diabetic wounds through KEAP1/Nrf2 signalling, with longer-term healed skin quality assessments reporting increased collagen deposition and improved biomechanical properties, extending EV efficacy endpoints beyond closure to include functional integrity of repaired tissue.¹⁰⁹ Endothelial progenitor cell-derived EVs further enhanced wound healing in both control and diabetic mice through activation of the AGE–RAGE signalling pathway.⁹⁸

Clinical Translation: Current State and Critical Gaps

While preclinical evidence supporting MSC-EVs is robust, clinical studies investigating their therapeutic application in AD and chronic wounds remain limited in both number and scale. Current evidence consists of early-phase trials, small prospective studies, and case series, with larger randomised controlled trials either ongoing or yet to be initiated. These studies are presented not as an exhaustive evaluation of clinical efficacy, but to illustrate the current stage of clinical translation and highlight the disconnect between preclinical innovation and clinical implementation. Overall, MSC-EVs exhibit favourable safety profiles with low immunogenicity and toxicity across reported studies,^34,128 though long-term human safety data remain insufficient to draw definitive conclusions.

Clinical Studies in Atopic Dermatitis

One early-phase clinical study evaluated the safety and efficacy of human umbilical cord blood-derived MSCs, rather than isolated EVs in 34 adult patients with moderate-to-severe AD over a 12-week period.¹⁴⁸ Patients received either low-dose or high-dose subcutaneous injections and the study demonstrated an acceptable safety profile, providing translational safety data relevant to EV-based approaches despite not directly assessing EVs.¹⁴⁸ More directly relevant to EV therapy, clinical evidence suggests potential benefits of adipose-derived MSC exosomes in managing dupilumab-related facial redness (DFR), a common adverse effect in AD patients receiving biologic therapy. Case reports documented successful treatment of refractory DFR using electroporation-assisted topical application of human adipose tissue-derived MSC exosomes, with marked improvement in facial erythema and patient satisfaction.^128,149,150 These findings were substantiated by a 12-week prospective study involving 20 adult AD patients with DFR, in which quantitative assessment revealed significant reductions in erythema index by week 4, with decreases of 31, 27, 13, and 25 units observed on the forehead, chin, right cheek, and left cheek respectively.¹⁵¹ Molecular analysis of stratum corneum samples from this study demonstrated suppression of pro-inflammatory mediators including IL-1α and TSLP, alongside increased expression of FLG and VEGF, indicating simultaneous immunomodulatory and barrier-restorative effects following exosome treatment.¹⁴⁹ These studies are nonetheless limited by small sample sizes and the absence of randomised controls.^128,150

Clinical Studies in Chronic Wounds

Clinical translation of EV-based therapies for wound healing has advanced further than for AD, with a Phase II randomised trial now reporting on wound healing outcomes. A recent study evaluated a scalable ligand-based exosome affinity purification (LEAP) method for producing clinical-grade platelet-derived EVs (P-EVs).¹⁵² LEAP-purified P-EVs exhibited consistent size and cargo profiles and significantly enhanced fibroblast proliferation, migration, and angiogenesis in vitro, though no statistically significant difference in mean wound healing time was observed compared to placebo in the clinical arm, validating LEAP as a manufacturing approach while highlighting the need for optimised delivery and dosing.¹⁵² More encouraging results were reported in a Phase II randomised clinical trial (ClinicalTrials.gov identifier: NCT06812637) assessing Wharton’s jelly-derived MSC exosomes for DFUs.¹⁵³ In this study, 110 patients with chronic DFUs were randomised to receive standard of care alone, standard of care with placebo vehicle, or standard of care with weekly topical Wharton’s jelly MSC exosome application for four weeks. The exosome-treated group demonstrated significantly accelerated wound healing, with 62% of patients achieving complete closure by study completion, a mean epithelialisation time of six weeks compared to 20 weeks in controls, and marked ulcer size reduction as early as two weeks post-treatment.¹⁵³

Registered Clinical Trials

Registry data from ClinicalTrials.gov indicate growing clinical interest in EV-based therapies for chronic ulcers and genetic skin disorders. Phase I studies are evaluating conditioned media containing microvesicles or exosomes from Wharton’s jelly MSCs in chronic cutaneous ulcers (NCT04134676), the combined use of MSC-derived exosomes with nutritional correction in diabetic chronic skin ulcers (NCT05243368), and autologous exosome-rich plasma for intractable cutaneous wounds (NCT02565264). A Phase I/IIa randomised multicentre trial is investigating AGLE-102 (allogeneic MSC-EVs) for epidermolysis bullosa lesions (NCT04173650).¹⁵⁴ Across all registered trials to date, EV or conditioned media therapies are administered as free preparations such as topically, subcutaneously, or intravenously, without biomaterial-assisted delivery vehicles.

