Extracellular Vesicles as Therapeutic Tools for the Treatment of Chronic Wounds

Chronic wounds develop when the orderly process of cutaneous wound healing is delayed or disrupted. Development of a chronic wound is associated with significant morbidity and financial burden to the individual and health-care system. Therefore, new therapeutic modalities are needed to address this serious condition. Mesenchymal stem cells (MSCs) promote skin repair, but their clinical use has been limited due to technical challenges. Extracellular vesicles (EVs) are particles released by cells that carry bioactive molecules (lipids, proteins, and nucleic acids) and regulate intercellular communication. EVs (exosomes, microvesicles, and apoptotic bodies) mediate key therapeutic effects of MSCs. In this review we examine the experimental data establishing a role for EVs in wound healing. Then, we explore techniques for designing EVs to function as a targeted drug delivery system and how EVs can be incorporated into biomaterials to produce a personalized wound dressing. Finally, we discuss the status of clinically deploying EVs as a therapeutic agent in wound care.


Introduction
Cutaneous wound healing is complex, consisting of overlapping processes: hemostasis/coagulation, inflammation, proliferation, and remodeling [1]. This requires intercellular communication among resident cells and entering immune cells through soluble, membrane-bound, and extracellular matrix (ECM) molecules [1,2]. Wounds that fail to heal in a timely process are called chronic wounds [3]. A 2004 meta-analysis found that in the United States skin ulcers and wounds were associated with USD 9.7 billion in annual direct medical costs [4]. For patients, chronic wounds cause pain, loss of productivity, a profound impact on quality of life, and increased mortality [4][5][6]. Risk factors for the development of chronic wounds include advanced age, diabetes mellitus with associated peripheral vascular disease and peripheral neuropathy, as well as chronic kidney disease and immobility [5,7]. The societal burden of chronic wounds will increase as the population ages and the prevalence of co-morbid chronic conditions continues to rise. Current advanced therapies, including topical application of growth factors [8], extracellular matrix products [9], and skin substitutes [10], are not always effective [11]. Therefore, it is imperative that cutting-edge therapeutics be identified to treat chronic wounds.
The goals of this review are as follows: (1) briefly discuss the benefits and limitations of mesenchymal stem cell (MSC) therapy for treating chronic wounds and how MSC

Extracellular Vesicles
Extracellular vesicles (EVs) are lipid bilayer vesicles that can be secreted by all cell types [48]. The term "extracellular vesicle" is generic, referring to any lipid bilayer secreted vesicle. EVs are a heterogeneous population consisting of exosomes, microvesicles, and apoptotic bodies (Figure 1a). They differ in size, morphology, density, cargo, biogenesis, and biologic activity. Given the heterogeneity and challenges distinctly classifying these populations, we will use the term EV throughout the paper to refer to all classes of vesicle. Microvesicles are released by ectocytosis and budding from the plasma membrane. Apoptotic bodies form from cells undergoing apoptosis and may contain fragmented organelles. (b) Depiction of select EV contents that contribute to wound healing. Additional EV contents are discussed in the main text and Table 1. (c) Released EVs interact with a recipient cell through membrane receptors thereby initiating intracellular signaling. EVs can also deliver their cargo to the recipient cell following endocytosis and back fusion within an MVE or by direct fusion with the plasma membrane. Lysis of EVs in the extracellular space releases contents that then act on the recipient cells.
Exosomes are secreted intraluminal vesicles (ILV). Inward budding of the endosomal membrane results in the formation of ILVs in a multivesicular endosome (MVE). Exosomes are released by fusion of the MVE with the plasma membrane [49]. Microvesicles form as outward protrusions (ectocytosis) of the plasma membrane [50]. A cell undergoing apoptosis breaks down its cellular components and organelles and packages them into apoptotic bodies [51,52].
Tetraspanin proteins are enriched in the membrane of EVs and regulate membrane structure, trafficking, and fusion with recipient cells [66]. EVs also contain adhesion molecules, ESCRT proteins, heat-shock proteins, cytoskeletal proteins, enzymes, and proteins involved in antigen presentation, membrane trafficking, and signal transduction [48,53]. Additionally, proteins such as Wnt3a are associated with the exterior of EVs [67].
The biodistribution of EVs is dependent on their cell of origin and expression of surface molecules [56]. Their half-life in circulation ranges from minutes to hours [71][72][73]. Clearance by the reticuloendothelial system can be prolonged by the expression of antiphagocytic surface proteins: CD47, THBS-1, and SIRPα [74]. Upon reaching a target EVs may bind surface receptors initiating intracellular signaling or deliver their contents by endocytosis or fusion with the plasma membrane ( Figure 1c) [48]. Additionally, the lysis of EVs releases their cargo into the extracellular space [75,76].

