Clinical Translation of Extracellular Vesicles

Extracellular vesicles (EVs) occur in a variety of bodily fluids and have gained recent attraction as natural materials due to their bioactive surfaces, internal cargo, and role in intercellular communication. EVs contain various biomolecules, including surface and cytoplasmic proteins; and nucleic acids that are often representative of the originating cells. EVs can transfer content to other cells, a process that is thought to be important for several biological processes, including immune responses, oncogenesis, and angiogenesis. An increased understanding of the underlying mechanisms of EV biogenesis, composition, and function has led to an exponential increase in preclinical and clinical assessment of EVs for biomedical applications, such as diagnostics and drug delivery. Bacterium‐derived EV vaccines have been in clinical use for decades and a few EV‐based diagnostic assays regulated under Clinical Laboratory Improvement Amendments have been approved for use in single laboratories. Though, EV‐based products are yet to receive widespread clinical approval from national regulatory agencies such as the United States Food and Drug Administration (USFDA) and European Medicine Agency (EMA), many are in late‐stage clinical trials. This perspective sheds light on the unique characteristics of EVs, highlighting current clinical trends, emerging applications, challenges and future perspectives of EVs in clinical use.


Introduction
Initially dismissed as secretory vesicles released by cells for eliminating cellular waste, EVs have gained considerable attention interactions have been observed, including self-homing, which is of particular interest when considering EV-based applications for biomarker and drug delivery to enable selective targeting to cells and tissues. [9]For instance, Qiao et al. isolated EVs from two different cancer cell lines and reintroduced them into mice bearing a subcutaneous tumor specific to one of the cancer cell lines. [10]The results demonstrated that EVs derived from one type of cancer cell preferentially associate with their parent cancer cells in vitro, and upon intravenous administration, demonstrating enhanced homing to their tissues of origin.Upon loading these EVs with Doxil, a doxorubicin-loaded chemotherapeutic liposome, the engineered EVs showed enhanced retention and tumor debulking, and as such, were coined "Trojan horses". [10]This self-homing nature is not limited to cancer cell derived EVs but appears to extend to EVs from healthy cells.Webb et al. demonstrated the tendency of neural stem cell-derived EVs to cross the blood-brain barrier and home to injured brain tissue in murine embolic stroke models. [11]While it has been proposed that EV membrane proteins are key modulators of driving site specific interactions and organotropic behavior of EVs, a recent study by Magoling and colleagues corroborated this hypothesis. [12]In their study, they demonstrated that depleting CD44 on EVs derived from breast cancer cells (which typically overexpress CD44 on their surface) led to significantly reduced tumor delivery of EVs and subsequent reduction in tumor growth, which not only highlighted the tumorigenic potential of cancer-cell derived EVs, but also validated the role of EV surface proteins in modulating organotropism. [12]t is worth noting that EVs injected systemically at high concentrations over a short time are unlikely to mimic biodistribution profiles of endogenous EVs.Recently, a new approach that more closely mimics physiological conditions was reported, involving continuous chronic venous infusion of EVs with a miniosmotic pump. [13,14]Isolating, labeling, and storing EVs may also cause modifications that impact site-specific interactions. [15,16]herefore, in order to exploit the intrinsic targeting ability of EVs, care should be taken to preserve their physicochemical characteristics and endogenous conditions.Nonetheless, the inherent and unique characteristics of EVs points toward a promising future in the clinic. [17]Thus, this perspective aims to provide a brief overview of the characteristics and physiological roles of EVs, while focusing on current developments and challenges in the clinical translation of these unique and naturally derived advanced materials for diagnostic and therapeutic applications.

Physiological Roles of EVs
Despite several reports assessing the composition and function of EVs, there has been limited research that unambiguously elucidates the physiological roles of EVs, owing partly to a lack of standardized methods for EV isolation, characterization, and assessment of functional effects in cell culture and animal models.For example, recent reports demonstrated that EV interactions with recipient cells are substantially affected by the levels of lipoproteins in cell culture medium, [18,19] and multiple studies fail to use conditions that mimic a physiological setting. [19]espite recommendations from the International Society of Extracellular Vesicles (ISEV) published as the Minimal Information for Studies of EVs in 2018 (MISEV2018, Table 1), [20] and other scientific organizations, [4] progress in the field is hindered due to an inability to consolidate findings and make inter-study Parameter Guideline Nomenclature Use of the generic term "extracellular vesicle" along with either size/density or subcellular origin specifications Collection and pre-processing All culture and harvesting conditions should be specified • For conditioned medium, reporting the protocol for EV depletion from the medium, type of vessels, volume of medium, culture conditions, passage number, number of cells/area, cell viability at time of collection, and the frequency of collection • For tissues and biofluids, mass or volume, collection conditions, pre-treatment, and handling temperature/time Storage and recovery (thawing) conditions for the EV source before and after isolation should be outlined EV  Reporting of obtained EV data Submission of method data to EV-specific databases, such as EV-TRACK comparisons.23][24][25]

EVs in Cell-To-Cell Communication
The ability of EVs to modulate cellular functions by acting as vehicles that transfer biological cargo was reported as early as 2006 by two independent groups who demonstrated that spontaneously released microvesicles were highly enriched in messenger RNA (mRNA).However, whether this cargo was functionally active remained to be determined. [26,27]A study by Valadi et al. was one of the first to demonstrate the presence of functional RNA in EVs, exposing EVs as "RNA-carrying shuttles" capable of regulating protein production in recipient cells. [28]Since then, EV mediated intercellular communication has been extensively reported to corroborate the transfer of functional nucleic acids, proteins, carbohydrates, and lipids.Focus has predominantly been on the transfer of microRNAs (miRNAs) by EVs with novel engineering strategies such as use of clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein 9 (Cas9)-based reporting to analyze EV internalization. [29]However, whether the RNA cargo delivered by EVs is functional upon reaching the cells is likely to be highly context-dependent.Once EVs are internalized, the mechanisms involved in unpacking and delivery of cargo to a designated intracellular site are poorly understood.Understanding the physiological consequences of EVmediated cell-to-cell communication would require the use of relevant cell models in a 3D environment.Use of in vivo model organisms, such as zebrafish, to probe mechanisms involved in packaging and targeting EV cargo for intercellular communication provides a promising avenue. [30]Coupling these strategies with novel in vivo visualization and tagging techniques could further improve the understanding of EV mediated cell-to-cell communication.