The Delivery Translational Gap: Free EV Administration Across All Current Trials

A systematic review in 2024 reported that the meta-analysis of 21 reports on EV human therapy confirms that while preclinical proof-of-concept for engineered systems is high, however, the clinical trials are only “emerging slowly” and primarily focus on safety using more traditional delivery routes.¹⁵⁵ Recent study of both AD and chronic wounds shows a consistent and striking pattern which every clinical trial to date has administered EVs as free preparations, whether via topical application, subcutaneous injection, or intravenous infusion. The shift to biomaterials in chronic wounds is seen as essential for stability but has not reached clinical stages and neither for AD. No registered or published clinical trial has incorporated biomaterial-assisted EV delivery, such as hydrogel encapsulation, scaffold integration, or sustained-release dressing systems, despite this being the dominant and increasingly standard delivery strategy in the preclinical literature.^36,156–158

At the preclinical level, biomaterial-assisted delivery is not a fringe strategy but a well-established paradigm with consistent mechanistic and efficacy advantages. ADSC-exosomes delivered via alginate-based hydrogel significantly improved wound closure, collagen synthesis, and neovascularisation compared to free EV controls.³⁹ Human umbilical cord MSC-derived exosomes delivered through genipin-crosslinked hydrogels enhanced early wound repair and reduced inflammatory responses while offering practical advantages over direct MSC transplantation.¹⁵⁹ PF-127/hADSC-exosome thermosensitive hydrogel complexes maintained therapeutic efficacy with less frequent administration, directly demonstrating the pharmacokinetic benefit of sustained release.¹⁴⁰ Keratinocyte-derived EVs delivered via GelMA hydrogel enhanced angiogenesis through PI3K/AKT-mediated signalling,¹⁶⁰ while a multifunctional sprayable hydrogel incorporating hUCMSC-derived exosomes achieved near-complete wound closure by day 14 alongside immunomodulatory and ROS-scavenging activity.¹⁶¹ Across these studies, the rationale for biomaterial-assisted delivery is clear: free EVs are rapidly cleared from the wound site, limiting their effective therapeutic window; encapsulation within hydrogels or scaffolds prolongs local retention, provides sustained cargo release, and protects EV bioactivity from the hostile wound microenvironment.

This translational disconnect may help explain, at least in part, why clinical outcomes with free EV administration have been modest or inconsistent relative to preclinical expectations. The Phase I platelet-derived EV trial, for example, validated the manufacturing process but failed to demonstrate a statistically significant improvement in wound healing time,¹⁵² a result that may reflect suboptimal EV retention at the wound site rather than a fundamental limitation of EV biology. Conversely, the more favourable outcomes reported in the Phase II trial of Wharton’s jelly MSC-exosomes, 62% complete wound closure and a mean epithelialisation time of six weeks,¹⁵³ may partly reflect the structured weekly topical dosing regimen. This approach likely compensates for the rapid clearance of EVs through repeated replenishment, rather than relying on sustained local release.

The implication is not that free EV administration is inherently ineffective, but that the field has yet to systematically test, in a clinical setting, the delivery strategies that preclinical evidence most strongly supports. Bridging this gap by translating biomaterial-assisted EV delivery systems from bench to clinical trial, represents both a key unmet need and a logical next step for the field. Future trials should consider incorporating controlled-release EV delivery vehicles, validated against free-EV comparator arms, to determine whether the pharmacokinetic advantages demonstrated preclinically translate into improved and more durable clinical outcomes. The contrast between preclinical and clinical delivery approaches is summarised in Figure 4. A consolidated summary of current clinical studies and registered trials is presented in Table 2.

Table 2 Summary of Clinical Studies and Registered Trials of EV Therapy in AD and Chronic Wounds

Figure 4 Translational comparison of preclinical vs clinical delivery strategies. This figure compares EV delivery strategies across preclinical and clinical settings, highlighting the widespread use of biomaterial-assisted systems such as hydrogels, microneedles, and scaffolds in preclinical studies, contrasted with the predominant use of free EV administration in clinical trials. The figure emphasises the translational gap between experimental delivery optimisation and current clinical practice.

Across reported studies, MSC-EVs demonstrate favourable safety profiles with low immunogenicity and toxicity,^34,128 and some animal models demonstrate preservation of body weight compared with corticosteroid-treated controls.⁸² Key translational challenges include potential immune responses to allogeneic or xenogeneic EVs, although these appear substantially lower than those associated with whole-cell therapies,^34,128 as well as unintended off-target effects mediated by EV-associated miRNAs.¹²⁶ Additional barriers include heterogeneity in dosing regimens and administration schedules,¹⁶² and the need for standardised storage conditions, as EV bioactivity may decline during prolonged storage.¹⁶³ Although subcutaneous injection remains the most commonly employed delivery route, intravenous and topical applications have demonstrated comparable efficacy,³⁴ though the biodistribution and rapid clearance of systemically administered EVs by the liver and spleen warrant further investigation.¹⁶⁴ Collectively, these findings underscore the necessity for rigorously designed, adequately powered clinical trials with standardised protocols and incorporating validated biomaterial-assisted delivery systems alongside free-EV comparator arms to fully establish the safety, efficacy, and translational feasibility of EV-based therapies in dermatologic and wound-healing applications.

