Graft-versus-host disease (GVHD) is a major complication associated with morbidity and mortality following allogeneic hematopoietic stem cell transplantation (HSCT) characterized by dysregulation of inflammatory cytokines and activated donor cells that attack recipient organs and tissues. In the pilot study using MSCs for GVHD, Le Blanc et al. reported the successful use of third-party MSCs to treat steroid-refractory GVHD following allogeneic stem cell transplantation (36). This initial study provided a rationale for the use of MSCs in GVHD and encouraged further pre-clinical and clinical studies of MSCs.

Unlike in vitro studies that have clearly demonstrated MSCs’ immunosuppressive effects against allogeneic T cell responses, in vivo studies have been more ambiguous. Pre-clinical studies revealed that various factors, including cell dose, timing of infusion, and pre-activated state of MSCs contribute to the therapeutic effects of MSCs. While some studies have suggested that MSCs can effectively treat GVHD in a dose-dependent manner (40, 41), others have shown that therapeutic effect could not be obtained at any dose (42–45). In addition to cell dose, studies proposed that the timing of treatment might play a more critical role in MSC-mediated immune suppression. Regarding the timing of administration, Polchert et al. revealed that MSCs were ineffective before GVHD development or when GVHD was too severe. This study attributed the inconsistency of MSC treatment to the difference in IFN-γ levels at different time points of GVHD development (43). Th1 cytokine, IFN-γ play a pivotal role in the immunomodulatory function of MSCs. MSCs pretreated with IFN-γ could prevent GVHD even when administered at the time of transplantation. Nonetheless, several pre-clinical studies failed to exhibit therapeutic potential of MSCs regardless of various timing and dose (2, 44, 46).

With the dramatic outcome of the first GVHD patient treated with MSCs, MSC therapy has undoubtedly showed promising results for the treatment of GVHD. In Ringden’s subsequent study (47), MSCs administered in patients with grades III to IV steroid-refractory GVHD showed a complete remission rate of 75%. The European Group for Blood and Marrow Transplantation obtained similar results in a multicenter phase II study using allogeneic and third-party bone marrow derived MSCs (48). Patients with steroid-refractory GVHD were treated with MSCs from various donor sources and had a significantly reduced level or transplantation-related mortality. While the small sample size and heterogeneous treatment protocols and MSC products limit the characterization of MSC efficacy, it is becoming clear that MSC therapy may be more effective in specific environments indicating that mixed results of MSCs for GVHD may arise depending on age, organ involvement, and severity of GVHD patients. In a multi-center trial, a greater proportion of pediatric patients responded to MSCs than adults (48). Furthermore, patients with skin-involved GVHD generally had a higher response rate to MSC treatment (49–51) whereas some studies have suggested that gastrointestinal GVHD may respond better (52, 53). In addition, studies that included both acute and chronic GVHD patients, the response was higher in acute GVHD (49, 51). It is important to recognize, however, that there is generally a lack of studies on de novo GVHD, chronic GVHD, and GVHD prophylaxis. Majority of studies involved patients resistant to conventional steroids and failed at least their first-line of treatment (13, 25, 43, 50, 54, 55) suggesting a better outcome in these settings. In contrast, there is some evidence that MSCs may be less effective in cGVHD (31) and GVHD prophylaxis (50).

Despite these potential variables in identifying optimal clinical settings for MSC therapy, the efficacy of MSCs is still unpredictable. In the phase III industry-sponsored trial (NCT00366145) using Prochymal for the treatment of steroid-refractory GVHD, public reports announced that Prochymal failed to achieve increased complete response rate compared to placebo controls (39, 52).

Despite rapid medical advancements in the cardiovascular field, inflammatory reactions that occur upon heart diseases and failures continue to be associated high morbidity and mortality rates. Initially, MSCs were used following myocardial infarction due to their regenerative and tissue repair properties. Pre-clinical studies focused on the evaluation of MSC therapy through the measure of cardiac function and effects on cardiac remodeling (56–58). However, with the discovery that only a small portion of the injected MSCs remained present in the heart (59, 60), interest shifted toward MSC-produced paracrine factors as a critical role post myocardial infarction (61, 62). Despite different reported mechanisms of actions, MSC therapy in cardiovascular diseases of animal models have showed overall positive results.