EVs as Modulators of Immune Responses
A study by Raposo and co-workers was one of the first to elucidate the role of EVs in the modulation of the immune response.The authors observed that the fusion of the late endocytic compartments harboring major histocompatibility complex (MHC) class II molecules with the plasma membrane in B lymphoblastoid cells resulted in the formation of vesicles that were enriched in MHC class II protein. [31]Following this, multiple studies demonstrated the capacity of dendritic cell (DC)-derived EVs to modulate T-cell responses. [32,33][36] Immunomodulatory properties are also evident in the case of T-cell derived EVs.For example, a recent study showed that mitochondrial DNA in T-cell EVs induces anti-viral responses in DCs. [37]Moreover, miRNA-rich cargo of T regulator cell derived-EVs induces the suppression of T cell proliferation which can be advantageous in transplant patients, but disadvantageous in cancer patients where this mechanism promotes immune evasion of tumor cells. [38,39]Omics analysis of antigen presenting cell (APC)-EVs has led to the identification of numerous proteins associated with EV-based immune modulation outlined in the ExoCarta database, a library of EV specific proteins, lipids, and RNA. [40]dministration of EVs derived from humans has not shown major adverse events in experimental animal models.For example, small EVs isolated from human umbilical cord mesenchymal stromal cells (MSCs) when administered intravenously to animal models, including rats, guinea pigs, and rabbits, did not cause adverse events, such as elevated body temperature, hemolysis, anaphylaxis, and impaired muscle, renal or liver function. [41]ther studies also demonstrated the lack of toxicity and immunogenic effects following the administration of human EVs to mice, [42] and human EVs did not change the activation or proliferation of mouse splenocytes in vitro. [14]These studies suggest that EVs released from immune cells can induce immune responses in target cells, however, EVs from other cell types lack immunotoxicity.These observations open the possibility of using EVs as therapeutics or as carriers of therapeutic agents in a wide range of conditions.

EVs in Cargo Delivery
It is established that the unique cargo shuttled by EVs plays an important role in cell signaling and modulation of immune responses. [43]However, the fate of the cargo upon delivery is poorly understood.It is hypothesized that EV interactions with recipient cells can occur through several different mechanisms, which can be predominantly divided into EV endocytosis, EV fusion with the cell membrane, and surface interactions without EV internalization (kiss-and-run mechanism [44] ).In all cases, transmembrane proteins on the EV surfaces may be involved in ligand-receptor interactions with the cell membrane.For example Segura et al. found that EVs derived from mature dendritic cells were enriched in transmembrane proteins such as intercellular adhesion molecule 1 (ICAM-1) and MHC class II, and elicited significantly potent T cell responses. [45]While surface cargo on EVs can affect signaling pathways in recipient cells in the absence of internalization, the activity of internal EV cargo is dependent on direct fusion with the plasma membrane or endocytosis followed by subsequent lysosomal escape.A recent study by Joshi and co-workers used green fluorescent protein (GFP)-CD63-expressing EVs to trace intracellular cargo delivery and found that acidification-dependent fusion of EVs with endosome/lysosome membranes led to the release of EV cargo in the cytosol. [46]Peripheral sites of cargo delivery, such as the endoplasmic reticulum (ER), have also been hypothesized, especially for functional RNA cargo.In a study by Heusermann et al., GFPexpressing CD63 EVs were found to be taken up as single vesicles that accumulated in filopodia-rich regions of the cell, where they, along with endosomes, associated with the ER before being sorted into lysosomes. [47]n improved understanding of the molecular mechanisms involved in EV function at the cellular and sub-cellular levels has highlighted the potential of EVs as next generation of materials for diagnostics and therapeutics for many clinical indications.

EVs as Biomarkers in Disease Diagnosis
EVs are present in all biofluids, including blood, saliva, ascites, urine, breast milk, and semen; and are hypothesized to carry cellular markers from distant, even inaccessible anatomical locations.This property of EVs can be exploited by utilizing them as biomarker candidates for disease diagnosis.Released EVs may mirror the diseased state of the parental cells by bearing an inventory of RNA, proteins, metabolites, and lipids, which can be utilized as biomarkers for diagnosis and disease progression.
In particular, circulating EVs are a valuable non-invasive (for example, saliva and urine) or minimally invasive (for example, blood) strategy to assess various physiological disorders.Transcriptomics, proteomics, glycomics, and lipidomics of EVs has led to the identification of numerous biomarkers with the potential to be used for cardiovascular, autoimmune, and neurodegenerative disorders, and most notably cancer. [7]Although EV-based assays have yet to receive approval from the FDA in the United States, several have been classified as laboratory developed tests (LDTs) that are available for clinical use in individual laboratories regulated under Clinical Laboratory Improvement Amendments (CLIA). [48,49]It is important to highlight that CLIA and FDA regulatory schemes have distinct, but complementary purposes.Specifically, the CLIA program regulates laboratories to ensure accurate and reliable test results of patient specimens.In contrast, The FDA regulates manufacturers and devices to ensure safety and efficacy.
This article includes a brief overview of the prognostic and diagnostic potential of EVs as biomarkers, focusing on current clinical studies assessing EV-based diagnostics.