Role for MSC Extracellular Vesicles in Wound Healing
There is accumulating pre-clinical evidence that MSC-EVs are beneficial in cutaneous wound healing (Table 1). In this section, we will discuss how MSC-EVs can influence key components of the wound healing process.   MT-EVs were more potent than EVs at polarizing macrophages to M2 phenotype.

Study EV Source Model Findings
Yu et al. 2020 [98] Human BM-MSC

Rat diabetic wound -Local injection
EVs from MSCs treated with atorvastatin (ATV-EVs). EVs promoted wound healing and angiogenesis. ATV-EVs were more effective.

In vitro endothelial cells (HUVECs)
EVs promoted proliferation, migration, and tube formation, increased VEGF secretion, and activated AKT/eNOS signaling. ATV-EVs produce a larger magnitude effect compared to standard EVs. ATV-EVs proposed to work by upregulating miR-221-3p in endothelial cells.

Mouse burn wound -Topical in alginate hydrogel
EVs accelerated wound closure, increased epithelial thickness, collagen deposition, and neovascularization.

Rat burn wound -Local injection
EVs accelerated re-epithelialization, reduced scar width, promoted collagen maturation, and stimulated neoangiogenesis. Effects depended on EV transfer of Wnt4.

In vitro fibroblasts and endothelial cells (HUVECs)
EVs stimulated proliferation and migration, stimulated Collagen I and III, and elastin secretion, and promoted tube formation.