Bridging the Gap: Delivery Systems as the Missing Link

Why Delivery Matters: The EV Clearance Problem

The therapeutic potential of EVs is substantially constrained by their pharmacokinetic limitations following administration. When administered systemically, approximately 99% of EVs are eliminated within 24 hours, primarily sequestered by the liver and spleen through phagocytic clearance.^104,164 Local administration is similarly compromised by rapid dispersal from the wound site, enzymatic degradation within the protease-rich wound microenvironment, and dilution through wound exudate.¹⁶⁵ These clearance dynamics create a fundamental mismatch between the sustained signalling requirements of skin barrier repair, which unfolds over days to weeks, and the brief residence time of free EV preparations. In the context of both AD and chronic wounds, where repeated or prolonged therapeutic signalling is necessary to overcome established pathological cycles, this limitation is not a minor inconvenience but a fundamental barrier to clinical efficacy.

This clearance problem explains, at least in part, why the otherwise impressive preclinical pharmacology of MSC-EVs has not yet translated into consistently strong clinical outcomes. The integration of EVs with biomaterial delivery systems represents the most rationally justified strategy for overcoming this limitation, and the preclinical evidence for its superiority over free EV administration is now substantial.^166,167

Biomaterial Platforms for EV Delivery

A diverse range of biomaterial platforms has been investigated for EV delivery in skin barrier repair, each offering distinct pharmacokinetic, mechanical, and disease-specific advantages. These platforms are not interchangeable, and their optimal application differs meaningfully between AD and chronic wound contexts.

Hydrogels

Hydrogels remain the most extensively investigated biomaterial platform for EV delivery in wound healing and skin barrier applications.^168,169 These three-dimensional hydrophilic polymer networks encapsulate EVs while maintaining their bioactivity, providing a moist wound environment conducive to tissue regeneration, and enabling sustained, controlled EV release over extended periods.^150,170 A wide array of natural and synthetic polymers has been employed, including chitosan, gelatin methacryloyl (GelMA), collagen, hyaluronic acid, alginate, silk fibroin, and PF-127.^39,171,172 Beyond sustained release, EV-loaded hydrogels protect EV cargo from enzymatic degradation and the harsh, protease-rich chronic wound microenvironment, improve EV retention at the application site, and can be formulated to incorporate additional therapeutic agents for synergistic effects. Their tuneable physical properties, including mechanical stiffness, porosity, degradation kinetics, and gelation temperature, allow hydrogel composition to be matched to the mechanical and biochemical requirements of the target tissue.^173,174 Preclinical studies have consistently demonstrated enhanced therapeutic outcomes with EV-loaded hydrogels compared to free EV administration, including accelerated wound closure, enhanced angiogenesis, and increased collagen deposition.^{39,140,159–161,171,172}

Microneedle Arrays

Microneedle (MN) arrays represent a particularly important delivery platform for AD, where the thickened and structurally abnormal stratum corneum represents a formidable physical barrier to topically applied EVs. Conventional topical application deposits EVs on the skin surface, from which penetration into the viable epidermis is negligible given the size of EV particles.^175,176 Dissolving, solid, and hydrogel-forming MN patches penetrate the stratum corneum via micron-scale channels, delivering EVs directly into the viable epidermis or dermis, substantially improving bioavailability while avoiding systemic exposure.^177,178 Bui et al demonstrated that dissolving MN patches enable long-term storage of EVs within the polymer matrix while maintaining bioactivity upon dissolution, addressing a key practical challenge for clinical product stability.¹⁷⁷ Critically, an EV-loaded hyaluronic acid-based MN platform has demonstrated direct intradermal EV delivery with sustained release and minimal skin irritation, establishing proof of concept for the MN-EV approach in skin applications.¹⁷⁹ In the context of AD, MN-delivered S. epidermidis-derived EVs loaded with the antibiotic levofloxacin have demonstrated combined antimicrobial activity, immune modulation, and enhanced skin penetration in a murine AD model, illustrating the combinatorial potential of MN platforms.¹⁷⁹ EV-loaded MN patches have also been demonstrated for diabetic wound treatment, with dissolvable MN-based dressings enabling transdermal and continuous delivery of platelet-derived exosomes with anti-inflammatory and pro-angiogenic effects.^180,181