EVs as Cancer Biomarkers
[52][53][54] A seminal study by Skog et al. proposed the use of EVs as diagnostic tools upon observing that microvesicles released by glioblastoma cells were enriched in RNA and angiogenic proteins characteristic to gliomas. [55]This was the first study to reveal that RNA cargo in tumour-derived vesicles could be isolated from serum samples of patients, thus, paving the way for research into the use of EV cargo derived from biofluids as liquid biopsies for early detection of cancer.Profiling of the RNA content of EVs has been a turning point for early cancer detection with potential RNA biomarker databases constantly being updated by studies evaluating circulating tumor EVs.For instance, a recent four-stage study that aimed to identify potential miRNAs in plasma EVs of colorectal cancer patients, successfully identified a panel of seven miRNAs that were overexpressed in the carcinoma, four of which (miR-17-5p, miR-18a-5p, miR-18b-5p, and miR-181a-5p) were highly enriched in plasma-derived EVs. [56]Examination of miRNA signatures has also been reported to correlate with treatment responses as reported in a study by Kral et al., where overexpression of the miR-17/92 cluster in plasma-derived EVs correlated with adjuvant chemotherapy treatment responses in rectal cancer patients. [57][60][61][62][63] In a recent study by Hoshino and colleagues, proteomic analysis of plasma and tissue explant-derived EVs from cancer patients led to the identification of a panel of cancer-specific EV proteins which could be used as not only potential biomarkers for detection and diagnosis of cancer, but also aid in monitoring treatment response in patients. [60]

EVs as Biomarkers for Other Disorders
Numerous studies have demonstrated the potential of EV miRNA signatures, such as miR-126, [64] miR-30e, miR-92a, [65] miR-192, miR-194, and miR-34a, [66] as biomarkers in early detection of cardiovascular diseases.However, most EVs biomarker studies for the detection of cardiovascular diseases are early-stage feasibility studies comparing patients with advanced disease to healthy controls, which does not reflect the intended use population.In addition to circulating EV miRNAs, the miRNA cargo in urine EVs has been extensively explored in kidney-related disorders.For example, Miller and colleagues demonstrated that urine EVs positive for miR-125a-5p could serve as early predictors of acute kidney injury. [67]Urine-derived EV mRNA is another cargo molecule with promising potential as a diagnostic biomarker.In a cohort of 175 patients, Fekih and colleagues demonstrated the potential of urine EV mRNA signatures as a screening tool for kidney allograft rejection. [68]imilar to the enclosed RNA cargo within EVs, comprehensive analysis of lipid profiles of EV membranes has also led to the identification of disorders based on the differences between the enrichment of lipids in EVs derived from patients with disease versus healthy controls. [69,70]A recent study by Su and colleagues analyzed brain-derived EVs from Alzheimer's disease patients to identify changes in their lipid profiles in Alzheimer's disease. [71]esults from the study showed that these patient-derived EVs had significantly altered lipid profiles compared to controls, thus highlighting the advantage of using EVs as biomarkers for Alzheimer's disease. [71]While evidence from these primary studies points toward a positive outlook for the diagnostic potential of EVs in non-cancer disorders, future studies in larger patient cohorts and intended use populations are necessary for successful translation of EV-based biomarkers.

