Acute graft-versus-host disease (aGVHD) is encountered after allogeneic hematopoietic stem cell transplantation (HSCT) and is a systemic illness manifested by skin rash, gastrointestinal (GI) inflammation presenting with nausea, anorexia, diarrhea, and hepatic dysfunction.1 Clinical outcomes are influenced by the extent of GVHD with overall survival directly correlated with maximal GVHD score.2 The underlying biology of aGVHD is a systemic inflammatory process that occurs across temporal and spatial boundaries.1 Multiple cell populations have been implicated, including various T-cell subsets, both host and donor dendritic cells, B cells, NK cells, mast cells, and underlying host endothelial cells. Multiple cytokine mediators have been implicated, including IL-1, IL-2, IL-12, TNF-α, IFN-γ, TGF-β, IL-17, and IL-23 as well as critical adhesion and costimulatory interactions that are involved in the process. However, the molecular epicenter of the disease is the T-cell receptor, where adoptively transferred donor T cells will react via their receptor to polymorphic peptides presented within the cleft of HLA major histocompatibility complex (MHC) Class I or Class II molecules. The critical need for time and amplification of aGVHD has been elegantly studied and modeled by Ferrara and colleagues.1 In their proposed model, the conditioning regimen used for the transplantation process also plays a critical role in systemic aGVHD in which tissue damage will contribute by activation of the innate immune system, possibly due to bacterial translocation across an injured small intestinal epithelial wall with lipopolysaccharide triggering the innate immune system. As a consequence, a host of inflammatory signals are amplified by the adoptively transferred mature donor T cells reacting to target MHC antigens. This cascade, coined “cytokine storm,” contributes to the systemic inflammatory changes associated with aGVHD. Mesenchymal stromal cells (MSCs), a ubiquitous cell resident within the marrow and a critical component of the marrow microenvironment, have been proposed as a therapeutic cellular product that could be used to treat GVHD.3, 4 They may aid in tissue repair by transdifferentiation or fusion events. MSCs also inhibit the T-cell proliferative response and dampen the innate immune response. MSCs have the capacity to home and traffic to areas of inflammation. However, the impact of MSCs in aGVHD may be potentiated as they appear to function in a MHC-unrestricted manner, and the lack of a variety of surface ligands may provide them immune sanctuary. Thus, MSCs may potentially impact the cytokine storm cascade by influencing and dampening both innate and adaptive immune responses. MSCs have been utilized in clinical trials since the mid-1990s when autologous MSCs were first grown ex vivo and administered to patients with breast cancer undergoing autologous stem cell transplantation, with the principal goal of augmenting hematopoietic engraftment.5, 6 The earliest studies suggested that there was some improvement in engraftment, and these naturally led to studies of the application of MSCs to the allogeneic transplant setting. In the original multicenter study of allogeneic MSC use, published by Lazarus and coworkers,7 HLA-identical, sibling, culture-expanded MSCs were administered concomitantly with HSCs for patients undergoing HSCT for hematologic malignancies. This study represented the first application of ex vivo expanded, related-donor MSCs in human allogeneic HSCT. Safety was established in this study with no infusional toxicities or late MSC-associated adverse effects identified. There was no deleterious impact on engraftment but there was no clear evidence of enhancement either. No increased relapse rate of the underlying hematologic malignancy was identified. However, there was a suggestion of a trend toward less aGVHD. All patients received a modified two-drug GVHD prophylaxis with cyclosporine and only 3 days of methotrexate administered at standard doses of 15 mg/m2 on Day 1 and 10 mg/m2 on Days 3 and 6. Of 46 patients included in this study undergoing myeloablative matched-related donor transplantation, 50% had no aGVHD, 13 developed Grade II to IV GVHD, and seven had Grade III or IV aGVHD. Of 42 patients evaluable for chronic GVHD (cGVHD), 14 patients developed limited and eight developed extensive cGVHD. Since aGVHD rates observed in this study were equivalent to patients receiving four doses of methotrexate as a more standard prophylaxis schedule, it was questioned whether the MSCs may have provided some benefit. Frassoni and coworkers8 performed a matched-pair analysis of patients who were treated on the original Phase I study with a cohort from the EBMT. Their conclusions were that cotransplantation of expanded MSCs with HSCs were safe, had improved disease-free survival and overall survival, and did not enhance relapse and also that the incidences of both aGVHD and cGVHD were reduced. They did conclude that randomized trials were critically needed in this field. In the United States, continued efforts