Electrospun Nanofibre Scaffolds

Electrospun nanofibre scaffolds, fabricated from materials including polylactic-co-glycolic acid (PLGA), collagen, polycaprolactone (PCL), and hyaluronic acid, offer a distinct biomaterial architecture that simultaneously provides physical ECM-mimicking support and controlled EV delivery.¹⁸² The fibrous morphology of these scaffolds guides fibroblast and keratinocyte migration directionally, a property not available from hydrogels alone. EV loading is typically achieved by tethering electronegative EVs to electropositive fibre surface modifications, enabling slow, degradation-controlled release as the scaffold breaks down in vivo.⁹⁵ Furthermore, hUCMSC-EVs preconditioned under serum and glucose deprivation were tethered to PLGA nanofibrous scaffolds decorated with electropositive components, enabling sustained miR-29a-3p-mediated Wnt/β-catenin activation and enhanced angiogenesis in diabetic wound models.⁹⁵ This approach demonstrates how scaffold architecture and EV cargo can be co-designed to produce composite systems where the structural and pharmacokinetic properties of the delivery platform reinforce the biological action of the EV payload.

Three-Dimensional Bioprinted Constructs

Three-dimensional bioprinting technology enables spatially precise distribution of EVs within engineered tissue constructs, an advantage uniquely suited to chronic wounds with complex or irregular geometries.^138,183 Bioinks can be formulated to incorporate EVs alongside cells and ECM components, creating layered architectures that recapitulate the structural organisation of native skin.¹³⁸ This approach holds particular promise for generating personalised skin grafts with tailored EV compositions, concentrations, and spatial release profiles matched to patient-specific wound characteristics and healing requirements.¹⁸⁴ The key challenges are maintaining EV bioactivity during the printing process, which may involve shear stress, temperature variation, and UV crosslinking exposure, and validating that printed constructs maintain intended structure and function in vivo.^185,186 Despite these challenges, 3D bioprinting represents a frontier delivery platform that, as manufacturing processes mature, could enable a level of therapeutic personalisation not achievable with conventional EV delivery approaches.¹⁸⁷

Nanoparticle Carriers

Nanoparticle (NP) carriers, including lipid nanoparticles (LNPs), PLGA nanoparticles, and magnetic iron oxide nanoparticles, offer an alternative strategy for enhancing EV stability, protecting EV cargo from degradation, and enabling targeted delivery to specific cell types within the skin.^188,189 Cell-derived nanovesicles (CDNs), produced by mechanically shearing MSCs through membranes of decreasing pore sizes, mimic the proteomic and functional properties of MSC-EVs while circumventing the low-yield and batch-variability limitations of conventional EV isolation.¹⁸⁸ Surface functionalisation of NP carriers with targeting ligands, antibodies, or cell-penetrating peptides can direct EVs preferentially to keratinocytes, fibroblasts, or immune cells, enabling more precise therapeutic interventions than non-targeted free EV administration.^36,189 For wound applications, magnetic NP-labelled exosomes have been demonstrated to enable site-directed delivery under an external magnetic field, enhancing wound-site accumulation and reducing off-target uptake.³⁶ EV-mimetic nanovesicles, such as melatonin-loaded EV-mimetic nanoparticles, have demonstrated improved AD symptoms on topical administration compared to free drug,¹³⁶ illustrating the principle that NP-based EV engineering can simultaneously improve delivery and therapeutic payload capacity.

In addition to naturally derived EVs, EV mimetics have emerged as an alternative strategy to overcome limitations in yield and scalability. These vesicle-like structures are typically generated through mechanical extrusion, sonication, or synthetic assembly, enabling the production of large quantities with higher drug loading efficiency.^190,191 EV mimetics retain key membrane properties of natural vesicles while allowing improved control over composition and cargo incorporation.¹⁹² Although still in early stages of development, they represent a promising platform for scalable and cost-effective delivery systems in both AD and chronic wound applications.

Disease-Specific Delivery Considerations

The distinct pathophysiology of AD and chronic wounds creates fundamentally different delivery requirements that should inform the selection and design of EV carrier systems. A platform optimally suited to one condition may be suboptimal or inapplicable to the other, and the preclinical literature increasingly reflects an understanding that delivery strategy and disease context must be co-optimised.