Current Clinical Landscape of EV Biomarkers
Encouraged by evidence from numerous preclinical studies highlighting the diagnostic potential of EVs, more than 50 clinical trials have been registered (www.clinicaltrials.gov) to evaluate the potential of EVs as biomarkers for disease diagnosis.Although 13 studies have been reported as completed (Table 2), six studies have published results, which are discussed below.
An observational study undertaken in 2017 by Motawi and Colleagues at Cairo University aimed to assess the differential expression of three miRNA signatures, miRNA-136, miRNA-494 and miRNA-495, in EVs isolated from peripheral blood of 100 patients with pre-eclampsia (PE) against a control group of 100 patients with normal pregnancies (NCT03562715).Results from the study demonstrated that these miRNA signatures were overexpressed in PE patients compared to the control group, thus, highlighting the potential of EV cargo as biomarkers for early identification of women at risk to develop PE. [72] A phase 2 interventional clinical trial completed in 2019 used DNA isolated from bronchoalveolar fluid EVs to confirm mutation status for assessment of safety and efficacy of Olmutinib, an epidermal growth factor inhibitor used for the treatment of T790M mutation positive non-small cell lung cancer (NSCLC) (NCT03228277). [73]In the same year, an interventional study by Chanteloup et al. aimed to evaluate the feasibility of measuring EV-associated heat shock protein 70 (HSP70) for early diagnosis of cancer.Although the status of this study is registered as completed on clinicaltrials.gov(NCT02662621), a recent publica-tion indexed to the trial reported that the study is still ongoing. [74]hough the in vivo data look promising, the EVs used in the trial were isolated by ultracentrifugation, which is challenging to implement in a clinical setting due to need for specialized equipment, lengthy and complex protocols, low EV purity, high methodological variability, and limited yields.
[77] The EPI assay, developed at Exosome Diagnostics by McKiernan and colleagues, is a non-invasive urine assay that identifies three EV mRNA biomarker signatures associated with high-grade prostate cancer. [78]Results from clinical trials (NCT02702856, NCT03031418, NCT04720599, and NCT03235687) (Figure 2) demonstrated that the EPI assay could accurately predict the risk of occurrence of prostate cancer and identify patients who do not need an invasive prostate biopsy. [78]In 2016, the EPI test was commercialized and offered for use through the Exosome Diagnostics' CLIA certified laboratory in Cambridge, Massachusetts in the United States. [79]In early 2019, the test was also included in the National Comprehensive Cancer Network's Clinical Practices Guidelines as a recommended test for early detection of prostate cancer. [80]In the same year, this assay was granted a breakthrough device designation by the FDA, expediting its regulatory review process in the United States.Recently, a version of the EPI test (EPI-CE) was approved in Europe under Conformité Européenne (CE) regulation for clinical use. [81]nother EV-based diagnostic assay being developed by miR-Scientific is the miR Sentinel TM PCC4 assay (miR Sentinel TM Prostate Test).This Prostate Cancer test analyses the expression profiles of 442 small non-coding RNAs (sncRNAs) packaged within urinary EVs to classify the patients into four disease statuses as no evidence of prostate cancer, or evidence of low, intermediate, high-risk, or potentially lethal stage of prostate cancer.An earlier version of this prostate cancer test (PCC4 assay), which analyzed expression profiles of 280 sncRNAs within EVs was clinically validated [82] and received a "breakthrough device" designation from the FDA in 2020, enabling an accelerated regulatory review process.However, it is unclear whether the current test has been clinically validated, given the change in the sncRNA signatures being evaluated, which would require its revalidation before commercialization.Nonetheless, this test was made commercially available in 2022 out of laboratories based in United States, Puerto Rico, and several other international markets, [83] and a clinical trial has also been registered in Puerto Rico to evaluate the clinical impact of this Sentinel TM PCC4 assay (NCT04661176).miRScientific is also working towards the development of miR Sentinel BCa test which is a urine-based diagnostic test that analyses the expression profiles of over 280 small non-coding RNAs packaged within EVs, to accurately detect the presence or absence of bladder cancer in patients presenting with haematuria.The diagnostic test is currently being evaluated in an observational study in a large cohort of 3000 patients (male and female) (NCT04155359).
Given that EVs are found in all bodily fluids, other sources of EVs are also being clinically evaluated for their potential for early detection, however, the results from these trials are still pending (Table S1, Supporting Information).Of the 40 currently ongoing clinical trials in this space, nearly half evaluate  Reproduced with permission. [77]Copyright 2018, Elsevier.B) Decision tree and outcome for the EPI test in clinical trial NCT03235687.Reproduced with permission. [76]Copyright 2020, Springer Nature.C) Detection via biopsy as a function of scores obtained from EPI test.Reproduced with permission. [76]opyright 2020, Springer Nature.D) Regulatory approval of EPI test as a LDT under FDA regulation based on its superior sensitivity compared to other detection tests.Reproduced with permission. [78]Copyright 2022, Springer Nature.E) Regulatory approval of EPI test as an IVD under CE regulation based on its superior sensitivity compared to other detection tests.Reproduced with permission. [81]) Copyright 2022, Springer Nature.
the potential of circulating cancer/tumor-derived EVs (TEVs) as liquid biopsies.One of the carcinomas being targeted in these trials is colorectal cancer (NCT04852653, NCT04523389, NCT04394572).For example, the University Hospital of Bordeaux in France is assessing the ability of TEVs to identify poor versus good responders to neoadjuvant chemoradiotherapy by analyzing the protein content and DNA cargo (NCT04852653).A similar study in France is analyzing characteristic EV proteins, such as tetraspanins, integrins, and matrix macromolecules (NCT04394572).
EV biomarkers of lung cancer, especially NSCLC, have also been evaluated in numerous clinical trials.For example, one prospective study focused on the anaplastic lymphoma kinase (ALK) fusion gene, which is highly expressed in malignant NSCLC patients. [84]The trial is monitoring the dynamic modifications of EML 4-ALK fusion in EVs to determine whether this information can be used to predict the efficacy of administered therapeutics in a cohort of 75 NSCLC patients.As a secondary objective, the trial will also assess the sensitivity of EVbased diagnostics compared to FDA-approved assays for identifying ALK fusion (NCT04499794).Another interventional study, which is currently recruiting, is attempting to predict the efficacy of immunotherapy (pabolizumab and nafulizumab cocktail) for NSCLC patients by monitoring the difference in the expression profiles of programmed death-ligand 1 (PD L-1) and miRNA in EVs (NCT04427475).Additionally, a pilot study by a group in China is currently evaluating whether RNA profiles from circulating EVs can be used for non-invasive detection of lung metastases in high-grade osteosarcoma (OS) in a 40-patient cohort (NCT03108677).The study was designed based on evidence from RNA profiling of circulating EVs, which suggested that metastatic EVs had a significantly higher mutation burden compared to non-metastatic ones. [85]

Challenges Associated with Clinical Translation of EV-Based Diagnostics
Although several CLIA-certified EV-based diagnostic assays, classified as LDTs, are in clinical use, none have been approved by the FDA.These commercialized LDTs are helping numerous patients, however, their widespread use is limited, as they are operating in individually approved laboratories.
Additionally, although a plethora of RNA and protein signatures have been identified (and cached in the ExoCarta database), many of them lack rigorous validation.Traditional EV isolation methods, such as ultracentrifugation, often yield pellets that are enriched in protein/lipoprotein contaminants, which accentuates the complexity of EV-based biomarker validation leading to poor performance and high analytical variability.Additionally, EV quantification methods in clinical trials are poorly standardized, which hinders inter-study comparisons.Some studies report particle counts, while others quantify proteins or report the particle/protein ratio. [86]astly, EV based diagnostics cannot be generalized in a "one size fits all" manner.EV release and cargo is hypothesized to vary from patient to patient due to influences from a multitude of factors such as genetics, lifestyle, sex, and age.For example, studies by Eitan et al. demonstrated that EV release decreases, while EV internalization by B cells increases with aging. [87]This discovery points toward opportunities and challenges for developing personalized EV-based diagnostics.(Figure 3)