for assessing the role of allogeneic MSCs were mostly outsourced to industry, led by Osiris Inc., which began to develop a universal MSC donor approach. In Europe, MSC therapies became the purview of academic institutions. In what is now a landmark publication in the field, Le Blanc and coworkers9 reported a young patient with severe steroid-refractory gastrointestinal and hepatic aGVHD who had failed to respond to calcineurin inhibitors, PUVA, infliximab, and daclizumab. He had apparent resolution of both diarrhea and jaundice after being treated with marrow-derived haploidentical MSCs. With clinical improvement, calcineurin inhibitors were discontinued. However, he had a relapse of his hepatic and gastrointestinal aGVHD, which responded to retreatment with haploidentical MSCs from his mother, was successfully weaned from corticosteroids, and was maintained on calcineurin inhibitors long term. Correlative science studies of interest included the identification of female cells within the gastrointestinal tract as determined by chromosomal fluorescence in situ hybridization studies. These observations suggested that MSCs may be able to traffic to the GI tract or other sites of inflammation. Several subsequent pilot studies and two large Phase II studies were encouraging for the efficacy of MSCs for treatment of aGVHD.10-13 Le Blanc and colleagues14 reported an EBMT registry series where 55 patients received related or unrelated donor MSCs. Thirty of 55 patients had a complete response (CR) and nine had partial responses. Kebriaei and coworkers15 reported the data of a Phase II study sponsored by Osiris Inc. using allogeneic, unrelated, universal-donor MSCs. Not all patients on the study received product from the same donor, but similar release criteria and affirmation that cell surface phenotype were consistent with MSCs were achieved.16 In this study, 31 patients were treated, of whom 24 patients had a CR (77%) and five of 31 were reported as partial responses (16%). Of the patients who achieved CR, 19 of the 24 had sustained CR for 90 days without the need for addition of second-line therapy. There also was a suggestion of a dose response. Several other smaller Phase I and/or II GVHD treatment trials, both in de novo aGVHD and in steroid-refractory settings have demonstrated infusional safety and varying degrees of clinical response (reviewed in Newell et al.3 and Kebriaei and Robinson4). Following up on these promising results, Osiris Inc. performed two Phase III studies in steroid-refractory and newly diagnosed aGVHD with their universal-donor product, Prochymal.17 The first study in steroid-refractory patients included all patients with Grade B to D aGVHD who demonstrated no improvement after 3 days of corticosteroid therapy and/or had duration of not more than 14 days while on at least 1 mg/kg steroid therapy. Their primary endpoint was a complete remission sustained at at least 28-day duration. MSCs were administered twice weekly for 4 weeks, and if there was a CR or the patient was a nonresponder, therapy was discontinued on Day 28. For partial or mixed responders, four more weekly doses were given. A Phase III trial for MSCs in newly diagnosed aGVHD was also performed, with the primary endpoint being complete remission followed by 28 days of sustained response without steroid increase, with no second-line therapy, and followed for a 90-day survival endpoint. The treatment schedule was MSCs administered twice weekly for 2 weeks and then weekly for 2 additional weeks. The results of these studies have since engendered great attention, both negative and positive. Both Phase III trials failed to demonstrate a benefit in the primary endpoint. On an intent-to-treat analysis, 87 patients were treated with placebo, with 26 CRs identified (30%) versus 173 patients treated with MSCs with 60 CRs identified (35%; p = 0.3). These studies are yet to be published, but data from the abstracts detailing study outcomes18, 19 suggested significant improvement in gastrointestinal and hepatic endpoints. For patients with steroid-refractory hepatic disease, there was 76% response versus 47% in patients receiving placebo. Similarly, for steroid-refractory GI disease, 82% response in MSC cohort, and 68% with placebo (p = 0.03). For patients with three-organ aGVHD involvement, overall response rates with MSCs were 63% versus 0% with placebo. In pediatric patients, an overall response rate of 64% was noted in patients receiving MSCs versus 36% in those receiving placebo.19 The 100-day survival was also improved, with 79% for MSCs and only 50% for those receiving placebo. Additional observations were that there was no evidence of dose response, that there did appear to be improved hematopoietic engraftment, and that there was no increase in relapse or infectious morbidity. The reported failure to reach primary endpoints has led to significant scrutiny of the design of the clinical trial and an examination of the true role of MSCs in aGVHD. The study primary endpoint selection and lack of study specific guidelines for immune