Mesenchymal stromal cell‐derived extracellular vesicles for regenerative therapy and immune modulation: Progress and challenges toward clinical application

Abstract Extracellular vesicles (EVs) derived from mesenchymal stromal cells (MSCs) have emerged as a promising form of regenerative therapy and immune modulation. Fundamental advances in our understanding of MSCs and EVs have allowed these fields to merge and create potential cell‐free therapy options that are cell‐based. EVs contain active cargo including proteins, microRNA, and mRNA species that can impact signaling responses in target cells to modify inflammatory and repair responses. Increasing numbers of preclinical studies in animals with various types of injury models have been published that demonstrate the potential impact of MSC‐EV therapy. Although the emergence of registered clinical protocols suggests translation to clinical application has already begun, several barriers to more widespread clinical adoption remain. In this review, we highlight the progress made in MSC‐derived small EV‐based therapy by summarizing aspects pertaining to the starting material for MSC expansion, EV production, and isolation methods, studies from preclinical models that have established a foundation of knowledge to support translation into the patient setting, and potential barriers to overcome on the path to clinical application.


| EXTRACELLULAR VESICLES-DEFINITIONS, CLASSIFICATION, HISTORY
Extracellular vesicles (EVs) have a phospholipid bilayer membrane encapsulating cargo from their cell of origin and are released as paracrine effectors to influence target cells as a form of intercellular communication. Mesenchymal stromal cells (MSCs) give rise to stromal and microenvironmental supporting cells found in many tissues of the body and release small EVs as a means of supporting tissue growth such as hematopoiesis in the bone marrow but also modulate inflammatory responses and facilitate tissue repair across a range of tissue injury and organ dysfunction. Since the initial observations that small EVs could facilitate tissue repair, we have learned that most cells release EVs and that signaling molecules such as microRNA, mRNA, proteins, and other small bioactive compounds are packaged as cargo. 1 A recent classification system has been proposed by the International Society of Extracellular Vesicles that provides a more robust framework for consistent classification than existed previously. 2 Terms such as exosomes and microvesicles imply knowledge of the origins of the particles but are generally identified by size and surface marker expression. As such, the new guidelines advocate that the terms "small EVs" and "large EVs" be used when origin is not certain.
Large EVs (traditionally referred to as microvesicles or microparticles) are often shed from cells under stress following radiation treatment, chemotherapy, ischemic injury, or other stressors. 3 They may arise from membrane blebbing and are characterized by similar outer cell membrane surface markers as their cell of origin with cytoplasmic contents packaged within their cargo. Large EVs can serve as biomarkers of tissue injury or prognostic markers of adverse events but are rarely considered as potential cell-derived therapeutic products.
Indeed, some evidence suggests that large EVs may counteract the beneficial effects of small EVs. 4 Small EVs (often referred to as exosomes), in contrast, may arise within endosomes following a programmed cell-intrinsic pathway that can be altered by external stimuli. 5 Small EVs can be isolated and used as a therapeutic cellderived product for repair of tissue injury or modulation of immune responses and have generated much excitement regarding their potential role in clinical applications. 3 In this review, we highlight the progress made in MSC-derived small EV-based therapy by summarizing aspects pertaining to the starting material for MSC expansion, EV production and isolation methods, studies from preclinical models that have established a foundation of knowledge to support translation into the patient setting, and potential barriers to overcome on the path to clinical application.