In AD, the primary delivery challenge is epidermal penetration. The thickened, hyperkeratotic stratum corneum that characterises AD lesions, paradoxically a consequence of deficient barrier proteins such as FLG rather than a structurally reinforced barrier, nonetheless presents a physical obstruction to topically applied nanoparticles and EVs.^45,46 EV particle sizes of 30–200 nm cannot passively diffuse through intact stratum corneum, and the inflammation-driven structural abnormalities of AD skin create an unpredictable and variable permeation surface. Microneedle arrays are the most rationally justified primary delivery platform for AD as they bypass the stratum corneum entirely, delivering EVs directly into the viable epidermis where keratinocytes, Langerhans cells, and resident immune cells reside. Hyaluronic acid-based dissolving MNs are particularly attractive for this indication because hyaluronic acid has intrinsic skin barrier and moisture-retention properties that may complement EV-mediated ceramide and FLG restoration.¹⁷⁹ Where MN delivery is not feasible, electroporation-assisted topical delivery as employed in clinical studies of MSC exosomes for dupilumab facial redness^149,151 represents an alternative approach to transiently disrupt the stratum corneum and improve EV penetration. Secondary platforms, including hydrogels, may be appropriate for overnight occlusive application to chronically inflamed lesions, providing both sustained EV release and a moist, protective environment.

In chronic wounds, the delivery challenge is sustained wound bed coverage under hostile microenvironmental conditions. Open wound surfaces are exposed to elevated protease activity, acidic pH, reactive oxygen species, and bacterial colonisation, all of which degrade free EV preparations rapidly.^50–55 Hydrogels are therefore the primary delivery platform of choice for chronic wounds: they physically encapsulate and protect EVs from enzymatic degradation, maintain a moist wound environment conducive to healing, provide conformable wound coverage, and enable tunable sustained release matched to the multi-week timescale of chronic wound repair.^168,169 Sprayable or injectable hydrogel formulations, such as the Exo@AMCN system,¹⁶¹ are particularly practical for irregular wound geometries where solid scaffolds cannot be fitted. For deeper or tunnelling wounds, electrospun nanofibre scaffolds offer additional structural support and directional cell guidance alongside sustained EV delivery.^95,182 3D bioprinted constructs represent the most sophisticated option for large, complex chronic wounds where personalised graft geometry and spatial EV distribution are required.¹⁸⁴ Table 3 provides a comparative summary of EV delivery platforms and their disease-specific application rationale.

Table 3 Biomaterial-Assisted and Engineered EV Delivery Strategies for Skin Barrier Repair

The excipient composition of EV delivery formulations also differs meaningfully between AD and chronic wound applications, reflecting the distinct physicochemical environments and therapeutic requirements of each condition. In AD formulations, hyaluronic acid is the most commonly employed excipient, selected for its intrinsic skin barrier and moisture-retention properties, its capacity to form dissolving microneedle matrices compatible with EV cargo, and its established safety profile in topical dermatological preparations.¹⁷⁹ PF-127 has been used as a thermosensitive hydrogel excipient in wound applications, where its sol-gel transition at body temperature enables easy application as a liquid followed by in situ gelation at the wound site, providing conformable wound coverage without requiring surgical placement.¹⁴⁰ Genipin serves as a natural crosslinking agent for collagen and chitosan hydrogels in wound delivery systems, offering lower cytotoxicity than glutaraldehyde while providing sufficient mechanical stability for wound bed adherence.¹⁵⁹ Alginate is preferred in wound applications where divalent cation-mediated ionic crosslinking enables rapid gel formation under physiological conditions, and its high-water content supports the moist wound environment required for effective re-epithelialisation.³⁹ Dissolving microneedle matrices have been specifically engineered to enable long-term storage of EVs within the polymer matrix while maintaining bioactivity upon dissolution in the viable epidermis, providing a practical clinical advantage for EV product stability and shelf life.^177,178

The Translational Gap: Absence of Biomaterial-Assisted Delivery in Clinical Trials

Despite the compelling preclinical evidence, no clinical trial of EV-based therapy in AD or chronic wounds has incorporated biomaterial-assisted delivery. This represents a translational paradox: the delivery strategy most strongly supported by the preclinical literature is entirely absent from the clinical trial landscape.

Several interlocking factors contribute to this disconnect. Regulatory uncertainty is perhaps the most significant. EV-based therapeutics already occupy an ambiguous regulatory position, classified variably as biological products, advanced therapy medicinal products (ATMPs), or drug delivery systems depending on jurisdiction, source, and manufacturing process.^194–197 The addition of a biomaterial carrier introduces a second regulatory layer. EV-hydrogel or EV-scaffold combinations may be classified as combination products or medical device-drug combinations, triggering dual regulatory pathways with separate safety, efficacy, and quality requirements. The regulatory complexity of combination products substantially extends development timelines and increases costs, creating a powerful disincentive to incorporate biomaterial delivery into early-phase trials even when the pharmacological rationale is strong.