In a pilot clinical trial, patients with acute myocardial infarction received intracoronary injection of autologous MSCs compared to saline (63). The administration of MSCs showed significant clinical improvement compared to the control group. Through cardiac electromechanical mapping, Chen et al. were able to detect viable MSCs up to three months after infusion and increased cardiac functions as demonstrated by cardiac echocardiography. Since then, further clinical trials attempted to assess the safety and efficacy of MSC transplantation for the treatment of cardiovascular diseases, including myocardial infarction and chronic ischemic cardiomyopathy (64–69). Similar to GVHD trials, the heterogeneity of clinical protocols has limited the direct comparison of MSC efficacy in cardiovascular diseases.

The different administration routes by MSC therapy may especially affect the outcomes of cardiovascular diseases. The delivery methods of MSCs include percutaneous coronary intervention, intracoronary injection, intramyocardial injection, and intravenous injection. However, direct comparison between the administration routes and following clinical outcomes is not available. There is evidence, however, that the intracoronary delivery may cause microinfarction by microvascular obstruction (70) or undesired tissue differentiation, such as bone, in the myocardium (71). Interestingly, the intravenous injection of MSCs in the randomized placebo-controlled dose-escalating trial by Osiris Therapeutics, led to an improved clinical outcome compared to the placebo group (65). However, in this study, 6-months follow-up was too short to evaluate the role of MSC in the recovery of cardiac function. Similarly, patients receiving intracoronary injection of MSCs initially showed significant improvement in left ventricular ejection compared control; however this difference did not last up to 18 months (38). Therefore, while both direct and systemic infusions of MSCs are clinically feasible for cardiovascular diseases, the efficacy of MSCs remains to be elucidated due to limited follow-up.

Crohn’s disease, also referred to as inflammatory bowel’s disease (IBD), is a chronic inflammatory autoimmune disease in which the immune system attacks the gastrointestinal tract. In pre-clinical studies, MSCs were hypothesized to exert immunomodulatory effects that would improve the pathogenesis of IBD. Previously, we reported for the first time that MSCs are not clinically beneficial in IBD treatment (72). Other pre-clinical studies have suggested the positive outcomes of MSCs for the treatment of IBD; however, in these studies, MSCs were not used alone, but rather genetically modified to IL-12p40 (73), coated with antibodies (74), or genetically depleted with autoimmune regulators (75). In a single study that did claim the positive effects of MSCs, improved clinical IBD score consisting of weight loss, stool consistency and stool bleeding were noted but immune modulation of inflammatory cytokines was not described (76).

In contrast to the clinical application of MSCs for other diseases, the use of MSCs for IBD treatment demonstrates the most homogeneity in the source of MSCs and administration protocols. Majority of trials used either allogeneic (77–82) of autologous MSCs (83) derived from adipose tissue; however, the use of bone-marrow derived MSCs has also been reported (84, 85). Furthermore, majority of clinical trials used direct intralesional infusion of MSCs mixed with fibrin glue (77–84) whereas only a single study reported the intravenous injection of MSCs (85). While MSC therapy showed local healing of Crohn’s fistula through intralesional infusions, the efficacy of MSC therapy in refractory IBD patients treated with systemic infusions is ambiguous. Only a small portion of patients showed clinical response six weeks post-treatment (85). The observations that intralesional treatment of MSCs mainly improved local symptoms rather than the basic pathogenesis itself support the notion that the systemic delivery of MSCs may not be efficacious indicating the important discrepancy of MSC therapy based on administration routes.

