Immune modulation by MSCs
Some of the first evidence that MSCs could actively blunt immune responses originated from the results of mixed lymphocyte reaction (MLR) assays performed ex vivo [30–36]. These assays are based on the observation that T cells from preparations of immunologically mismatched peripheral blood mononuclear cells proliferate rapidly when mixed together under appropriate conditions [37, 38]. Results from MLR assays showed that T-cell expansion could be inhibited by the addition of MSCs to MLRs. While the majority of cell culture studies to date agree that such observations are mediated by MSC-derived soluble factors that do not cause T-cell apoptosis, several alternative mechanisms have also been proposed. Di Nicola et al.  employed a series of antibody blocking assays to implicate the role of transforming growth factor beta (TGFβ) and hepatocyte growth factor (HGF) whereas Aggarwal et al.  proposed a role for prostaglandin E2 (PGE2) based on their ability to ablate inhibitory responses with cyclooxygenase 2 (COX2) inhibitors. Aggarwal et al. further proposed that the secretion of PGE2 and related factors induced dendritic cells to up-regulate the anti-inflammatory cytokine interleukin (IL)10 while reducing the secretion of pro-inflammatory tumor necrosis factor alpha (TNFα) and IL12. This, in turn, initiates a shift in the ratio of T helper (Th) cells from a pro-inflammatory Th1 subtype to an anti-inflammatory Th2 subtype. This was accompanied by the differentiation of naive T cells to an immunoregulatory regulatory T cell (Treg) phenotype, thereby reducing the overall number of Th cells. Similarly, Akiyama et al.  showed that MSCs could induce apoptosis of inflammatory T cells through activation of the Fas–Fas ligand axis. During this process, MSCs recruited additional T cells by secretion of monocyte chemotactic protein-1 (MCP-1) as part of a positive feedback loop. Apoptotic T-cell debris then activated phagocytes to secrete TGFβ, resulting in the differentiation of naive T cells into Treg cells that can promote systemic immune tolerance . In an alternative model, Meisel et al.  proposed an intriguing mechanism whereby MSC-derived indoleamine-2,3-dioxygenase (IDO) catalyzes the conversion of tryptophan to kynurenine in an interferon gamma-dependent manner. In turn, the kynurenine inhibits T-cell proliferation [40, 41]. This mechanism was later confirmed by utilizing the IDO antagonist 1-methyl-L-tryptophan . In a series of experiments performed by Waterman et al. , it was reported that MSCs could be induced to express enhanced levels of IDO and PGE2 by transient stimulation of toll-like receptor (TLR)3 with polyinosinic-polycytidylic acid (poly I:C). MSC-mediated IDO activity has also been shown to enhance kidney allograft tolerance in mouse models through a mechanism involving Treg up-regulation, demonstrating that IDO-mediated mechanisms of immune modulation can indeed occur in vivo . Nitric oxide , galectin-1 and semaphorin-3A  have also been implicated as MSC-derived modulators of T-cell proliferation, but it is noteworthy to add that nitric oxide has only been shown to function as an MSC modulator in the murine system.
MSCs also have the capacity to modulate the activity of macrophages. This effect was initially described ex vivo using macrophage cultures stimulated with TLR ligands such as lipopolysaccharide (LPS), zymozan, or polyinosine-polycytidylic acid (poly I:C); these simulate the effects of bacterial or viral infection [47, 48]. When macrophages are challenged with such agents, they secrete inflammatory factors such as TNFα, IL1β, IL6, and reactive oxygen species. In the presence of MSCs, however, the ability of activated macrophages to secrete inflammatory factors was attenuated [32, 49]. Of interest, these observations were explained, in part, by MSC-mediated secretion of the extracellular protein TNFα-stimulated gene protein (TSG)6 . In this model, exposure to zymozan caused cultured macrophages to secrete high levels of TNFα and other inflammatory mediators via the TLR2–nuclear factor kappa-B (NFkB) axis. TNFα activates TSG6 expression by MSCs and engages a negative feedback loop by inhibiting NFkB via activation of the CD44 receptor. Several in vivo studies have confirmed that MSC-derived TSG6 acts via the CD44 receptor to inhibit NFkB activity in macrophages, dendritic cells, and Th cells in models of peritonitis , diabetes , and corneal transplant rejection . In addition to the action(s) of TSG6, MSC-derived PGE2 has also been demonstrated to have potent effects on macrophages in vivo. In a murine model of sepsis, Nemeth et al.  demonstrated that, upon activation by LPS or TNFα, MSCs secreted PGE2. This caused the release of anti-inflammatory IL10 by macrophages and improved cell survival. Indeed, the role of PGE2 in MSC-mediated macrophage modulation is a common theme in many culture models [54, 55]. In an alternative mechanism proposed by Chen et al. , placental human MSCs inhibited the interaction of TLR4 with a key effector molecule, MyD88 , resulting in inhibition of secretory factors by macrophages. This process was inhibited by addition of a COX2 inhibitor, suggesting that the process was PGE2-dependent.