EVs as Advanced Materials for Therapeutics
An enduring challenge in the development of EV-based therapeutics is an incomplete understanding of the intricate mechanisms involved in the interactions between EVs and the biological environment.The assembly and packaging of EVs vary according to cell type, cell condition, and environmental stimuli.These varying factors, make standardization onerous, but not impossible as evidenced by the success of EV-based vaccines for meningococcal disease such as VA-MENGOC-BC ® (licensed in Cuba) and Bexsero ® (approved by the European Medicines Agency/EMA), which consist of bacterium-derived outer membrane vesicles and recombinant proteins.However, therapeutic use of eukaryotic EVs has yet to achieve cementing clinical evidence to warrant translation into the clinic.
Assessment of the clinical application of both unmodified native EVs and therapeutic cargo loaded EVs has been gaining momentum.The first phase 1 clinical studies using EVs as therapeutic agents were reported in 2005, where the safety and efficacy of administering autologous DC-derived EVs were assessed. [88,89]n one of these studies, Escudier and colleagues used autologous DC-derived EVs pulsed with melanoma antigen 3 (MAGE3) peptides to assess safety and feasibility in patients with melanoma.The results from their study showed that the EVs lacked toxicity, however, the trial failed to achieve the desired therapeutic outcomes. [88]The second phase 1 clinical trial assessed the safety and efficacy of autologous DC-derived EVs loaded with MAGE tumor antigens in NSCLC, and the outcomes of the study demonstrated patient tolerance to the EVs.Moreover, this study reported an uncharacteristic increase in natural killer cell activity in nearly half of the patients.This observation led to the hypothesis that the EV therapy may stimulate both the innate and adaptive arms of the immune system, leading to enhanced anti-tumor effects in carcinomas. [89]Although both of these studies highlighted the biocompatibility of EVs, they failed to broadly demonstrate therapeutic potential.Inspired by the safety profile of EVs in these trials, Dai and colleagues evaluated the anti-tumor immunity of ascites-derived EVs in conjunction with granulocyte-macrophage colony stimulating factor (GM-CSF) as an adjuvant in patients with colorectal cancer. [90]The study illustrated the tolerability of ascites-derived EVs and the efficacy of GM-CSF as an adjuvant for the induction of anti-tumor responses.However, a lack of understanding of the precise cellular origin of the ascites-derived EVs was considered a drawback, as well as the lack of efficacy in the EV-alone treatment group. [90]Clinical studies in the following years probed the potential of ultrafiltration for isolation of mature DC-derived EVs with enhanced immunomodulatory capacities and assessed their use as a maintenance immunotherapy by administering them in patients with NSCLC in the first phase 2 clinical trial using therapeutic EVs (NCT01159288).However, the administered EVs did not induce any detectable immune responses. [91]here has been considerable interest in the benefits of MSCs and their use as therapeutics following mounting evidence of their efficacy in numerous preclinical studies and clinical trials. [92][95] Characterization of the RNA content and the surface molecules of MSC-derived EVs by Bruno et al. concluded that the transfer of the RNA content to target cells contributed to the regenerative potential of MSCs. [94]This conclusion was further supported by a study by Zhang et al. that showed that MSC-derived EVs were immunologically active and contributed to indirect regulation of T cell induction, while simultaneously inducing macrophage polarization, thus activating an immune response. [96]This discovery gave way to a single patient treatment experiment where Kordelas et al. administered MSCderived EVs intravenously to a graft-versus-host disease (GvHD) patient in increasing doses for two weeks.The study demonstrated positive outcomes where the patient displayed a substantial reduction in GvHD-associated symptoms and remained stable for as long as four months after treatment. [97]Most recently, Warnecke et al. evaluated the feasibility of first-in-human intracochlear application of umbilical cord-derived MSC-EVs in a single patient treatment study. [98]Although the patient showed no signs of adverse effects post EV-administration over a five-day observation period, results from a single patient study fail to provide robust evidence of the safety profile.
Apart from the evaluation of therapeutic capacity of native EVs, EVs loaded with exogenous cargo are also being evaluated as drug delivery systems.In terms of drug delivery, EVs have potential to outperform synthetic nanoparticles, such as liposomes, due to inherent targeting properties. [2,99]Liposomes are structurally comparable to EVs in terms of having a lipid bilayer surrounding an aqueous core.Clinically approved liposomes have been on the market since 1995 with the introduction of FDA-approved Doxil ® . [100]Polyethylene glycol (PEG)ylated lipid nanoparticles containing ionizable cationic lipids in the core have also been used in the recently developed coronavirus disease of 2019 (COVID-19) vaccines to deliver messenger RNA (mRNA). [101]However, all clinically approved nanoparticles are structurally simple, usually consisting of around four to five different components. [2]Although synthetic drug delivery systems that incorporate multifunctional and multistage components have been assessed in preclinical studies, [102][103][104][105] clinical translation of complex drug carriers, including inorganic nanoparticles such as silica, metal organic frameworks (MOF), and porous silicon, has failed either due to manufacturing challenges and/or unmet expectations of benefits from added complexity. [48,106]Additionally, the complex interactions that occur between biological surfaces and these nanomaterials are not fully understood, making it challenging to exploit synthetic drug delivery systems for site-specific delivery.EVs are biologically complex materials carrying a repository of bioactive and structural molecules.][109] Therefore, screening methods to identify ideal EV subpopulations for targeted delivery could circumvent the need for in-depth understanding of biological surface interactions and subsequent rational design of synthetic nanoparticles, and instead exploit evolutionary optimization of EV surfaces for intercellular communication over short and long distances.
[112] While PEGylation is a preferred method for overcoming this hurdle, the extent of its effectiveness and the possibility of increased toxicity bring into question the feasibility of this approach. [113,114][117] Although EVs face a similar fate upon intravenous administration, increasing evidence points toward the existence of distinct EV subpopulations that express surface proteins such as CD47 which act as "don't eat me" signals. [118,119]herefore, native EVs within these identified subsets could be used for enhanced physiological retention.In comparison to synthetic nanoparticles, EVs have also been hypothesized to have a predominant advantage in terms of biocompatibility.In fact, early-stage clinical studies evaluating the safety of EV-based therapeutics have reported a lack of adverse events in patient cohorts of the targeted disease. [120]Nonetheless, these characteristics of EVs make them a promising alternative to synthetic nanoparticles from a drug delivery perspective.Of particular interest, is the loading of anti-tumor drugs such as doxorubicin and paclitaxel, [121,122] anti-inflammatory agents such as curcumin, [123,124] and RNA therapeutics such as small interfering RNA (siRNA) and miRNA [125][126][127] within EVs.[130] Endogenous loading methods utilize cell engineering approaches to exploit the natural cellular machinery for loading of therapeutic proteins or RNA. [131,132]uch techniques are promising compared to exogenous loading methods such as electroporation, where loading efficiency is low.Studies by Kamerkar et al. assessed the efficacy of electroporated EVs termed "iExosomes" loaded with siRNA targeting the oncogenic Kras (Kirsten rat sarcoma) in a pancreatic ductal adenocarcinoma mouse models. [127,133]These iExosomes were shown to evade phagocytic clearance, reduce tumor burden and improve survival.A head-to head comparison of similarly engineered liposomes (iLiposomes) demonstrated the superiority of iExosomes, which was attributed to the characteristic CD47 integrin-associated transmembrane protein. [127]However, it is worth noting that electroporation can result in the formation of nanosized siRNA aggregates, which makes it challenging to demonstrate that delivery is EV-mediated. [134]Favorable outcomes from this study led to the initiation of a pilot study (NCT03608631) primarily investigating the maximum tolerated dose and safety of iExosomes with the secondary objective of analyzing their pharmacokinetics, however the trial is yet to report results.
Chemotherapeutic drug-loaded EVs are also being evaluated, including the encapsulation of paclitaxel in the EV lipid bilayer, which has been shown to suppress pulmonary metastases in mouse models by demonstrating superior colocalization with metastatic lung lesions compared to free Taxol. [121] recent study by Patel et al. demonstrated that cannabidiolloaded EVs significantly sensitized triple negative breast cancer cells to doxorubicin (DOX) treatment, and a combination therapy of cannabidiol and DOX remarkably reduced tumor burden in mouse models. [135]A phase 2 clinical trial investigating the efficacy of DOX-loaded tumor-derived EVs targeting malignant ascites was registered in 2013, however, the status and outcome of this trial remain unknown (NCT01854866).In addition to EV-based delivery of chemotherapeutics, biomolecules such as curcumin, a natural anti-inflammatory agent with limited bioavailability, are also being evaluated in clinical settings for disorders such as inflammatory bowel disease.However, the fate of EVs is also highly dependent on the administration route, and additional research is needed to evaluate the biodistribution and toxicity of drug loaded EVs delivered via different routes of administration