| STARTING MATERIAL
MSCs can be readily expanded from bone marrow, adipose tissue, and placental tissues such as Wharton's jelly or umbilical cord blood. The International Society for Cellular Therapy established minimal criteria to distinguish MSCs from other stromal cells or plastic-adhering fibroblast-like cells based on specific cell surface marker expression and retention of multilineage differentiation capacity. 6 Moreover, the use of early passage MSCs for studies of regenerative therapy has proven essential to avoid issues related to cellular senescence of more extensive MSC passaging. 7 The development of serum-free conditions appears encouraging for translation of MSC-based therapies, and numerous clinical trials of MSCs for regenerative therapy and immune modulation have been conducted using approved media and other reagents. 8 Heterogeneity, however, in the source of MSCs, dosing, administration schedules, and outcome reporting continues to hamper definitive conclusions regarding the efficacy of MSCs in certain indications such as graft-versus-host disease (GVHD). 9 The observation that conditioned media from MSCs retains much of the therapeutic effects of MSCs themselves has contributed to the development and interest in MSC-derived small EVs as a therapeutic tool, highlighted by initial studies in acute kidney injury 10 and myocardial ischemia/reperfusion injury. 11 Indeed, we recently updated a systematic search of preclinical trials of MSC-based small EVs in animals with organ injury or immune dysfunction and identified a marked increase in published studies from 17 to 205 between 2013 and 2018 ( Figure 1). 12 Many of these studies used human cells in animal models, and the most common source of cells for the expansion of MSCs is from bone marrow. Most preclinical studies are reporting benefit in one form or another which is encouraging but suggests possible publication bias. A more in-depth analysis and meta-analysis of results from these studies is ongoing.
The role that specific ex vivo culture conditions exert on MSCderived EV production is interesting and may be a major component of future EV manufacturing and production methods. Hypoxia, for instance, can activate signaling pathways in MSCs to enrich EV content of specific signaling molecules that augment their capacity to promote angiogenesis and facilitate tissue repair in animal models. 13,14 Hypoxia can also enhance immune modulation by MSCs through increased secretion of indoleamine-pyrrole 2,3-dioxygenase (IDO) and increased Treg induction. 15 Inflammatory conditions, meanwhile, were associated with increased IDO secretion by MSCs, but Treg induction was unchanged. 15 Other signaling pathways can also be influenced and its ligand, c-kit, inducing greater angiogenic repair activity, 16 while exposure of adipose MSCs to basic fibroblast growth factors attenuated the angiogenic activity. 17 The tissue of origin can also influence the content of MSC-derived EVs. MSCs derived from bone marrow as compared with cord blood, for example, retain secretome signatures that have a greater influence on bone growth and differentiation, as recently summarized by Rosu-Myles et al 18 and others. 19 Adipose tissue-derived MSCs, however, appear to maintain similar immunemodulatory function compared with bone marrow-derived MSCs. 20 Combining strategic choices in the starting material to expand MSCs with selective growth conditions may be the optimal approach to influence the therapeutic effects of EVs. Gene transfection and other approaches that target packaging or enrichment of EV production also remain under active study and appear likely to influence EV-based therapies. MSCs can be easily transduced using lentiviral vectors for overexpressing specific proteins such as angiopoietin 1, 21 insulin growth factor 1, 22 and akt 23 to impact tissue remodeling and organ function, for instance, and other groups have shown enhanced efficacy of MSC-derived EVs by altering their packaged content with proteins such as CXCR4 24 and GATA-binding transcription factor-4, 25 or microRNAs such as mIR133b, 26 miR223, 27 and miR140. 28 One important consideration for cell culture conditions is the potential contamination of bovine EVs from serum in growth media. 29 Hence, the use of serum-free conditions when isolating cell-derived EVs are likely necessary to align manufacturing methods with the ever-increasing stringency of regulatory requirements for accreditation standards.

| Isolating and characterizing EVs
Initial studies of EV isolation used ultracentrifugation (UC) to separate EVs based on their size. Very high forces in excess of 100 000g for prolonged periods are needed to separate the small particles that are <200 nm in size. Unfortunately, UC has been associated with contamination by non-EV cellular material and reduced yield of RNA compared with density gradient isolation methods which could impact therapeutic efficacy. 30 More study is needed to understand the impact of different isolation methods on EV content and treatment efficacy. UC remains the most common approach to EV isolation, but is cumbersome and time consuming and not ideally suited to largescale production. For these reasons, there has been much interest in developing alternatives. Several kits have been developed for different types of fluids using proprietary polymers to combine and precipitate small EVs. Polymer-based methods may be used for very small sample volumes and has been associated with biomarker and profiling studies. 31 The issue of potential contamination with other small particles remains a concern, and impact of polymers on downstream targets that could alter efficacy has been raised in preclinical animal models, which limits enthusiasm. Tangential flow filtration (TFF) has emerged as a relatively easy method of isolating "pure" small EVs that are not complexed with any other molecule and represents a more gentle procedure that appears unlikely to physically alter the EVs. 32 High yields have been reported 33 and the potential to apply TFF for larger scale EV production is under active study. Methods reported to date, however, in published preclinical trials of MSC-EV treatment remain largely based on UC with increasing numbers of studies using TFF in studies published since 2015 ( Figure 2). Registered clinical trials that are actively recruiting patients do not provide sufficient technical information, and we will have to await publication of these results or study protocols to gain more insight regarding the clinical relevance of different isolation methods.