Inflammation
Wound healing is initiated immediately following tissue injury. Vascular injury and serum-derived factors promote clot formation and hemostasis at the site of trauma. There is rapid local production of pro-inflammatory cytokines (e.g., IL-1, IL-2, IL-6, IL-8, TNFα, interferons (IFNs), and prostaglandins) and growth factors (TGF-β, EGF, PDGF, and FGF) [102]. These factors promote the migration of inflammatory cells into the wound environment.
Neutrophils are the first inflammatory cell recruited. They are critical for controlling the invasion of bacteria through the compromised cutaneous barrier ( Figure 2). Neutrophils remove bacterial seeding by phagocytosis, producing reactive oxygen species (ROS), and releasing cytotoxic molecules [103]. The molecules released by neutrophils also promote the breakdown and clearance of cellular debris. Dysfunctional neutrophils may contribute to the formation of chronic wounds. Neutrophils from patients with diabetes mellitus (DM) have an impaired respiratory burst, a weaker chemotactic response, and are more prone to apoptosis [104,105]. The function of genetically defective neutrophils can be improved with EV treatment [106]. In this context EVs may be able to restore impaired neutrophil function associated with diabetes, potentially resulting in the recruitment of fewer neutrophils during the wound healing process. Additionally, excessive neutrophil recruitment is also found in chronic wounds [107]. MSC-EVs can inhibit the infiltration of neutrophils into corneal wounds [108]. It remains to be determined if inhibition of neutrophil infiltration is due to EVs acting on neutrophils, or if it is a response to reduced inflammatory cytokine secretion into the wound environment. The effects of MSC-EVs on neutrophils, and other cells to be discussed, may seem contradictory. But it is important to consider that like MSCs, MSC-EVs are likely working to restore tissue homeostasis and do not act on any one cell in isolation. Macrophages play a dual role in wound healing. Murine studies suggest that macrophages initially assume the M1 pro-inflammatory phenotype. M1 macrophages release proinflammatory cytokines and phagocytose bacteria, ECM, and apoptotic cells. After damaged tissues have been cleared, the wound progresses into the proliferative phase. For this transition to appropriately occur, macrophages must also transition from their proinflammatory M1 phenotype to their anti-inflammatory M2 phenotype. M2 macrophages act to resolve inflammation through the secretion of anti-inflammatory cytokines such as IL-10 and IL-1RA. M2 polarized macrophages are a key source of growth factors (EGF, TGF-β, IGF-1) that regulate the proliferative phase and promote fibrosis. Importantly, inappropriate macrophage activation has been linked to scarring and the development of chronic wounds [109].
Numerous studies have investigated the influence of EVs on macrophages, reviewed elsewhere [110]. In the context of wound healing and tissue repair, EVs promote polarization to the M2 macrophage phenotype [80,84,94,97,101]. Acquisition of the M2 phenotype is associated with reduced expression of pro-inflammatory cytokines (TNF-α, IL-1, IFN-γ) and increased expression of anti-inflammatory cytokines (IL-4, IL-10). He et al. found intravenously (IV) injected BM-MSCs home to the wound site, promote M2 macrophage polarization, and improve wound healing [80]. These BM-MSCs failed to promote wound healing if macrophages were depleted, or if the BM-MSCs were unable to secrete EVs. Finally, they proposed that the effect was due to the transfer of miR-223 to macrophages [80]. EVs may also promote M2 polarization through the transfer of miR-let7 [94,111], miR-181c [101], and miR-182 [112].
Apoptosis of transplanted MSCs can inhibit inflammation and hypertrophic scarring [113]. It is increasingly recognized that apoptosis of MSCs is a critical component of their therapeutic efficacy [114]. Direct application of apoptotic bodies derived from MSCs promotes wound healing and M2 macrophage polarization [84]. Additionally, macrophages preconditioned with MSC apoptotic bodies secrete paracrine factors that promote fibroblast migration and proliferation [84].
Toll-like receptors (TLRs) are a key component of the innate immune system that recognizes pathogen-associated molecules. While TLRs are important in the acute phase for the clearance of pathogens, their sustained activity can be maladaptive [115]. Chronic venous leg ulcers have higher levels of TLR-2 and TLR-4 [116]. MSC-EVs can modulate macrophage reactivity to LPS (TLR-4 ligand) by transfer of miR-let7b [94] and miR-181c [101], resulting in attenuated TNF-α and IL-1β production and stimulating the production of antiinflammatory TGF-β and IL-10 [94,101].
Progressive mitochondrial dysfunction is associated with aging and chronic inflammation [117], which contributes to chronic wound formation [118]. An intriguing additional mechanism for promoting M2 polarization is through the transfer of mitochondria [119]. Inflammatory M1 macrophages rely on glycolysis, whereas the anti-inflammatory M2 phenotype is more dependent on mitochondrial oxidative phosphorylation [120]. Additionally, in a murine model of acute oxidative stress, MSC-EVs can reduce ROS-associated skin inflammation in response to ultraviolet irradiation and protect mitochondria from oxidative stress [121].
T-lymphocyte recruitment occurs late in the inflammatory phase. Regulatory Tcells (T regs ) function to limit inflammation, thereby protecting viable cells from immunemediated damage. T regs promote neutrophils secretion of anti-inflammatory molecules and promote neutrophil apoptosis. They also can polarize macrophages towards the M2 phenotype [122]. Amphiregulin is an EGF-like growth factor that can induce the local release of bio-active TGF-β. Tissue resident T regs have been proposed to maintain an environment conducive for proper wound healing through this localized amphiregulin/TGF-β cascade [123]. Tissue resident γδT-cells secrete keratinocyte growth factors and IGF-1 to promote keratinocyte proliferation and survival [124]. Mice deficient in B-cells and T-cells have been shown to have scar-free healing [125]. Furthermore, depletion of T-cells impairs collagen deposition and decreases wound strength [126]. These findings indicate an important role for T-cells in the proliferation and remodeling phases.
Dendritic cells (DCs) are the primary antigen presenting cell of the immune system and are a key link between the innate and adaptive immune responses. MSC-EVs impair DC antigen uptake and expression of co-stimulatory molecules [127]. DC treatment with MSC-EVs reduced the secretion of IL-6 and IL-12p70 inflammatory cytokines, reduced the expression of CCR7 chemokine receptor, and increased secretion of anti-inflammatory TGF-β. These effects were attributed to EV-mediated transfer of miRNAs, in particular miR-21-5p [127]. Through their action on DCs, MSC-EVs are able to attenuate the production of inflammatory T-cells and shift production towards FOXP3 + regulatory T-cells [128,129]. MSC-EVs were also shown to inhibit inflammatory T-cell differentiation, proliferation, activation, and IFN-γ production [130].
The inflammatory response in cutaneous wound healing must remain in homeostasis. The initial burst of inflammation is critical for clearing pathogens and debris. Then the inflammation must resolve to make way for the next phases of the healing process. An excessive inflammatory response will damage surrounding healthy tissues and a prolonged response will delay wound closure. MSC-EVs display promising immunomodulatory effects for promoting an inflammatory environment conducive to effective wound healing.

