Therapeutic Potential of Multipotent Mesenchymal Stromal Cells and Their Extracellular Vesicles

Introduction

Mesenchymal stromal cells (MSCs) have attracted increased attention because of their considerable therapeutic potential. Their plasticity that gives rise to a spectrum of tissues of mesoderm origin, ¹ together with their immunomodulatory properties ^{2 –5} and low immunogenicity, makes them ideal candidates for different cellular therapies. The therapeutic value of MSCs is evident by the number of new and ongoing trials registered on clinicaltrials.gov targeting a variety of conditions (Table 1). As we become more reliant on the potential of these cells in the clinic, conclusions from ongoing randomized trials are needed in order to establish the true effectiveness of MSC therapy. In bone and cartilage repair, MSCs are expected to replace the damaged tissue, while in other therapies they modulate a response by the secretion of bioactive molecules. While the first reports describing the secretory functions of MSCs were published nearly 20 years ago, ⁶ the discovery that many of these effects can be attributed to extracellular vesicles (EVs) is relatively recent. Le Blanc and colleagues have contributed to pioneering studies of using bone marrow (BM)-MSCs to treat hematopoietic stem cell transplantation (HSCT)-related complications, ⁷ including graft-versus-host disease (GvHD) ⁸ and severe tissue damage, ⁹ and in cotransplantation with HSCs as prevention of complications and promotion of engraftment. ¹⁰ In addition to HSCT-related complications, Le Blanc et al. have successfully used MSC treatment in other diseases with an inflammatory component such as acute respiratory distress syndrome (ARDS) ¹¹ and newly diagnosed diabetes. ¹²

Sources of MSCs for Clinical Applications

In situ, MSCs are progenitor cells residing in stroma, where they are thought to participate in tissue wound healing and to modulate immune cell function. ⁵ Endogenous MSCs reside in close contact with the basement membrane and surround endothelial cells at perivascular locations throughout the vascular network of the body. In these locations, they orchestrate cellular interactions in order to control immune responses and tissue repair. ¹³ Because of their immuno- and trophic-regulatory capacity, these cells recently merited the description “site-regulated drugstores.” ¹⁴ BM is currently the main source of MSCs for clinical use (Table 1), but the cells can be isolated from many other adult and embryonic tissues, including adipose tissue, ¹⁵ amniotic fluid, ¹⁶ skin, ¹⁷ dental pulp, ¹⁸ umbilical cord, ¹⁹ and cord blood ²⁰ and menstrual blood. ²¹ MSCs derived from different tissues have similar but not identical functional potential, ^22,23 which needs to be considered when selecting the source of cells for therapy. Regardless of tissue source, the use of MSCs in clinical applications requires isolation and in vitro cell expansion, which raises challenges in order to produce cells with an optimal therapeutic efficacy. ²⁴ The generally accepted minimal criteria to define human MSCs being isolated from various tissues using different isolation and expansion methods were formed by the International Society for Cellular Therapy in 2006, as a response to inconsistencies in the field. ²⁵ These criteria include plastic adherence capacity together with trilineage multipotency (adipocyte, osteoblast, and chondrocyte) and a phenotypic definition that requires expression of CD73, CD90, and CD105 together with the lack of expression of hematopoietic cell surface markers CD45, CD34, CD14, CD11b, CD79α, CD19, and HLA-DR.

Cell Fate Following Systemic Administration

The major drawback in MSC therapy is the incomplete understanding of cell fate following systemic administration as well as the mechanisms by which these cells correct disease. ^{26 –30} Contradictory results in clinical trials using MSCs to treat steroid-resistant GvHD suggest that certain variables in the MSC production are critical in maintaining the multifunctional nature and thereby the therapeutic effect of the cells. ³⁰ Phase II clinical trials performed in Europe using BM-MSCs have reported a beneficial outcome, ^{8,31 –33} while a phase III trial performed in the United States using an industrially produced MSC source demonstrated no beneficial effects. ³⁰ Studies by Le Blanc and associates indicate that the use of very early passage MSCs mediates a better outcome than late passage cells, though the mechanism is unclear. ^34,35 Endogenous BM-MSCs can be mobilized into the circulation under certain circumstances, suggesting that MSCs have the capacity to be in contact with blood. ^{36 –39} However, culture-expanded MSCs may change their characteristics, which possibly affects their hemocompatibility. ³⁵