Current Clinical Landscape of EV-Based Therapeutics
Early clinical evidence of the therapeutic potential of EVs has led to a recent increase in clinical assessment of EV-based therapies, with over 40 clinical trials registered (www.clinicaltrials.gov)for disorders such as Crohn's disease, ulcerative colitis, type-I diabetes, and COVID-19.Of these, nine trials have been registered as completed, from which studies with published results have been discussed below.(Table 3).
Promising results from preclinical efficacy and Phase I clinical safety evaluation (NCT04313647) of MSC-derived EVs by Shi and colleagues in a study has led to the use of a novel approach for the administration of these MSC-derived EVs as therapeutics for COVID-19. [136](Figure 4) Based on the results of this study, Zhu and colleagues investigated safety and efficacy of aerosol inhalation of human adipose tissue-derived MSC EVs on a small cohort of seven patients in the recently completed MEXCOVID trial (NCT04276987) [137] In this interventional trial, upon five-day consecutive daily administration of the EVs, there was no occurrence of adverse events among patients, revealing that the treatment was well tolerated and lacked dose-relevant toxicity.However, further trials are of prime importance to evaluate the safety profile in larger patient populations.Additionally, a major drawback of this study was the simultaneous administration of antiviral agents along with nebulization with EVs, which makes it difficult to pinpoint the therapy responsible for the favorable outcomes in patients. [137]A similar interventional trial in Russia (NCT04491240) also investigated the safety and tolerance of a nasally administered aerosol formulation of MSC-derived EVs in a larger study group of 30 COVID-19 patients.The clinical outcome of this study has also reported no occurrence of adverse events.
MSC-EVs are also being assessed as intravenously administered therapeutics in other clinical trials for a range of diseases, such as acute respiratory distress syndrome (ARDS) (NCT05127122), Crohn's disease and irritable bowel syndrome (NCT05130983), ulcerative colitis (NCT05176366), chronic post-COVID-19 syndrome (NCT05116761) and the treatment of organ rejection (NCT05215288).However, the results from these trials are still pending.
Compared to the plethora of clinical studies using MSCderived EVs, usage of plasma-derived EVs as therapeutic agents have been limited, despite several preclinical studies, [138] Table S2, Supporting Information).Only three clinical trials using plasma/serum-derived EVs as therapeutics have been registered to date, with two of the studies estimated to be completed in 2022 (NCT04652531, NCT04849429 targeting ulcers and degenerative disease, respectively), whereas the status of one of the trials is listed as unknown (NCT02565264, targeting ulcers).The lack of clinical trials employing plasma as a source material could be due to the limited therapeutic effects of this heterogenous population of EVs and challenges in isolating a pure sample without co-isolation of lipoprotein contaminants.It is worth noting that sample purity does not necessarily correlate with improved therapeutic effects.On the contrary, EVs and contaminants could have additive or synergistic therapeutic activity.For example, some lipoproteins have potential as anti-inflammatory therapeutics and drug delivery systems. [139]From a regulatory standpoint, the presence of lipoprotein contaminants is not a hurdle per se, if quality assurance and batch-batch consistency can be demonstrated.Similarly, despite the significant preclinical interest in chemotherapeutic cargo loaded EVs, clinical evidence has been limited as evidenced by the low number of clinical trials investigating cargo loaded EVs as therapeutics, none of which have been completed till date.