Proliferation
The proliferative phase involves creating a new foundation upon which the epithelial barrier will rest. In the dermis, this involves angiogenesis, fibroblast proliferation, and provisional ECM deposition to create granulation tissue. The wound environment is metabolically active and requires new blood vessel formation to supply these demands. Failure to supply adequate metabolic nutrients can delay or disrupt the healing process [131]. Additionally, the high glucose environment of diabetes mellitus can inhibit endothelial cell and fibroblast proliferation and promotes their apoptosis [132] (Figure 2b).
MSC-EVs stimulate the expression of repair associated growth factors that promote neoangiogenesis in murine wound models (Table 1). In vitro, MSC-EVs can promote endothelial cell proliferation, migration, tube formation, and secretion of VEGF [81,98,100]. It was demonstrated that MSC-EVs stimulate the AKT/eNOS pathway to promote angiogenesis, in part through the transfer of miR-221-3p [98]. Transfer of miR-31, miR-125a, miR-126, and circRNA mmu_circ_0000250 have also be shown to support endothelial cell proliferation and tube formation [91,96,133,134]. Endothelial progenitor cells cultured in high glucose conditions undergo premature senescence. MSC-EVs can protect endothelial progenitor cells from senescence by inhibiting the expression of inflammatory cytokines and limiting ROS production [95].
MSC-EVs also stimulate fibroblast proliferation, migration, and ECM production in vivo (Table 1). MSC-EVs have been shown to carry EGF, FGF2, Wnt3a, and Wnt4, which can be delivered to dermal fibroblasts, stimulating their migration and collagen synthesis [67,[135][136][137]. Cultured fibroblasts treated with MSC-EVs increase the expression of growth factors (EGF, FGF2, VEGF, PDGF) and ECM molecules (Fibronectin, Collagen 1, Collagen III, Elastin) [81,138]. The function of fibroblasts derived from chronic wounds can be enhanced by treatment with MSC-EVs in a dose-dependent manner, which may be mediated by EV transfer of STAT3 [139]. MSC-EVs can also stimulate AKT and ERK signaling in fibroblasts which have been correlated with enhanced repair functions [79,81].
MSC-EVs may also promote repair through stimulation of tissue resident stem cells, though less is known if this occurs in cutaneous wound healing. MSC-EVs can increase the stemness of human dermal fibroblasts through the transfer of OCT4 and NANOG [140]. BM-MSCs and MSC-EVs undergo an age-related decline in reparative capacities [141]. It was shown that MSCs from aged rats expressed lower levels of pluripotency markers OCT4 and NANOG [142]. Incubation of old MSCs with MSC-EVs from young rats increased expression of OCT4 and NANOG and decreased expression the senescence marker Vinculin [142]. Additionally, it was shown that EVs from young MSCs can delay premature senescence, improve stemness, and stimulate glycolytic metabolism in old MSCs [143]. Finally, MSC-EVs can promote tendon repair by suppressing apoptosis of tendon stem cells [144]. Additional studies will be needed to determine how MSC-EVs influence cutaneous stem cell populations.

Remodeling
The remodeling phase is critical for strengthening the repaired wound. In this phase the provisional ECM is replaced with thicker and more organized collagen bundles, resulting in an increase in tensile strength over a period of months [102]. The wound will also contract, which is mediated by myofibroblasts. If any phase of the healing process is disrupted, atrophic scars, hypertrophic scars, keloids, and chronic wounds can result.
The effect of MSC-EVs on fibroblasts has been reported to either increase or decrease function between studies or within a study at different time points. One potential explanation for how this paradoxical effect may occur is through the generation of regulatory macrophages. Regulatory macrophages are anti-inflammatory and anti-fibrotic, whereas M2 macrophages are pro-fibrotic [146]. While MSC-EVs can enhance the anti-inflammatory phenotype of regulatory polarized macrophages [147], it is unknown if MSC-EVs enhance the anti-fibrotic effects.

Tailoring EVs to Heal Chronic Wounds
Pre-clinical work has demonstrated great promise for the use of MSC-EVs for treating chronic wounds. Numerous studies have found ways to further enhance the wound healing efficacy of EVs, which will be discussed in the following sections. As we learn more about the pathophysiology of chronic wounds, it can be envisioned that MSC-EVs can be personalized to an individual patient based on wound etiology, co-morbidities, and any underlying biological defect in the wound healing process.