Caplan and coworkers were the first to show that ex vivo-expanded MSCs are safe for systemic infusion ⁴⁰ and that infused MSCs are cleared from the circulation within 60 min. ⁴¹ Adult MSCs express intermediate levels of HLA major histocompatibility complex (MHC) class I molecules. In resting MSCs, MHC class II is not expressed on the surface, ⁴² which reduces the inherent immunogenicity of the cells. ⁴³ This supreme attribute allows allogeneic MSC transplantations. However, systemic infusion of MSCs handled in vitro leads to immediate activation of the complement and coagulation systems, a reaction termed “instant blood-mediated inflammatory reaction” (IBMIR) ³⁵ resulting in cell lysis and graft destruction. This innate immune attack is known to damage and compromise the engraftment also of islet cells as well as hepatocytes after intravenous administration. ^{44 –47} It has been shown that protection against the blood complement and coagulation cascade systems is attenuated upon culture expansion as well as cryopreservation, again highlighting the importance of low passage for optimal functionality of the cells. ^{48 –50} In addition, long-term engraftment of infused cells is extremely low to undetectable as assessed in autopsy material. ^{51 –53} In patients with advanced cirrhosis who were infused with radiolabeled MSCs, it was shown that initially the cells accumulated in the lungs but were gradually redistributed to the liver and spleen, where they could be detected up to 10 days postinfusion. ⁵⁴

Antidonor immune responses of various degree elicited by allo-MSC transplantations have been reported in preclinical models ⁵⁵ and can potentially explain fast clearance and low engraftment. However, no documented adverse events related to an antidonor immune response exist although an extensive number of human subjects have received allo-MSC intervention for various conditions in clinical trials. ²⁶ Overall, these observations, including the fast clearance from circulation and low engraftment, suggest that MSCs mediate their long-term effects in a paracrine manner. MSCs possess a phenotypic plasticity and harbor an arsenal of bioactive molecules that can be released upon sensing signals in the local milieu. ⁵⁶ The reported paracrine effects comprise many of the important functions of MSCs, including supporting HSCs in the BM, ^{57 –59} promoting angiogenesis, ⁶⁰ and modulating the immune system. ^5,61

Immunomodulatory Mechanisms of MSCs

The immunomodulatory properties of MSCs are being utilized in the treatment of a vast number of immune-based disorders, including GvHD, ARDS, and multiple sclerosis, and in newly diagnosed diabetes. Even though the mechanisms by which cells provide the therapeutic effect are not completely known, research has documented that MSCs display dual immunomodulatory properties, including proinflammatory and anti-inflammatory actions depending on the challenging signal, ^5,62,63 suggesting that the status of circulating signals at the time of infusion is critical to determining the nature of their immune properties. There have been several recent comprehensive reviews on MSCs covering their immunomodulatory competence and interaction with individual immune cell subsets, which include T-cell subsets, B-cells, NK cells, dendritic cells, and neutrophils. ^{2 –5,14,64 –66} Since immunomodulation mediated by MSCs requires preparative activation, the actual stimuli present in the local milieu will determine the nature of the MSC response. The proinflammatory MSC phenotype promotes secretion of IL-6, IL-8, IFNβ, MIFs, GM-CSF, CCL2, CCL3, and CCL12 upon exposure to bacterial products (lipopolysaccharide [LPS] or peptidoglycane [PGN]) via toll-like receptors (TLRs) expressed on the MSC cell surface. The released factors are shown to increase neutrophil and monocyte migration to a site of injury. The recruited monocytes differentiate into and are maintained as M1-proinflammatory macrophages in this milieu, which promote bacterial clearance in the area. ⁶⁷

The immunosuppressive MSC phenotype is developed in response to proinflammatory cytokines released by immune cells, primarily Th1 cells. Both human and rodent MSCs have been reported to inhibit proliferation and effector functions upon stimulation of peripheral blood mononuclear cell (PBMC) subpopulations in co-culture experiments in vitro. ² Again, the MSC-mediated activity mainly acts through secretion of soluble molecules with immunosuppressive function. ⁴ For human MSCs, the IFNγ-regulated enzyme indoleamine 2,3-dioxygenase (IDO) has consistently been reported to mediate the T-cell inhibitory mechanisms. ⁶⁸ This enzyme catalyzes the conversion of tryptophan to kynurenine causing depletion of local tryptophan but most importantly accumulation of toxic breakdown products in the area, resulting in T-cell apoptosis. ^4,69 However, both IDO- and IFNγ-independent mechanisms have been reported for human adult MSC-mediated immunosuppression, ^70,71 offering an explanation for the sustained activity of MSCs lacking a functional IFNγ receptor 1 and inhibited IDO production. Species-specific mechanisms of immunosuppression have also been reported where nitric oxide that affects macrophage and T-cell function is described as a major immunosuppressive molecule in mice. ⁷²