Manufacturing complexity is the second major barrier. Producing clinical-grade EVs that meet the Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines for purity, particle size consistency, and functional potency is already technically demanding.^{110,198–200} Incorporating these EVs into a biomaterial carrier, whether by hydrogel encapsulation, nanofibre loading, or MN patch fabrication, introduces additional manufacturing steps, each of which must be validated for GMP compliance, sterility, and preservation of EV bioactivity. The thermosensitive hydrogel and MN patch fabrication processes, for example, involve conditions including mechanical stress, polymer crosslinking, temperature cycling, that may compromise EV membrane integrity and cargo bioactivity if not carefully optimised.^185,186 Establishing GMP-compatible manufacturing processes for EV-biomaterial combination products remains an open engineering challenge, and few academic or spin-out EV therapy developers have the manufacturing infrastructure to address it at clinical scale.

Cost and development resources represent the third barrier. EV therapy development is capital-intensive, and early-phase clinical trials are typically funded through academic grants or early-stage venture investment with limited budgets. The additional cost of developing, validating, and manufacturing an EV-biomaterial combination product, relative to a free EV preparation, is substantial. Developers rationally choose the simplest viable formulation for proof-of-concept clinical studies, deferring biomaterial integration to later development phases.

The preclinical evidence that biomaterial-assisted delivery produces superior outcomes is mechanistically grounded and reproducibly demonstrated, not merely incremental. Free EVs are suboptimal delivery vehicles, not because EV biology is insufficient, but because the pharmacokinetic limitations of unencapsulated EVs prevent their biological potential from being realised at the wound site over the timescales required for tissue repair. Biomaterials are not accessories to EV therapy; they are enablers that determine whether EV pharmacology can be translated from a controlled preclinical environment in which EVs are applied in large quantities, repeatedly, under optimised conditions, to a clinical wound bed where retention, protection, and sustained activity are prerequisites for efficacy. The field cannot afford to wait for mature combination product regulatory frameworks before beginning to test these systems in clinical settings; early engagement with regulatory agencies through pre-Investigational New Drug (IND) meetings and scientific advice procedures, as recommended by the European Medicines Agency (EMA) and United States Food and Drug Administration (FDA) is the necessary next step.^155,197

Beyond Delivery: EVs as Drug Carriers

In parallel with advances in biomaterial-assisted EV delivery, EVs themselves can function as sophisticated drug delivery vehicles, leveraging their natural biocompatibility, capacity for biological barrier traversal, and low immunogenicity to transport therapeutic molecules to target cells.^201,202 EVs can be loaded with small molecule drugs, therapeutic proteins, miRNAs, or gene-editing constructs through strategies including incubation of donor cells with drugs to promote endogenous packaging, electroporation or sonication for direct loading of isolated EVs, chemical transfection, and genetic engineering of producer cells to incorporate therapeutic molecules into EV cargo.^203,204 Engineered EVs loaded with small molecule drugs have demonstrated enhanced wound healing outcomes in preclinical models, including VH298-loaded EVs released from GelMA hydrogels, which facilitated diabetic wound healing through HIF-1α-mediated enhancement of angiogenesis,¹³⁸ and melatonin-loaded EV-mimetic nanovesicles, which improved AD symptoms on topical administration.¹³⁶ Surface modification of EVs with targeting peptides, antibodies, or other ligands can further direct EVs toward specific cell populations including keratinocytes, fibroblasts, or immune cells, enabling more precise therapeutic interventions and reducing off-target exposure.^205,206 When combined with biomaterial scaffolds, drug-loaded engineered EVs represent a multi-modal therapeutic strategy in which the delivery system, the EV carrier, and the EV cargo each contribute distinct and complementary therapeutic functions.^207,208

EVs also can be further enhanced through genetic engineering of parent cells to enable selective loading of therapeutic cargo, including miRNAs, siRNAs, and proteins.^209,210 This approach allows targeted modulation of key signalling pathways such as NF-κB, AKT, and VEGF, thereby enhancing immunomodulatory and regenerative effects.¹⁵⁷ Engineered EVs have demonstrated improved specificity in regulating inflammatory responses in AD and promoting angiogenesis and tissue repair in chronic wounds.^211,212 Compared to native EVs, this strategy offers greater control over cargo composition and therapeutic function, representing a promising direction for precision medicine applications.²¹³

Challenges and Future Directions

Standardisation and Scalability

A critical challenge in the clinical translation of EV-based therapies is the substantial variability in isolation methods, characterisation protocols, and dosage regimens across studies.^110,198 Current isolation techniques such as ultracentrifugation, density gradient centrifugation, size-exclusion chromatography, ultrafiltration, precipitation, and affinity capture, yield EV populations with differing purity, yield, and functional properties, limiting cross-study comparability.^199,200,209 The MISEV guidelines provide a framework for standardised characterisation including EV marker quantification (CD9, CD63, CD81, TSG101, Alix), particle size distribution analysis by nanoparticle tracking analysis or transmission electron microscopy, and assessment of non-EV contaminants,¹¹⁰ but adherence remains inconsistent.¹⁹⁸ Dosage variability is similarly unresolved, with preclinical studies employing widely varying EV doses, frequencies, and treatment durations in the absence of validated functional potency assays.^34,162,163