Extracellular Vesicles: Source
As previously noted, MSCs are known to be a highly heterogeneous population, and unsurprising, EVs derived from MSCs also show significant variability. EV production is influenced by the source cell, passage number, growth media, atmosphere, culture substrate, and collection conditions. Successful clinical implementation of EVs will also require a means to produce enough EVs. Fortunately, MSCs are one of the most active producers of EVs [148]. EV production can be enhanced by various stimuli, such as hypoxia [149], low pH [150], 3D cell culture [151], acoustic-, electrical-, and mechanicalstimulation [152][153][154][155]. Methods for enhancing intrinsic MSC production of EVs have been reviewed elsewhere [156]. Given the prevalence of chronic wounds, economical largescale production methods will be needed to generate MSC-EVs for this to be a broadly applicable therapy. Standard cell culture vessels are inefficient for large-scale MSC-EV production. Bioreactor systems provide a scalable system for generating large quantities of clinical-grade EVs [157,158].
MSC-EV cargo and downstream effects vary depending on where MSCs are harvested from. With regard to wound healing, Hoang et al. evaluated how MSC source influences EV function. They found that BM-MSC-EVs contained the highest levels of FGF2 and PDGF-BB and displayed the strongest effect on fibroblasts. Whereas, UC-MSCs contained the highest levels of TGF-β and produced the greatest effect on keratinocytes [135]. Comparative analysis of BM-MSC-EV and AD-MSC-EV content revealed that both types are enriched in miRNAs targeting EGF, PI3K/AKT, TGF-β signaling pathways [93]. AD-MSC-EVs are enriched in proangiogenic miRNAs that target HIF-1 and other angiogenic proteins (TGF-β, FGF, PDGFR, TNF, ANGPT1). BM-MSC-EVs contained more abundant proteins linked to integrin and cadherin signaling and metabolic processes [93].
Production of EVs by MSCs is also age-dependent. MSCs from older individuals and late-passage cultures produce more EVs [159,160]. Importantly, these EVs have different cargos and may not produce the desired therapeutic effects [161,162]. Qui et al. showed that adult BM-MSCs pre-treated with neonatal serum EVs have enhanced wound healing potential. Furthermore, these "rejuvenated" BM-MSCs secreted EVs that are superior at promoting wound healing, inducing endothelial cell proliferation, and stimulating AKT/eNOS signaling [85]. Comparison of MSC-EVs from young and aged mice identified enrichment of miR-126 in young MSC-EVs [163]. Overexpression of miR-126 in aged MSCs, results in the production of EVs with potent angiogenic potential, equivalent to EVs from young MSCs [163]. These findings have implications when designing therapies for chronic wounds. MSCs would ideally be harvested from younger donors and MSCs would not be expanded beyond an early number of passages. When this is not feasible, it may be possible to use young MSC-EVs or molecules to "rejuvenate" sub-optimal MSCs to produce EVs with better biologic activity.
Environmental stimuli also influence MSC-EV characteristics. Growing MSCs in a hypoxic atmosphere or the use of hypoxia-mimetic molecules increases EV yield and increases the angiogenic potential of isolated EVs [164][165][166][167][168]. Hypoxia increases VEGF, EGF, FGF, VEGF-R2, VEGF-R3, MCP-2, and MCP-4 in AD-MSC-EVs, which correlates with more robust angiogenic potential [167]. EVs from MSCs treated with dimethyloxaloylglycine stimulate angiogenesis by activating AKT/mTOR signaling [168]. The hypoxia-mimetic deferoxamine when added to BM-MSCs results in the production of EVs with increased wound healing and pro-angiogenic properties [96]. It was shown that in part this was through EV delivery of miR-126 to recipient cells, resulting in PTEN suppression [96].
Inflammation stimulates MSCs to generate immunosuppressive EVs [169]. EVs from MSCs stimulated with TNF-α and IFN-γ promote M2 macrophage polarization, potentially through changes in miRNA content, resulting in IRAK1 inhibition [170]. Additionally, MSCs preconditioned with TNF-α and IFN-γ generate EVs with elevated COX2, leading to the generation of anti-inflammatory PGE 2 [171] . Ti et al. showed, in a diabetic wound model, that EVs from LPS preconditioned MSCs decreased inflammatory cell infiltration into the wound and polarized macrophages towards the M2 phenotype. LPS preconditioned MSC-EVs were enriched with let-7b, miR-1180, miR-183, miR-550b, and miR-133a. Transfer of let-7b to macrophages leads to M2 polarization through inhibition of TLR4/NF-kB and stimulation of STAT3 and AKT signaling [94].
The culture substrate is another modifiable factor when generating tailored MSC-EVs [172]. MSCs grown on a fibrous scaffold or as spheroids enhance their secretion of paracrine mediators that promote wound healing [173,174]. Growing MSCs in 3D culture enhances the secretion of galectin-1, promoting the proliferation and migration of keratinocytes and fibroblasts [175]. The role of EVs in these studies was not specifically addressed, but EVs would have been present in the MSC conditioned media based on the methods reported. A recent study found that 3D culture of UC-MSCs generates EVs that promote fibroblast proliferation and migration [176].
Based on the preceding findings, when MSCs are stimulated by factors found in the chronic wound environment they produce EVs with more potent wound healing potential. When MSCs are exposed to hypoxia, they generate EVs that promote angiogenesis, and when they are exposed to inflammatory molecules, they produce immunomodulatory EVs. These observations are congruent with MSCs, and by extension with MSC-EVs, being critical regulators of tissue homeostasis. It should be explored if a combination of environmental factors can further enhance the bioactivity of MSC-EVs for chronic wound applications.
The cargo of MSC-EVs can also be influenced by targeting MSC receptors. Melatonin promotes MSCs to produce EVs with enhanced anti-inflammatory and wound healing activity [97]. Melatonin MSC-EVs enhance wound closure, Collagen I and III expression, and M2 macrophage polarization compared to untreated MSC-EVs. Melatonin MSC-EVs attenuate inflammation by suppressing AKT signaling [97]. EVs collected from atorvastatin treated MSCs display enhanced angiogenic effects, mediated by miR-221-3p upregulation and AKT/eNOS activation in endothelial cells [98]. It is intriguing to note that MSCs express light-sensing proteins that are typically expressed by retinal photoreceptors. MSCs stimulated with blue (455 nm) light released EVs with more potent angiogenic potential [177]. Blue light stimulation was noted to increase miR-135b and miR-499a packaging into EVs [177].
Multiple techniques exist for isolating EVs including ultracentrifugation (differential, density-gradient, and sucrose cushion), size-exclusion chromatography, immunoaffinity, microfluidics, and others [178]. The advantages and disadvantages of each technique have been reviewed elsewhere [179,180]. For example, Wnt3a is bound to the exterior of BM-MSC-EVs. Traditional ultracentrifugation dislodges Wnt3a, but a combination of polyethylene-glycol enrichment with sucrose cushion ultracentrifugation allows for the recovery of EVs with bound Wnt3a [67]. The type of isolation method employed must consider cost, safety, and the quantity, quality, and biologic-activity of recovered EVs.