Prostaglandin E₂ (PGE₂) is one of the bioactive molecules being secreted by MSCs in response to inflammatory cytokines like IFNγ or TNFα. PGE₂ has been shown to elicit both immunosuppressive as well as proinflammatory effects. ^{73 –75} The immunosuppressive potential of PGE₂ is based on its inhibition of T-cell proliferation and IL-2 production ⁷⁶ as well as the inhibition of additional immune cells, including NK cells, granulocytes, and dendritic cells (reviewed in ref. ⁷³ ). PGE₂ has also been shown to induce FOXP3 gene expression and T regulatory function from naïve T-cells. ⁷⁷ By providing an opportunity to affect many immune cell subtypes, PGE₂ thus qualifies as one of the major immunomodulatory molecules secreted by MSCs. However, the proinflammatory effects include a PGE₂-facilitated expansion of both human and mouse Th17 cells. ⁷⁸ By blocking specific PGE₂ receptors, an immunosuppressive effect has also been reported. ⁷⁹ The dual outcome of PGE₂ secretion may possibly depend on other regulatory factors in the local milieu as well as the PGE₂ concentration.

Microrna Regulation of Paracrine Effects of MSCs

It has recently become evident that microRNAs (miRNAs) are an integral part of the molecular networks coordinating many aspects of inflammatory responses. ^80,81 These small noncoding RNAs affect inflammatory processes primarily by having an impact on activated subsets of cells and by controlling the levels of cytokines produced. In MSCs, miRNAs have recently been shown to be critically involved in regulating differentiation, proliferation, survival, as well as the paracrine activity. ⁸² These miRNAs modulate transcription and control the levels of proteins being secreted by MSCs. Importantly, and as discussed later, they can also be released in EVs and exert their regulatory function in target cells. As mentioned above, one of the major immunomodulatory molecules secreted by MSCs is PGE₂. Interestingly, miR-146a has been reported to target PGE ₂ ⁸² as well as stimulate production of IL-8 in MSCs in response to TNFα, ⁸³ suggesting regulatory involvement in the immunomodulatory capacity of MSCs. In addition, miR-155 is a TLR-induced miRNA, which is negatively regulated by IL-10, suggesting a central role in immunomodulation. ⁸⁴ Initial proinflammatory responses mediated by TLRs may partly be mediated by a fast upregulation of miR-155, while later in the response upregulation of miR-146 and miR-21 promotes anti-inflammatory activity with a facilitation of IL-10 production. ⁸⁵ Deep sequencing and array technology have provided information about which miRNAs are expressed in MSCs. It is clear that tissue origin differences exist ^{86 –88} ; however, there is a high representation of let-7 family members in all MSCs derived from various tissues, which have been shown to regulate IL-6 and IL-6-induced migration of cells. ⁸⁹ Several miRNAs (miR-27b, miR-126, miR-146a, and miR-886) have been reported to target stromal-derived factor 1 (SDF-1), ^{90 –94} a chemokine released by MSCs in order to attract immune cells, further supporting the miRNA control of MSC activity. ⁸² The fact that the miRNAs affect several targets generates a situation where small changes in miRNA levels can have a substantial impact on cell function.

EVs as Paracrine Mediators of MSCs Effects

The wide range of molecules involved in the paracrine effects of MSCs can be packaged in EVs and released, and have the potential to affect cellular processes in nearby cells. These molecules include proteins, mRNA, miRNAs, and other entities capable of activating a range of cell signaling pathways. ⁹⁵ Vesicle-mediated cell communication complements messages conveyed by secreted soluble factors and direct cell-to-cell contacts. The specific subsets of proteins, mRNAs, miRNAs, lipids, and metabolites harbored in the EVs represent the actual status of the releasing cell, and keeping in mind the plasticity of MSCs, the contained bioactive molecules most likely vary depending on the local milieu. EVs were at the time of discovery in reticulocytes regarded as secreted waste material with no real biological function. ⁹⁶ However, in 1996, Raposo et al. demonstrated that EVs stimulate adaptive immune responses. ⁹⁷