MSCs were reported to modulate the proliferation, differentiation, and immunoglobulin secretion of B cells without induction of apoptosis . Transwell assays separating the two cell types but allowing for exchange of secreted factors showed that such MSC-mediated effects derived, in part, from the paracrine activity of soluble factors secreted by MSCs. These experimental results have since been replicated using purified B cells and unpurified preparations of peripheral blood mononuclear cells [58–60]; however, the paracrine mechanism was recently challenged by a co-culture study that suggested physical interaction between T cells and MSCs was necessary for MSCs to inhibit the activities of B cells . Using a mouse model of allergy, Nemeth et al.  reported that MSC-derived TGFβ was critical in suppressing B-cell mediated allergic responses in vivo. They speculated that MSCs may recruit Treg cells that down-regulate allergy-specific cytokine and immunoglobulin production as well as lung eosinophil infiltration. Consistent with their immune-modulatory properties, efficacy with MSC treatment has been demonstrated in a variety of inflammatory models of disease, including arthritis , Crohn’s disease , multiple sclerosis [65, 66], myocardial infarction , diabetes [51, 67], graft versus host disease [34, 68, 69], and corneal rejection .
Promotion of cell survival by MSCs
In addition to the paracrine effects of MSCs on immune cells, they also secrete a diverse repertoire of factors that support cell survival, including growth factors, cytokines, and extracellular matrix (ECM). Together, the components of the MSC secretome have the theoretical capacity to rescue injured cells, reduce tissue damage, and accelerate repair. This is exemplified by their natural roles as reticular cells that support the hematopoietic stem cell niche [26–28, 70, 71] and as vascular pericytes that support endothelial cells [72, 73]. The observation that MSCs can be isolated from a wide variety of tissues, such as bone marrow, adipose, ligament, skin, placenta, dental pulp, synovium, placenta, umbilical cord, and other fetal tissues [72, 74], lends support to the concept that they function endogenously as stromal support cells.
The pro-survival effect(s) of the MSC secretome on other cell types was first recognized through studies of long-term bone marrow cultures [26–29, 75] and embryonic cells . Collectively, these cell culture studies provide for an attractive, paracrine-based explanation for the ability of MSCs to promote healing across a broad range of developmentally unrelated tissues and for myriad diseases and injury types. Detailed analysis of the MSC transcriptome and proteome has confirmed that they secrete a vast repertoire of paracrine pro-survival factors commonly referred to as trophic factors or mediators [77–82]. Of interest, the MSC-secreted factors comprise a diverse group of soluble peptides and proteins with complementary set(s) of biological activities that can accelerate progenitor cell self-renewal, stimulate angiogenesis, and minimize apoptosis and/or inflammation. Despite several decades of research and progress, the specific paracrine mechanisms by which administered MSCs improve cell survival and self-renewal under particular contexts of tissue rescue/repair remain largely undefined [75, 77].
In line with the traditional model of paracrine biology whereby cells secrete factors that regulate adjacent cells, it was initially thought that engrafted MSCs readily migrated into injured tissue and then remained to orchestrate repair. For many models of tissue injury, however, what was originally perceived as “MSC migration” turned out to be far less directed (e.g., non-specific, transient trapping of MSCs within the microvasculature and capillary network). Of particular interest, depending on their relative size (i.e., diameter), the majority of intravenously administered MSCs will typically lodge in the lung microvasculature upon the first pass through the circulation, regardless of the presence or absence of lung-specific injury. Notably, after intravenous MSC infusion, paracrine factors released into the blood by circulating MSCs or from trapped MSCs may indirectly influence survival signaling and the fate of distal cells previously compromised by injury or disease. Thus, for effect, paracrine factors produced by MSCs appear not to depend on long-term MSC engraftment, nor do they require the unlikely differentiation of mesodermal progenitors into tissues of ectodermal or endodermal lineages.
Some of the best evidence supporting an indirect role for MSCs in the repair of tissues/organs originates from studies of heart with infarction. In a rat model of myocardial infarction, MSCs modified with the gene encoding protein kinase B (a.k.a. Akt) engrafted into the myocardium, reduced pathological remodeling, and improved cardiac function . The observed efficacy was later attributed to a paracrine effect mediated by secreted frizzled related protein (sFRP), a Wnt signaling inhibitor that reduces cardiomyocyte apoptosis [84–86]. Since these studies, a number of additional mechanisms for the paracrine action of MSC-derived factors on cardiac repair have been proposed, including secretion of angiogenic factors [87–89], stromal cell derived factor-1 (SDF-1) , and Jagged/Notch signaling [89, 91]. Of interest, MSC-mediated improvements in cardiac function could be achieved without long-term engraftment of MSCs . Using a different approach, MSC-conditioned medium was employed to prime cardiac stem/progenitor cells prior to cardiac grafting in a rat model of myocardial infarction. The conditioned medium (CM) improved cardiac stem cell engraftment through mechanisms involving connective tissue growth factor and insulin signaling .