Extracellular Vesicles: Engineering
There is tremendous interest in selectively engineering EVs to maximize their delivery of bioactive molecules and to target them to specific cell populations [181][182][183][184][185][186][187]. The surface of EVs can be modified for display of therapeutic molecules or modulate cell targeting. MSCs can be genetically engineered to display peptide sequences, proteins, and antibody fragments fused to the extracellular domain of EV transmembrane proteins. The exterior can be further modified post-isolation by conjugating molecules to surface proteins (e.g., "click" chemistry) and insertion of amphipathic molecules into the lipid bilayer [187]. Modification of the EV surface has been largely unexplored in wound healing research, but it has the potential for substantial therapeutic benefit. For example, it may be possible to insert palmitoylated proteins such as Wnt proteins into isolated EVs.
The most frequently employed method for enriching EV cargo is to overexpress the coding DNA sequence in the EV source cells. This technique has been successfully utilized in wound healing studies. The transcription factor NRF2 provides protection against oxidative stress in diabetic models. EVs derived from NRF2 overexpressing AD-MSCs, compared to standard AD-MSC-EVs, promote faster wound healing in vivo and protect cultured endothelial progenitors from senescence by inhibiting ROS and inflammatory cytokines [95]. MSC-EVs loaded with lncRNA H19 can modulate the miR-152-3p/PTEN axis in fibroblasts grown from diabetic foot ulcers [90]. These H19 loaded MSC-EVs promoted wound healing in a mouse diabetic wound model, suppressed inflammation, and decreased apoptosis [90]. MSC-EVs enriched with TSG-6 showed superior ability to reduce scar formation compared to standard MSC-EVs [83].
Other methods have been developed to target proteins to EVs that are not normally loaded into EVs. The 'exosomes for protein loading via optically reversible protein-protein interactions' (EXPLORs) technique uses a light reversible linker to attach proteins to CD9, an EV associated tetraspanin molecule [188]. Another technique proposed is to capture proteins in self-assembling structures such as 'enveloped protein nanocages' [189].
Additional methods have been proposed to induce EV formation while bypassing active cargo sorting mechanisms, thereby producing EVs with a sampling of all cytoplasmic molecules. EVs generated by these means are also referred to as extracellular vesicle mimetics or cell-engineered nanovesicles [190,191]. Vesicle production can be induced by subjecting cells to hypotonic solution followed by osmotic vesiculation buffer [192]. Cytochalasin B is a pharmacologic agent that disorganizes the actin cytoskeleton. When cytochalasin B treated MSCs are then subjected to shearing stress (vortexing) they produce immunomodulatory and angiogenic EVs [193,194]. EVs can also be generated by extruding cells through 1 µmor 2 µm-pore polymer filters [195], or by ultrasonication [196].
Isolated EVs can be passively loaded with drugs that can pass through the lipid bilayer, whereas other molecules need additional assistance to enter EVs. Active methods such as electroporation, sonification, freeze/thaw, extrusion, saponin, and transfection reagents can allow additional cargos into EVs [183]. Most methods discussed are inefficient at incorporating large molecules into EVs. Engineered lipid nanoparticles can be loaded with high concentrations of therapeutic molecules, but have inferior biocompatibility compared to EVs [197]. Hybrid exosome-liposome vesicles can be generated through co-incubation, freeze-thaw, and sonication [197]. These hybrid vesicles possess the membrane proteins important for EV biodistribution and targeting while incorporating large molecules into the vesicle [198].