Today, EVs are recognized as instrumental in numerous biological and pathological processes. ^{98 –101} EVs have now been isolated from most body fluids and exert their effects on biological processes in a pleiotropic manner, directly activating cell surface receptors via protein and bioactive lipid ligands, merging their membrane contents into the recipient cell plasma membrane, and delivering effectors (Fig. 1). EVs are typically categorized based on their biogenesis. The three main classes of EVs are exosomes, microvesicles (MVs), and apoptotic bodies. The common denominator is that they are all enclosed by a lipid bilayer, ranging from 30 to 2000 nm in diameter depending on their biogenesis pathway. The term “exosomes” refers to membrane vesicles with a diameter of 50–100 nm, which are derived from the endo-lysosomal pathway. Inward budding and fission of the multivesicular body (MVB) membrane leads to the formation of intraluminal vesicles. When MVBs fuse with the plasma membrane, the intraluminal vesicles are released as exosomes. In contrast, MVs are generated by the outward budding and subsequent fission of the plasma membrane, which results in the release of generally larger vesicles. However, the dimensions and density of exosomes and MVs can overlap, making it technically challenging to analyze their individual properties separately. ¹⁰²

OPEN IN VIEWER

Figure 1.

Therapeutic potential of MSC-EVs can be mediated by multiple mechanisms. EV-associated enzymes can partake in ATP and adenosine production, leading to generation of energy needed for cell survival, and to activation of ERK and AKT signaling pathways in damaged cells, respectively. Other cell signaling pathways can be activated via ligand–receptor interaction with target cells. These interactions can affect damaged cells directly or indirectly by stimulating secretion of anti-inflammatory cytokines from immune cells. EVs, extracellular vesicles; MSC, mesenchymal stromal cell.

Therapeutic Potential of MSC-EVs in Preclinical Models

The most extensive and well-controlled preclinical evidence regarding the therapeutic potential of MSC-EVs has been demonstrated in models of acute kidney injury (AKI), ^{99,103 –106} myocardial infarction and reperfusion injury, ^{107 –110} hindlimb ischemia, ¹¹¹ and liver ¹¹² and lung injury, ¹¹³ recently summarized in systematic reviews. ^114,115 Many of the effects are related to specific ligand–receptor interactions with the target cells, and also the transport of tolerogenic molecules such as programmed death-ligand 1 (PD-L1), galectin-1 (Gal-1), transforming growth factor beta (TGF-β), and others. ¹¹⁶ By indirect ways, potentially through modulating cell-signaling pathways, MSC-EVs induce the secretion of anti-inflammatory cytokines from targeted lymphocytes. ¹¹⁶ The premise for tissue repair therapies is related to a range of enzymes associated with the MSC-EVs. For example, ATP-generating enzymes (e.g., glycolytic enzymes and various nonmetabolic kinases) have been postulated to provide energy needed to prevent cell death and to support the repair of tissue damage. ¹¹⁷ MSC-EVs mediate cell survival and proliferation signals to sites of damage via activating ERK and AKT signaling pathways. This may occur through enzymatic production of adenosine, a potent activator of both ERK and AKT, by EV-associated ecto-5′-nucleotidase (5′-NT, also known as CD73). ^95,117

In AKI, MSC-EVs have been described to mediate a wide spectrum of general effects, similar to that of their parent MSC. The effects include increased proliferation of tubular cells, ^99,104,106 reduction of apoptosis, ^99,103 amelioration of renal function and recovery, ^{99,103 –105} increased survival of mice, ¹⁰³ and reduction of oxidative stress. ¹⁰⁶ Most importantly, these treatments led to the improvement of several clinical parameters such as blood urea nitrogen and creatinine levels, as well as reduction of uric acid and proteinuria. In these studies MSC-EVs were characterized to display various cell surface markers, for example, CD29 (integrin β-1), CD44 (homing cell adhesion molecule, also known as phagocytic glycoprotein-1), and CD73, which may explain some preclinical effects of these EVs related to activation of ERK signaling pathway via enzymatic production of adenosine, as mentioned. Upon uptake in AKI models, MSC-EVs seem to increase the level of a range of anti-apoptotic Bcl2 family proteins and reduce the level of caspases to suppress apoptosis. ¹⁰³ Nevertheless, it has remained elusive which exact effector molecules are involved in these processes, but some evidence suggests that the effects are partly related to EV-mediated transfer of mRNA. ^99,104