| Rationale for how MSC-derived EVs can modulate immune responses and facilitate organ repair
Initial studies that demonstrated how MSCs could modulate immune responses reported a range of cell-cell contact dependent or independent mechanisms to induce anti-inflammatory signals and immune tolerance states. In brief, secreted factors such as IL6, IL10, prostaglandin E2 (PGE2), transforming growth factor β (TGFβ), hepatocyte growth factor and IDO have been implicated in affecting target cells of the innate and adaptive immune system. 19 Collectively, MSC-derived signaling can inhibit natural killer cell proliferation and activity, reduce B lymphocyte activation, and promote T cells to differentiate into IL17-producing effector T cells. 34,35 At the same time, MSC-associated signaling has been associated with the induction of CD25-expressing FoxP3-positive T regulatory cells. 36   tion of angiogenic supplements such as ischemic brain extract. 42 Assessing the content of EVs will undoubtedly emerge as an important aspect of regulatory oversight related to production methods for MSC-EVs.  We identified a total of six preclinical studies of kidney injury in mice (three studies) or rats (three studies) using human (four studies), mouse (one study) or rat-derived (one study) MSCs to generate EVs that were isolated by UC (five studies) or density gradient centrifugation (one study). 10

| Comparing MSC-derived EVs with MSCs
In our systematic review published in 2015, we identified four studies addressing acute kidney injury (two studies; 10, 53), hind limb ischemia (one study; 61) and tumor growth (one study; 47) that included MSCs as a control. The remaining studies used placebo as controls and/or used EVs derived from other cell types in the control group. In the studies using MSCs as controls, MSC-EVs were at least as effective as MSCs and in one case they were more effective, suggesting much of the therapeutic efficacy associated with MSCs can be ascribed to EVs, however, we and the authors recognize that MSCs retain homing abilities that may be superior to EVs and collectively,

| Clinical studies
There are scant reports of MSC-derived EV administration to patients. In one case report, a patient with steroid-refractory acute GVHD following allogeneic hematopoietic cell transplantation received MSC-derived exosomes. 62 Although a positive response was reported with a reduction in concomitant corticosteroid dosage, the patient died of pneumonia several months after receiving exosome treatment. In a second published report, 56  The International Society of Extracellular Vesicles recently published a position paper that provides a thorough discussion of important considerations regarding regulatory and safety issues for EV-based therapeutics in clinical trials. 5 Aspects of the starting cellular material remain highly relevant, however, increasing numbers of manufacturing facilities that are accredited by the Foundation for the Accreditation of Cellular Therapy produce MSCs that meet FDA and international regulatory standards. The isolation of EVs from MSCs will require greater standardization and efforts to scale up EV production to levels that can support studies in patients. Characterization of EVs will need to be robust, using methods and approaches that can be validated and approved by regulatory bodies. This will likely include a combination of flow cytometrybased methods for membrane surface markers and quantitation of EVs for assessment of purity and yield. If MSC-derived EV products can be used as a third party "off the shelf" product, it may make sense to manufacture and store the product in advance. Storage conditions will need to be optimized and validated to ensure post-thaw potency using assays that reflect the application being considered, whether it is vascular repair or immune modulation. Any intervention to augment or tailor EV content will require added stringency in terms of validation and approval. In addition to the product, early clinical trials should endeavor to demonstrate safety and tolerability. Efficacy studies will need to include appropriate control groups and standard outcome reporting for each disease domain to facilitate pooling of data for meta-analysis and development of evidence networks. It has been previously reported that identification of appropriate patient populations, such as in patients with acute kidney injury, can be a challenge for identifying opportunities for translation of preclinical trials. 63

| SUMMARY
MSC-derived EVs have been studied increasingly in preclinical models of organ injury and immune disorders and appear as a promising cell-free regenerative cell-based therapy for clinical application. More definitive preclinical trials that overcome potential threats to bias by using random allocation of animals to treatment groups and blinding of outcome assessments will accelerate the design of informative initial clinical trials. Translation to clinical studies that meet regulatory approval appears readily feasible. Manufacturing facilities capable of large-scale EV production will be needed in the near future to support this growing field of study.

D.
A. declares consultancy/advisory role with Canadian Blood Services. The remaining authors declare no potential conflict of interest.

AUTHOR CONTRIBUTIONS
All authors contributed to the manuscript and approved the final version.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.