In addition to Crohn’s disease, MSCs for the treatment of autoimmune diseases in various animal models, such as experimental autoimmune encephalomyelitis, systemic lupus erythematosus (SLE), and rheumatoid arthritis (RA) have been reported. Multiple sclerosis is a chronic inflammatory demyelinating disease of the central nervous system. Initially, the therapeutic potential of MSCs have been implicated in neurodegenerative diseases of the central nervous system including stroke (86), Parkinson’s disease (87), and spinal cord injury (88). Although the precise mechanisms of MSCs were not understood at the time, it was hypothesized that the immunomodulatory effects of MSCs could regulate various effector cells involved in the pathogenesis of neurodegenerative diseases. In experimental autoimmune encephalomyelitis, a model for multiple sclerosis, the intravenous administration of MSCs induced tolerance to myelin oligodendrocyte glyco-protein promoting improvement in clinical symptoms associated with the reduction of demyelination and CNS infiltration by lymphocytes (89). However, MSC therapy was only effective when used at disease onset or at the peak point during inflammation; however, not during chronic phase. Therefore, MSCs may be highly dependent on the timing of administration in multiple sclerosis, as in GVHD.

Clinical studies on MSCs for multiple sclerosis are limiting. The intrathecal administrations of autologous MSCs for these patients were feasible and safe; however, the clinical results remain unclear. While clinical improvements associated with reduced expanded disability status scale were reported in some patients, others showed no improvement or worsening of diseases (90–92). However, limited studies and small sample size prevent definitive conclusions on the effects of MSC therapy for multiple sclerosis.

Next, SLE is an autoimmune inflammatory disease with multi-organ involvement including the kidney, brain, lung, and hematopoietic systems. It has been previous suggested in both mice (93) and humans (94) that MSCs derived from the diseased individuals have abnormalities in terms of phenotype, proliferation, and differentiation. The transplantation of MSCs from healthy donors, however, was able to ameliorate the disorder (93). These observations support the use of allogeneic, rather than autologous, MSCs for SLE. In murine models, allogeneic MSCs or xenogeneic human cord-blood derived MSCs for the treatment of SLE. However, the results have not been consistent. While both human and allogeneic murine MSCs were able to alleviate renal functions associated with SLE (95), other clinical symptoms including proteinuria or double-stranded DNA levels remained unchanged (19). Furthermore, allogeneic MSCs could also enhance anti-double stranded DNA antibody production levels and worsen the disease (96).

Similarly, in the clinical setting, the use of allogeneic and umbilical cord-derived MSCs have shown therapeutic potential in active SLE patients correlated with ameliorated disease activity, improved serological markers, and stabilized renal functions (93). Furthermore, MSC therapy was feasible in patients with refractory SLE (97, 98). The results of these initial clinical trials are encouraging; however, the absence of larger randomized controlled clinical trials for SLE and other autoimmune diseases need to be resolved for further analysis.

RA is a T-cell-mediated autoimmune disease characterized by cartilage and bone destruction. Despite continued research on the role of MSCs in RA, therapeutic potential of MSCs is controversial. In our study, we observed that MSCs are ineffective for treatment of collagen-induced arthritis (CIA), a murine model for rheumatoid arthritis, promoting Th17 related cytokines and subsequently aggravating symptoms of CIA (99). In another study, the negative effects of MSCs in CIA were associated with the presence of TNF-a, which reversed the immunomodulatory properties of MSCs and worsened clinical symptoms (100). Various doses were injected at the time of immunization or booster injection; however, the aggravation of clinical symptoms was similar in both conditions. In contrast, therapeutic effects have also been demonstrated (100). A single injection of MSCs at the time of CIA induction could prevent the incidence of CIA. Interestingly, MSC treatment following the establishment of CIA could also prevent the exacerbation of disease in which 7 of 10 animals treated with MSCs showed significantly lower disease score (101).

The mixed results demonstrated in pre-clinical studies have delayed the application of MSCs for RA treatment in clinical trials. The use of MSCs in RA patients had not been reported in the clinic until recently. In 2013, Wang et al. reported the safety and efficacy of umbilical cord derived MSCs in the treatment of RA (102). In the ongoing cohort, 172 patients with active RA refractory to traditional treatments were allocated into two groups, in which patients received anti-rheumatic medication with or without MSCs. MSC therapy in RA were safe and effective in controlling the refractory disease. In MSCs-treated groups, improvements of clinical manifestations correlated with decreased levels of inflammatory cytokine and increased percentages of regulatory T cells. The clinical benefits were persistent up to 6 months with a single infusion and repeated infusions could be tolerated and further enhanced efficacy. Thus, in contrast to animal models, MSCs may provide significant benefits for RA.