The therapeutic effects of MSC-EVs in myocardial infarction and reperfusion studies can be robustly assessed by estimating the infarct size. Treatment with EVs has been shown to reduce infarct size by nearly 20% on average. ¹¹⁴ Some of the observed effects have been attributed to the EV-mediated restoration of bioenergetics, reduction of oxidative stress, and induction of prosurvival stimuli. ¹⁰⁷ In an in vivo hypoxia model it was shown that MSC-EVs promote angiogenesis and thereby prevent tissue damage. ¹¹¹ However, studies in hypoxia-induced pulmonary hypertension, on the other hand, indicate that hyperproliferative pathways should be under tight control in harmful oxygen-deprived conditions and that MSC-EVs could help cells to retain that control via modulating STAT3 signaling. ¹¹³ It remains unclear whether the latter hypoxia-related mechanisms could also be important in myocardial infarction. Many of the effects in ischemia/reperfusion models are hypothesized to be related to EV-mediated protein transfer, ATP production, and increased AKT signaling, ¹⁰⁷ similarly to some of the effects observed in AKI models above. However, delivery of EV-associated miRNA, especially pre-miRNA, has also been shown to be responsible for certain other protective effects. ¹¹⁰

Investigations into the preclinical potential of MSC-EVs collectively suggest that while the EV-mediated effects seem to be robust and that changes in tissues upon vesicle treatment can be described on a molecular level by analyzing cellular signaling events and monitoring apoptotic, necrotic, and fibrotic events in various models, it is still relatively unclear exactly which EV-associated molecules represent the key factors mediating the beneficial effects. It is likely that many of the molecular interactions are still unknown because the protein and RNA composition of MSC-EVs is extremely diverse. However, even though several key effects can be attributed to a handful of molecules and enzymes, it is likely that the modulation of entire networks of cellular components is crucial. Dissecting the dynamics of complicated networks is often difficult but nevertheless highly needed in order to understand the signature of an optimal MSC-EV preparation.

Conclusions and Advances in Treating Patients with MSCs and MSC-EVs

The cell therapy field has witnessed an increasing number of clinical and preclinical studies reporting on the potential of MSCs and MSC-EVs for treatment of a variety of conditions. As discussed above, the focus in the field to advance the use of MSCs in the clinic involves understanding the fate of the cells following systemic administration as well as identifying mechanisms by which these cells correct disease and, ultimately, identifying crucial characteristics of therapeutically active cells. Given the plasticity these cells harbor, together with the observations described in this review, it will be a necessary challenge to define an improved optimal standardized protocol for cell and EV production. As mentioned, the fast clearance from the circulation and low cell engraftment suggest that MSCs mediate their long-term effects through paracrine functions. It is an appealing hypothesis that the cellular debris after IBMIR in vivo generates EVs that can mediate the systemic long-term effects.

Leaders in the EV field have put extensive efforts into standardizing the production of MSC-EVs by establishing immortalized ES-derived MSCs that permit large-scale production of therapeutically active EVs in preclinical models. ¹¹⁸ Similarly, efficient means of purifying such EVs from large media volumes have been devised by using flow filtration and subsequent size exclusion chromatography. ¹¹⁹ However, despite tremendous progress in successfully moving MSC-EV into various preclinical trials, clinically relevant data are largely lacking from studies in human. Recently, a study performed at the University of Essen used MSC-EVs for the first time to treat a patient suffering from steroid-refractory GvHD. ¹²⁰ This is a hallmark achievement that ought to spur the MSC-EV field toward clinical translation. However, before realizing the full potential of EV therapy, several obstacles have to be overcome.

It will be very important to determine the mode of action of MSC-EVs and to understand the molecular content of EVs being responsible for the observed effects. It is also critical to establish optimized protocols for cellular expansion in terms of culture conditions as they might have a dramatic impact on the molecular composition and potency of EVs. Although EVs have the advantage over cells in that they are replication incompetent as they largely lack DNA, it remains to be shown how well EVs perform compared with their parental cells. To our knowledge, no study has addressed this systematically. Efforts should also be put into investigating whether the source of MSC-EVs (BM, adipose, ES-derived, etc.) is important for activity and thorough comparisons should be made with non-MSC cellular sources (e.g., fibroblasts) to ascertain that the effects observed are linked specifically to cellular multipotency. Finally, and importantly, it is crucial to establish GMP-compliant protocols for the purification of EVs to ensure effective translation from preclinical to clinical studies.