Scaling EV production from research-grade to clinical quantities remains technically challenging. Three-dimensional culture systems, including microcarrier-based and hollow-fibre bioreactors, offer improved scalability and can increase EV yields 10–20 fold compared to conventional two-dimensional cultures.^214,215 Tangential flow filtration has emerged as a promising large-scale isolation approach offering serum-free production and reduced batch-to-batch variation.²¹⁴ Standardised biomanufacturing protocols must address cell source selection and passage number limitations, culture medium composition, priming or preconditioning strategies, harvest timing, isolation methods, and storage formulations that preserve long-term EV bioactivity.^163,216 These scalability requirements become more demanding when EVs must be co-processed with biomaterial carriers for combination product manufacturing, underscoring that the clinical translation pathway for EV-biomaterial combination therapies is longer and more resource-intensive than for free EV preparations alone.

The clinical translation of EV-based therapeutics requires robust and reproducible manufacturing pipelines that comply with Good Manufacturing Practice (GMP) standards. GMP-compliant production encompasses three critical stages, including upstream processing, downstream purification, and quality control.^197,217 Upstream processes involve the use of well-characterised and traceable cell sources cultured under xeno-free or serum-free conditions to minimise variability and contamination risks.²¹⁸ Large-scale production increasingly utilises bioreactor systems to enhance yield and standardisation.²¹⁹ Downstream processing includes validated isolation and purification techniques such as ultrafiltration, size-exclusion chromatography, and tangential flow filtration, which are essential to ensure EV purity, integrity, and scalability.^197,217 Despite recent advances, major challenges remain, including batch-to-batch variability, low yield, and the lack of universally accepted potency assays, which continue to hinder large-scale clinical translation.¹¹¹

Rigorous characterisation is required to confirm the identity, purity, potency, and cargo composition of therapeutic EV preparations in accordance with the MISEV guidelines.^110,111 Particle size distribution and concentration are assessed by nanoparticle tracking analysis and transmission electron microscopy, which together provide quantitative size profiles and direct morphological confirmation of EV membrane structure.¹¹⁰ Expression of canonical EV surface markers such as the tetraspanins CD9, CD63, and CD81, alongside biogenesis-associated proteins TSG101 and Alix, must be confirmed by Western blotting or bead-based flow cytometry, while co-isolated contaminants including albumin should be assessed and reported.^110,111 Cargo characterisation is equally critical for therapeutic EVs intended for skin barrier applications. Small RNA sequencing or quantitative PCR panels targeting key therapeutic miRNAs, including miR-21-3p, miR-125a, miR-126-3p, and miR-29a-3p in MSC-EV preparations, provide cargo identity data and serve as potency indicators for immunomodulatory and angiogenic activity respectively.¹²⁹ Functional potency assays should include an in vitro readout directly relevant to the intended mechanism of action: for AD-targeted EVs, a keratinocyte cytokine suppression assay or Th2 polarisation inhibition assay; for wound healing EVs, a scratch wound closure or tube formation assay using endothelial cells.^34,110 These assays serve as release criteria and batch comparability tests, and the absence of validated, quantitative potency assays that correlate with in vivo efficacy endpoints remains one of the most significant regulatory science challenges for the clinical translation of EV-based dermatological therapies.¹¹⁰

Safety and Regulatory Hurdles

MSC-EVs generally exhibit low immunogenicity compared to whole-cell therapies; however, EVs carry surface molecules including MHC proteins that could potentially trigger immune responses with repeated or allogeneic administration.^126,220 Off-target effects mediated by the pleiotropic miRNA cargo of EVs represent an additional safety consideration, as each miRNA can regulate multiple target genes.²²¹ Long-term safety monitoring is particularly important given the limited data on chronic EV exposure; theoretical concerns include nucleic acid or protein accumulation in recipient tissues and potential effects on cancer development or progression,¹⁷¹ although current evidence supports favourable safety profiles.³⁴

Regulatory pathways for EV-based therapeutics remain poorly defined in most jurisdictions.¹⁰⁴ EVs occupy a regulatory grey zone, potentially classified as biological products, cell therapies, or drug delivery systems depending on source, manufacturing process, and intended use.¹⁹⁴ Establishing identity criteria adequate for the inherent complexity and heterogeneity of EV products, developing validated potency assays that predict clinical outcomes, and designing appropriate nonclinical safety studies represent the key regulatory science challenges.^195,196 For EV-biomaterial combination products, these requirements are compounded by the need to demonstrate that the biomaterial carrier does not adversely affect EV bioactivity, safety, or pharmacokinetics. The EMA and FDA have begun issuing guidance applicable to advanced therapy medicinal products,^197,222 and early engagement with regulatory authorities through pre-IND or scientific advice procedures is strongly recommended to clarify expectations and reduce development risk.¹⁵⁵

A Roadmap for Clinical Translation

The study supports a phased roadmap for translating biomaterial-assisted EV therapy from bench to clinic, structured around short-term, medium-term, and long-term research and regulatory priorities.