Extracellular Vesicles: Quality Control
Rigorous quality control metrics must be established prior to the clinical application of EVs [199]. Each batch of EVs needs to be assessed for its identity, purity, and potency to ensure safety and therapeutic efficacy. Multiple assays will be necessary to fully evaluate EVs given their complex biology. The identity and purity of a batch can be evaluated by measuring the ratio of MSC to non-MSC EV surface antigens and size distribution [199]. The quantity of any specific therapeutic molecules should also be assessed between batches.
One of the challenges with a clinical translation of EVs is optimal dosing. Many studies do not include a dose-response curve to optimize the proper concentration for efficacy. The question remains to be determined if higher dosing results in better/faster healing or if there is a plateau or negative effect from overdosing. MSC-CM and MSC-EVs have been shown to stimulate wound healing responses in a dose-dependent manner, though there is a ceiling to their effect [139,200].
Furthermore, the potency of an EV preparation must be determined to provide a consistent therapeutic dose. Most studies report EV dose as either the number of vesicles or protein content delivered. It would be more appropriate to calculate dose as biologically active "units" based on functional assays. Potency testing for wound healing could involve any combination of in vivo wound healing assays in model species, or in vitro assays to measure their effect on keratinocytes, fibroblasts, endothelial cells, and immune cells. An in depth discussion of functional assays for EVs can be found elsewhere [201].

Extracellular Vesicles: Delivery
Cutaneous wounds provide multiple options for MSC-EV delivery. Most murine studies injected EVs locally near the wound (Table 1). This method is not ideal clinically as it could cause significant pain and distress to the patient. Intravenous injection provides an alternative if the patient has multiple wounds, or a large body surface area is involved. Intravenous (IV) injection of MSC-EVs tagged with iron oxide nanoparticles can be directed to an injury site with a magnet [202]. Hu et al., demonstrated that IV injection of fluorescently tagged MSC-EVs in a mouse wound model showed fluorescence in the wound site on days 5-14 following injury and MSC-EV injection [78]. Their study showed that initially the MSC-EV fluorescence signal was restricted to the spleen on day 1 and then fluorescence accumulated in the injury site days later [78]. When considering the short circulatory half-life of EVs (minutes to hours, see Section 2.2), it is difficult to explain how EVs remain in circulation long enough to correlate with these findings. It may be that EVs rapidly accumulate in the spleen and then are slowly released back into circulation. Alternatively, EVs may act on splenic cells that are then released in response to inflammatory cues [203]. Further work will be needed to evaluate which cells in the wound environment are targets of MSC-EVs. This question could be addressed by identifying which cells in the wound accumulate tracer carried by EVs, though the signal may not reach the limit of detection with this method. An alternative would be to use EVs loaded with molecules capable of inducing stable changes in recipient cells (Cre recombinase or CRISPR/Cas9) [204].
Topical application is appealing because it minimizes patient discomfort, enables a high dose of EVs to be delivered directly to the wound, and allows EVs to be delivered as biomatrices to further enhance wound healing. Topical application of MSC-EVs loaded into carboxymethylcellulose, alginate, or Pluronic F127 hydrogels promote wound healing and neoangiogenesis [89,93,99]. EVs can also be loaded onto hydrogels designed with anti-microbial and adhesive properties suited for the wound environment [89]. Work done in our laboratories has shown that EVs can be stabilized in a collagen scaffold without diminution in efficacy for anti-inflammatory therapies in an osteoarthritis model as well as provide sustained release up to one week in vitro [205](and unpublished observations, D.A.G.). Wang et al., demonstrated that a complex hydrogel system (FHE-Pluronic F127, hyaluronic acid, and poly-ε-L-lysine) could provide pH-responsive sustained EV release and promote skin repair in a diabetic wound healing model [89].