Challenges Associated with Clinical Translation of EV-Based Therapeutics
EV-based therapeutics have the potential to exhibit higher biocompatibility, lower immunogenicity, and easier handling than cell therapy. [8,140]Current clinical trials assessing the therapeutic potential of EVs primarily encompass MSC-derived EVs.Other EV-producer cell types, such as HEK293 cells, are also promising for the design of EV-based therapeutics. [141]However, mechanistic understanding of the therapeutic effects of EVs is incomplete.Despite the increasing number of research groups attempting to understand the complex physiological roles of EVs, a lack of reliable methods to track and probe complex EV interactions in physiological settings further hinders mechanistic understanding.Methods, such as radioactive labeling of EVs to assess in vivo trafficking, are promising, but require specialized resources, and like most labeling probes, can affect the native properties of EVs. [142]n addition to a lack of physicochemical and mechanistic understanding, there are several challenges in the clinical translation of EV-based therapeutics.In particular, the heterogeneity  (Reproduced with permission under the terms of the CC-BY license. [136]Copyright 2021, the Authors.Published by Wiley Periodicals LLC. and biological complexity of EVs makes scalable manufacturing challenging.Additionally, standardization of isolation and characterization techniques (use of electron microscopy, mass spectrometry, FRET, etc) and identification of EV subtypes are major hurdles to be overcome.Commonly used EV isolation methods such as differential ultracentrifugation fall short in terms of maintaining batch-to-batch consistency, which is one of the key factors for the transition of EVs from bench to clinic.Other EV isolation methods, such as tangential flow filtration (TFF), outperform ultracentrifugation in terms of purity and provide more reproducible EV batches. [143]However, some lipoprotein contaminants co-isolate with this method. [144]Consecutive use of two or more EV isolation techniques significantly removes these contaminants, [145] but at the cost of EV yield.[148] Appropriate storage, handling, and transportation conditions for EV-based therapeutics should be established, and these conditions need to be outlined in regulatory application documents.There is no consensus yet on the suitable storage conditions that would preserve the structure and function of EVs.Recent systematic evaluation of the effect of different storage conditions on the integrity of EVs by Gelibter et al. have shown that freezing EVs at −80 °C results in time-dependent increase in EV size, while repeated freeze-thawing of EV samples led to significant reduction in their numbers over time. [149]Using lyophilization for more manageable storage and transportation of EV-based therapeutics could help mitigate this challenge and also extend the shelf-life of finished products.A recent study by Popowski and colleagues investigating the therapeutic potential of lung-derived EVs loaded with severe acute respiratory coronavirus 2 (SARS-CoV-2) spike protein encoding mRNA reported that lyophilized formulations of these therapeutic loaded EVs were stable at room temperature, retained the functionality of the mRNA cargo and were superior in therapeutic efficacy compared to their liposomal counterparts. [150]While results from this study are encouraging, further exploration of effects oflyophilizationn on the integrity of the EV membrane is critical.Clinical trials to date, have predominantly reported a lack of adverse events following EV administration; however, toxicity may occur in future studies due to the use of different EV sources, isolation/modification methods, dosage, and patient populations.For example, many publications focus on the clinical potential of cancer cell-derived EVs for drug delivery, however, there is a lack of consensus on whether cancer cell-derived EVs are suitable for clinical applications due to safety concerns relating to their potential to disseminate disease.Despite several ongoing clinical trials in the therapeutic space, the first EV product has yet to receive clinical approval from the FDA.In fact, due to the unapproved usage of EV-marketed products, the FDA released a public safety notice reporting multiple occurrences of serious adverse events in patients who were administered products falsely marketed as "exosome products". [151]Occurrences like these affect the confidence of the general public in EV-based products, further hindering clinical translation.
Development of EV-based therapeutics could benefit from taking into consideration the lessons learnt from the approval of cell-based therapies, which face similar hurdles in clinical translation such as off-target effects of the therapy, potential toxicity, and manufacturing challenges including scale-up, storage, and transportation.For instance, the risk of tumorigenicity involved when using cell-based therapies is assessed by employing extensive screening methods during pre-clinical testing, and a similar approach could be beneficial to therapies using tumor-derived EVs. [152]Although EVs pose a much lower risk of tumorigenicity as they cannot replicate, they do possess the capacity to transfer oncogenic cargo and promote metastasis through mechanisms that are still being understood.Where cell based therapies such as Provenge ® failed in terms of controlling the cost of therapy of a patient-specific dose due to not taking into consideration the scale-up cost during the development phase, [153] EV therapeutics could avoid the trap by using a bottom-up strategy to determine a commercially viable price-point and designing a scale-up strategy to achieve the same.Designing a scale-up strategy early-on could also help to predict and overcome large-scale production constraints, which are a bottleneck even for recently approved CAR-T therapies such as Breyanzi (approved by the FDA in February 2021) and Abecma (approved by the FDA in March 2021); for which the demand vastly exceeds the supply even now, causing a failure in both; patient reach and capitalization despite successful clinical translation and commercialization. [154]An interesting approach could also be to combine the best of both worlds by the development of EV therapies driven by cell therapies; as explored in a recent study, [155] as cell-based therapies essentially contain cells, which when introduced into the body are likely to be actively releasing EVs which may also be contributing to the actual therapeutic response generated by the cell-based therapy. [156]