Sustained expression of therapeutic genes through gene modification can significantly enhance the potency of MSCs independent of external inflammatory stimuli. IL-10 transduced MSCs have been used to treat GVHD (116). While untransduced MSCs were ineffective in suppressing the development of GVHD, IL-10 transduced MSCs significantly decreased mortality rates of mice, which correlated with decreased levels of pro-inflammatory cytokines. This clinical benefit also correlated with MSCs ability to deliver IL-10 to target inflammatory sites. Gene-modified MSCs have also been reported in various experimental autoimmune models. We have previously showed that MSCs transduced with TGF-β could potently suppress CIA models compared to untransduced MSCs (99). IL-12p40 expressing MSCs could also ameliorate murine colitis compared to normal MSCs (46). Also, MSCs engineered to overexpress IL-4 could attenuate experimental autoimmune encephalomyelitis (117).

However, genetically engineered MSCs have yet to be applied in the clinical setting for immune modulation. The use of genetically engineered MSC may raise critical safety issues for clinical application. In addition to genetic manipulation, the consistent secretion of anti-inflammatory cytokines may paradoxically cause pathological immune responses depending on the factors involved in disease progression. Therefore, efforts to transiently strengthen MSC-based immune modulation may be more clinically relevant.

The pre-treatment of IFN-γ has been previously discussed in GVHD (43). Furthermore, pre-treatment of MSCs with growth factors, including fibroblast growth factor, insulin-like growth factor, bone morphogenetic protein and stem cell derived factors have been reported to enhance cell survival and cytoprotective effects of MSCs exposed to hostile environments of hypoxic myocardial ischemia (118, 119). Moreover, activation of nucleotide-binding oligomerization domain 2 expressed on human umbilical cord-blood derived MSCs by muramyl dipeptide induces upregulation of PGE2 which significantly enhances the inhibition of mononuclear cell proliferation (120). NOD2 signaling could be useful in that it is specifically expressed on umbilical cord-blood derived MSCs rather than other cell sources.

One of major mechanisms of MSC-based immune modulation is the induction of regulatory cell subsets including Tregs, both in vitro and in vivo (40, 121). Therefore, MSC-mediated immune modulation is interdependent on the presence of endogenous Tregs. We postulated that soluble factors secreted by MSCs can promote the induction of Treg differentiation, and in turn, the cytokines produced by Tregs can promote immunosuppressive potential of MSCs. The combination of two immunosuppressive cell types, MSCs and Tregs, could synergistically support their functions and stabilize their plasticity. Supporting this hypothesis, a study reported that MSCs and Tregs, in combination, do not impair each other’s respective functions (122). In our study, we compared single cell therapy groups (MSCs or Tregs alone) with combined cell therapy group initially in an acute GVHD model (123). We observed that the combined-cell therapy approach had synergistic immunomodulatory effects in inducing long-term survival and reducing clinicopathological symptoms of GVHD which was associated with effective inhibition of both Th1 and Th17 responses compared to MSCs-treated alone. In subsequent studies, we further investigated combined-cell therapy of MSCs and Tregs in other transplantation models, including the induction of mixed chimerism following nonmyeloablative allogeneic HSCT (124), and the prevention of allogeneic skin-graft rejection (125), and observed similar results.

Based on reported clinical studies, MSC therapy seems to be most effective in immunosuppressant-resistant conditions, while randomized studies involving immunosuppressant-sensitive patients did not show therapeutic efficacy. Conventional immunosuppressants used to treat immune-mediated disorders involve potent inhibition of inflammatory responses. Although additive effects of MSCs and immunosuppressants may be expected in theory, the administration of MSCs during the use of immunosuppressants such as cyclosporine A (126) and dextametasone (127) has shown to disable the immunomodulatory functions of MSCs. This suggests that sufficient levels of cytokines or various paracrine factors involved in the micro-environment may be required for supporting MSCs’ functions. Therefore, either the application of MSCs in immunosuppressant-resistant patients or the timely administration of MSCs during periods of immunosuppressant tapering could be considered.