In the short term, the priority is systematic head-to-head comparison of free EV versus biomaterial-encapsulated EV administration in standardised animal models of both AD and chronic wounds. Such studies should use matched EV preparations and doses to isolate the pharmacokinetic effect of the delivery vehicle from the biological effect of the EV cargo, and should include functional barrier endpoints (TEWL, FLG expression, collagen deposition) in addition to gross wound closure metrics. The PF-127 hydrogel system¹⁴⁰ demonstrated that less frequent dosing of encapsulated EVs maintains equivalent efficacy to more frequent free EV dosing, and this pharmacokinetic framework should be replicated systematically across hydrogel and MN platforms to generate the comparative evidence base required to justify combination product clinical development. Standardised potency assays and release specifications for EV-biomaterial combination products should be developed in parallel, in consultation with regulatory agencies.

In the medium term, first-in-human trials of EV-biomaterial combination products should be initiated for the two indications with the strongest preclinical rationale. For chronic wounds, EV-hydrogel combination dressings represent the most clinically ready platform, building on the Phase II WJ-MSC exosome trial¹⁵³ and the extensive hydrogel preclinical literature.^{30,145,147,150,151} A Phase I/II trial comparing weekly free WJ-MSC exosomes (the current best-evidenced clinical regimen) against a single-application sustained-release EV-hydrogel dressing would directly test the translational hypothesis of this review. For AD, microneedle-EV combination products represent the most logical first clinical target, with MN patch technology already having a clinical regulatory precedent in vaccine delivery and MN-assisted topical drug delivery for skin conditions.¹⁸⁰ A Phase I safety and pharmacokinetic study of dissolving MN patches loaded with IFN-γ-primed MSC-EVs, compared to direct subcutaneous injection or topical application of matched EV preparations, would establish the clinical bioavailability advantage and safety profile of intradermal EV delivery in AD.

In the long term, defined here as beyond seven years, the regulatory pathway for “EV + biomaterial” combination products must be formally established. This requires proactive engagement between EV therapy developers, biomaterial manufacturers, and regulatory agencies to agree on classification, quality, and evidence requirements for combination products that do not fit neatly into existing biological product or medical device frameworks. Development of harmonised international guidelines for the quality and clinical evaluation of EV-biomaterial combination products would substantially reduce regulatory uncertainty and accelerate the timeline from bench to clinical application. Concurrently, large-scale bioreactor manufacturing processes capable of producing clinical-grade EV-biomaterial combination products at commercial scale must be validated, and health-economic models should be developed to assess the cost-effectiveness of combination product approaches relative to current standard of care in both AD and chronic wound populations.

Conclusion

This review integrates current mechanistic and preclinical evidence to demonstrate that EVs represent a versatile and biologically sophisticated platform for skin barrier repair. A central theme emerging from the literature is the context dependent plasticity of EV function. In AD, their primary therapeutic effect lies in restoring immune homeostasis through suppression of Th2 and Th17 driven inflammation and modulation of innate immune responses. In contrast, in chronic wounds, EVs predominantly facilitate tissue regeneration by promoting angiogenesis, fibroblast activation, and ECM remodelling, while immune modulation plays a supportive role in resolving persistent inflammation. This divergence underscores the importance of disease specific therapeutic targeting rather than a one size fits all approach.

However, a critical gap remains between preclinical success and clinical translation. While numerous studies demonstrate robust efficacy in animal models, these advances have not been matched by equivalent progress in clinical trial design. Limitations in delivery efficiency, retention at target sites, dosing standardisation, and large-scale production continue to hinder clinical application. The current paradigm, which often relies on direct EV administration, may be insufficient to achieve sustained therapeutic effects in complex human skin pathologies.

Future progress will depend on the integration of EVs with advanced delivery systems. Biomaterial assisted approaches, including hydrogels, scaffolds, and engineered matrices, offer the potential to enhance vesicle stability, localisation, and controlled release, thereby improving therapeutic efficacy. In parallel, standardisation of isolation methods, characterisation protocols, and potency assays will be essential to ensure reproducibility and regulatory compliance.

In conclusion, EVs remain a highly promising therapeutic modality for skin barrier repair, supported by strong mechanistic rationale and preclinical evidence. Their successful translation into clinical practice will depend not only on further biological understanding but also on innovation in delivery strategies and clinical trial design.