The role of MSCs in the protection of other damaged tissues has also been demonstrated. For example, intraperitoneally and intravenously administered MSCs from murine bone marrow and adipose tissue had a protective effect in a cisplatin-induced acute kidney injury (AKI) model , as evidenced by a reduction in the apoptosis of tubule cells and improved renal function. This effect appeared to be mediated by secreted factors since the results could be repeated by intraperitoneal administration of CM generated from the MSCs (MSC-CM). In contrast, Xing et al.  reported that murine MSC-CM containing HGF, vascular endothelial growth factor (VEGF)-A and insulin-like growth factor (IGF)-1 failed to protect the kidneys of mice against ischemia-reperfusion injury, whereas live MSCs had a significant protective effect. This is one of several examples in the field where apparently minor differences in the cell source, the culture conditions, duration of medium conditioning, and dosage can profoundly affect outcome. Such complexities have made elucidation of the mechanism(s) responsible for the protective effect of MSCs on kidney tissue challenging, but some progress has been made. For example, Zarjou et al.  demonstrated that the stress-responsive enzyme heme-oxygenase-1 (HO-1) played a role by utilizing MSC from bone marrow of HO-1-/- mice. In this study, HO-1+/+ MSC-CM rescued pathology associated with cisplatin-induced AKI, while HO-1-/- MSC-CM was ineffective. The authors attributed the difference in effect to enhanced levels of SDF-1, VEGF-A, and HGF in the HO-1+/+ MSCs. Indeed, immunological and transcriptional blocking experiments both confirm a protective role for VEGF-A [96–98] and IGF-1  in mice with AKI and for VEGF-A in rats with cerebral ischemia (stroke) .
The utility of MSCs and their secreted products to protect cells and to foster tissue repair has been demonstrated in numerous efficacy-based studies across a broad range of tissue injury and disease models. While a comprehensive summary of the associated literature is beyond the scope of this review, some key examples of MSC-derived benefits include facilitation of wound healing , improved treatment of diabetes , enhancement of bone repair [103, 104], and effect(s) on cancer .
Effects of MSCs on fibrosis
Fibrosis is generally defined as a an accelerated accumulation of ECM factors (predominantly collagen type I) that prevents the regeneration of tissue. It can occur in virtually any tissue as a result of trauma, inflammation, immunological rejection, chemical toxicity, or oxidative stress. Current clinical strategies generally have poor outcomes in terms of efficacy and adverse effects . Given the immunomodulatory and trophic properties of MSCs, they have become attractive candidates for the treatment of fibrosis and preclinical studies suggest they have a promising level of efficacy in a variety of models. While the anti-fibrotic effects of MSCs are likely to overlap with their anti-inflammatory and angiogenic properties, the specific mechanisms remain poorly understood. Nevertheless, a comprehensive review by Usuner et al.  suggests that their modes of action seem to fall under four categories: i) immune modulation, ii) inhibition of TGFβ-mediated differentiation of various cells types into ECM-secreting myofibroblasts by epithelial to mesenchymal transition, iii) inhibition of oxidative stress, and iv) matrix remodeling. For example, Ortiz et al. demonstrated that systemic murine MSC administration attenuated fibrosis in a bleomycin-induced lung injury model . This was achieved through MSC-mediated secretion of IL1 receptor antagonist, which reduced infiltration of lymphocytes and neutrophils and their production of inflammatory and fibrotic mediators such as IL1 and TNFα. Using the same model, it was recently reported that MSCs had the capacity to inhibit fibrosis through the action of the secreted protein stanniocalcin-1 (STC-1) . The authors demonstrated that STC-1 acted in multiple ways by reducing the secretion of collagen by fibroblasts, by reducing TGFβ output by endothelial cells and also through alleviating oxidative stress by uncoupling mitochondrial respiration via the induction of uncoupling protein 2. Using a model of chronic kidney injury, Huuskes et al.  demonstrated that MSCs improved kidney morphology and functionality when co-administered with the putatively anti-fibrotic hormone recombinant human relaxin (serelaxin). In this system, MSCs and serelaxin acted synergistically to reduce TGFβ-induced myofibroblast differentiation and collagen deposition while increasing the level of matrix metalloproteinase 2 (MMP2), a collagen-degrading enzyme.