Outlook and Future Perspectives
EVs have captured the attention of the scientific community for over five decades due to their promise in disease diagnosis and therapy, but the true potential of EVs as emerging material is now evident with over 100 EV-based products currently in clinical trials.Given that the clinical translation of EVs is still in progress, there are a number of areas in the EV field that could be improved to ensure successful and timely translation into the clinic.Given the limitations of current techniques for EV tracking and imaging within cells and tissues, the development of reliable EV tracers would advance understanding of extracellular and intracellular trafficking, which is integral to gaining mechanistic insights.The use of recently developed single particle analysis techniques such as Exoview [157] and total internal reflection fluorescence microscopy [158] could aid in characterizing EV subpopulations rapidly and more accurately.
Given that the heterogeneity of EVs is a crucial challenge for regulatory approval, nanoscale methods such as nanoflow cytometry, [159] imaging flow cytometry, [160] and affinity-based methods [161] to characterize and separate EV subpopulations could improve therapeutic and diagnostic capacity and reduce heterogeneity.However, these techniques require further technical optimization in terms of increasing EV yields.
Scalable manufacturing and batch-to-batch consistency of EVs are ongoing challenges within the field.Adopting cell culture and EV isolation methods that are suitable for large-scale applications, such as bioreactors and TFF, could overcome some of the challenges associated with scalable manufacturing.Recent studies have shown that EVs derived from cells cultured in bioreactors were highly consistent in terms of size concentration and surface glycan profiles. [162,163]Inspiration from mature orthogonal techniques that are already in use for therapeutics, such as liposomes and monoclonal antibodies could also benefit the EV field.
The lack of standardization and transparency in reporting of EV-related literature also needs to be addressed.Transparency in reporting has improved with the introduction of databases, such as EV-TRACK, EVpedia, and ExoCarta, which serve as repositories for EV-related research, and greatly accelerate the identification of potential biomarkers suitable for development of EVbased diagnostics.Ensuring compliance with ISEV recommendations, such as MISEV2018, would greatly aid in promoting standardization across the field while simultaneously educating the next generation of researchers.
The EV field can draw inspiration from the recent development and regulatory approval of the mRNA-1273 (Moderna) [164] and BNT162b1 (Pfizer) COVID19 vaccines, [165] which contain mRNA.These vaccines can serve as a potential reference frame for the development of mRNA-loaded EVs.However, from a drug delivery perspective, regulatory requirements for the approval of EVs are different from those of synthetic nanoparticles, as EVs are biological products.Hybridization of EVs with synthetic nanoparticles (organic and inorganic) could be another potential pathway for better clinical outcomes for EVs as theranostic and drug delivery carriers.The development of better scalable hybridization or coating techniques such as those based on microfluidics will pave the way for future EV research and clinical success.Employing emerging technologies such as artificial intelligence (AI) and machine learning in the design and engineering of EV based therapeutics could be an approach to advance the research in the field.[169] Use of above mentioned high throughput techniques such as microfluidics in combination with machine learning based EV analysis could accelerate the clinical translation of EV-based therapeutics.
Nonetheless, as with other clinical research, translation of EVbased products into the clinic would also require them to traverse and survive the proverbial "valley of death"; an intermediate phase between research of a product in academic environments and its successful commercialization; which is often adorned by roadblocks such as lack of funding to commercialize the product and market competitiveness, especially in the development of therapeutic products.Often times, researchers at academic institutions, where a majority of the research into product development occurs, are not equipped with the expertise and resources required for successful clinical translation. [170]Companies within the sector that are interested in commercializing the product are also under tremendous pressure to be profitable, leading to careful selection of products based on their market competitiveness; and sometimes shelving of new products that are perceived as competitors of existing products in the market.There is a high risk of failure involved while attempting commercialization of products, which often discourages academics from pursuing clinical translation.Introducing measures such as promoting academic entrepreneurship in universities by encouraging collaboration between research laboratories and business schools within academia; allocating funds for academics wishing to pursue commercialization and transparency within the corporate sector by adopting an open innovation model [171] could help bridge this translational gap.
In summary, EV research has substantially grown over the past decades, demonstrating the potential of EVs in diagnosis and therapy, as evidenced by numerous clinical trials that are underway.The development of EV-based diagnostic assays has the potential to enable timely detection of a broad range of diseases.Progress in the use of EVs as therapeutics and drug delivery systems is likely to substantially impact health care in the upcoming decades.

Figure 2 .
Figure 2.Validation, clinical evaluation, and regulatory approval of the EPI test for prostate cancer.A) Superior sensitivity and specificity of EPI test compared to other detection tests for prostate cancer in both previous and current validation cohorts as evidenced by the area under the curve.Reproduced with permission.[77]Copyright 2018, Elsevier.B) Decision tree and outcome for the EPI test in clinical trial NCT03235687.Reproduced with permission.[76]Copyright 2020, Springer Nature.C) Detection via biopsy as a function of scores obtained from EPI test.Reproduced with permission.[76]Copyright 2020, Springer Nature.D) Regulatory approval of EPI test as a LDT under FDA regulation based on its superior sensitivity compared to other detection tests.Reproduced with permission.[78]Copyright 2022, Springer Nature.E) Regulatory approval of EPI test as an IVD under CE regulation based on its superior sensitivity compared to other detection tests.Reproduced with permission.[81] ) Copyright 2022, Springer Nature.

Figure 3 .
Figure 3. Extracellular vesicles: pathway from discovery to clinical translation as biomarker and/or therapeutics.

Figure 4 .
Figure 4. Preclinical efficacy and safety evaluation of aerosol administration of human adipose-tissue MSC-derived EVs (haMSC-EVs) A) Scheme for the manufacture and characterization of haMSC-EVs B) Biodistribution of haMSC-derived EVs over 28 days after aerosol administration C) Post infection with Psuedomnas aeruginosa to induce murine lung injury, aerosol administration of EVs showed lower influx of white blood cells and neutrophils, as well as inflammatory cytokines IL-6, IL-10, and TNF- in the bronchialveolar lavage fluid D) Improved survival rate in mice treated with haMSC-EVs compared to control groups E) Reduction in histological severity of lung injury in murine models post nebulization with haMSC-EVs (F) Presence of lower inflammatory cells in the alveolar space and lungs of murine models post nebulization.(Reproduced with permission under the terms of the CC-BY license.[136]Copyright 2021, the Authors.Published by Wiley Periodicals LLC.

Table 1 .
Summary of MISEV guidelines established in 2018.
Quantification by particle, protein, and/or lipid amount and/or volume of initial fluid • Ratio of two quantification parameters to assess purity • Characterization via evaluation of at least one protein from the three categories: • Transmembrane or glycosylphosphatidylinositol (GPI) anchored proteins associated with the plasma membrane or endosomes • Cytosolic proteins • Contaminant proteins Single EV characterization can be achieved by imaging single EVs (such as, cryogenic electron microscopy) and non-imaging methods analyzing large populations of single EVs (such as nanoparticle tracking analysis) Functional studies of EVsReporting of dose response assessment, including quantitative comparison of functional activity of medium before and after EV isolation, non-EV fractions, and comparison to EV-enriched fractions.

Table 2 .
Current clinical trials of EVs for diagnostic applications.

Table 3 .
Current clinical trials for EV therapeutics.