Clinical Perspectives
Stem cell products including EVs are regulated and require FDA approval. Currently, the only stem cell products that are FDA-approved for use in the United States consist of hematopoietic progenitor cells that are derived from umbilical cord blood for use in patients with disorders that affect the production of blood.
There are currently no FDA-approved EV products (https://www.fda.gov/vaccines-bloodbiologics/consumers-biologics/consumer-alert-regenerative-medicine-products-including-stemcells-and-exosomes, accessed on 16 September 2021). At the time of this publication, clinicaltrials.gov (accessed on 16 September 2021) listed three trials using EVs to treat chronic wounds. Two clinical trials will evaluate if serum derived EVs can induce a change in wound size and associated pain (NCT02565264 and NCT04652531).
One clinical trial (NCT04173650) will evaluate MSC-EV dose-limiting toxicity and wound size in dystrophic epidermolysis bullosa. Recessive dystrophic epidermolysis bullosa is an inherited skin fragility disorder, due to mutations in the COL7A1 gene, resulting in defective anchoring of the epidermis to the dermis [206]. Affected children suffer from generalized skin blistering, ulceration, and scarring, for which there is no definitive cure. BM transplant and MSC treatment can increase Collagen VII in the skin [207][208][209]. Work done in our laboratories demonstrated that MSC-EVs are capable of transferring Collagen VII mRNA and protein to fibroblasts [210]. Additionally, EVs may provide an ability to rejuvenate skin cell damage [211].
MSC-EVs can also be considered for an adjuvant role to other modalities. Skin flaps and grafts are part of the clinical toolkit for treating wounds, but flap/graft failure is a major clinic challenge and can prolong the course of a chronic wound. In an in vivo model of flap ischemia-reperfusion injury, EVs increased the rate of flap survival, reduced inflammatory cell infiltrate, and induced neoangiogenesis [164]. Additionally, MSCs were shown to delay the rejection of MHC-mismatched skin grafts in immunocompetent baboons [212]. These findings indicate that MSC-EVs could be an adjuvant therapy when using allografts to reduce immune-mediated graft rejection.
MSC-EVs provide many benefits relative to their parent cell. MSC-EVs are more stable than MSCs. Unlike MSCs, experiments monitoring EV biodistribution have not reported significant pulmonary accumulation. Transplantation of genetically engineered MSCs carries a risk for tumorigenesis and ectopic tissue formation should they become stably incorporated into the host. EVs carry a finite quantity of bioactive molecules; thus, mitigating the risk. A wide array of modifications can be applied to EVs to enhance their intended therapeutic purpose. Limitations to MSC-EV therapeutics include scarcity of the source cell should BM-MSCs be used, limited yield of EVs per production batch, heterogeneity among EVs, and lack of standardized quality control and potency assays. All cell-based therapies have the potential to transmit infectious diseases. While most infectious diseases can be screened for, no approved method for detecting prions has been approved. Human platelet lysate appears to be an alternative to bovine serum in MSC culture, with MSC-EVs showing comparable immunomodulatory effects [213].
The International Society of Extracellular Vesicles has published a position paper outlining important considerations regarding the application of EVs in clinical trials [214]. The clinical application of MSC-EVs in wound healing will require the development of manufacturing strategies compliant with good manufacturing practices (GMP). Additionally, robust quality control and potency testing will be needed to fulfill regulatory requirements.

Conclusions
Pre-clinical data indicate that MSC-EVs can accelerate wound healing by modulating the immune response and by promoting angiogenesis, fibroblast function, and re-epithelialization. There are numerous methods available to modify the cargo of EVs, making them a versatile drug delivery system. MSC-EVs can be delivered intravenously, injected into the wound site, or applied topically to treat chronic wounds. This flexibility in the design and delivery of MSC-EVs opens the doors for creating personalized therapies for chronic wounds.

Conflicts of Interest:
The authors declare no conflict of interest.