suppression after a therapeutic response have both been criticized.20 A thoughtful review by Galipeau21 suggested that the selection of the MSC product may also have contributed dramatically to the trial failure. Specifically, there can be donor variance in IFN-γ responsiveness of the MSC product and that this can influence the functionality of the product. With continued passage (as is experienced in the prolonged manufacture of MSCs, rather than using early-passage MSC products, as was used in the European studies), it is conceivable that epigenetic reprogramming occurred during continued passage that leads to loss of efficacy of the product. Similarly, it was hypothesized that senescence can occur with ongoing product passage, contributing to a loss of therapeutic efficacy. It is also possible that the immunogenicity of the allogeneic product could have contributed to more rapid clearance. Finally, Galipeau suggests that based on his research studies, which demonstrate that fresh cultured MSCs have higher production of protein products than freshly thawed, previously cryopreserved MSC products, recognizing that MSCs that are cryopreserved, that at least 24 hours in culture are needed for full reexpression of their protein profile.22 In the absence of other Phase III trials, a meta-analysis to assess benefit of MSCs in steroid-refractory aGVHD has been published. Chen and coworkers23 reviewed results from 13 eligible studies of GVHD and MSCs. Their detailed review identified that MSCs used in the studies were typically heterogeneous products, and they recognized that it could be difficult to draw conclusions across studies. However, their findings demonstrated 1) a trend toward higher CR in children (p = 0.05); 2) complete remission was enhanced in skin aGVHD versus GI and liver (p = 0.04), although the overall response rates were equivalent; 3) complete remission was enhanced in Grade II versus III and IV (p = 0.02), although the overall response rate was equal. They demonstrated no difference whether the MSCs were grown in fetal calf serum versus platelet lysate (p > 0.05) and that the overall response rates in children and adults were equivalent (p = 0.51). Limitations of the meta-analysis include a paucity of randomized control trials, a small number of patients in some of the independent Phase II trials, no standard MSC passage number for the therapeutic product in different studies (usually one to five passages), and finally, as described above, no consistency in the management of concomitant immune-suppressive therapy. The Osiris Inc. MSC product line has been acquired by Mesoblast Inc. and Mesoblast reportedly plans to engage with regulatory authorities for a new, more targeted Phase III study to be performed in the subset of patients with liver and lower gut aGVHD.24 In the meantime, Prochymal has received conditional approval in Canada and in New Zealand, for the treatment of children with acute steroid-refractory GVHD. The product also remains available in the United States under an expanded access program for treatment. In an open-label pediatric trial for Grade B to D steroid-refractory GVHD reported by Kurtzberg and coworkers,25 75 patients, aged 2 months to 17 years, were treated with MSCs that were manufactured from seven independent stromal marrow donors. Their data reported 61% responders with 26% CR with gastrointestinal GVHD, 44% CR with cutaneous GVHD, and 33% CR in hepatic GVHD. For patients who responded and had responses persist to Day 28, the 100-day survival was 78% versus 31% for those who failed to respond, sustained at Day 28. Similarly, Ball and colleagues26 reported 37 patients with steroid-refractory Grade III and IV aGVHD pediatric patients and showed that long-term outcomes were favorable for responders. In children with sustained remission of the GVHD, for patients who achieved a CR, there was a 65% projected 3-year survival, but for patients with PR or nonresponders, that projected long-term survival was nil. Their study included MSCs obtained from related or unrelated donors and with only short-term passage of two to three individual passage events. There has been less focus on aGVHD prophylaxis with MSC therapies. The original study of Lazarus and colleagues remains the largest published study but in five other small Phase I and II clinical trials, MSC cotransplantation for prophylaxis was found to be safe and feasible.27-31 Only one study of 13 patients suggested an observed lowered degree of Grade 3 to 4 aGVHD (p = 0.05) while another study by Ning and colleagues30 of 30 patients, of whom 10 received MSC, there was a lower incidence of Grade II to IV aGVHD but a higher risk of relapse (p = 0.02) in MSC recipients compared to controls. In this setting, Maziarz and coworkers32 recently reported a large multiarm Phase I study assessing the potential efficacy of multipotent adult progenitor cells (MAP-C; Multi-Stem) in GVHD prophylaxis. The MAP-C product is a marrow-derived adherent stromal cell population with extensive in vitro expansion capacity. Studies of this product demonstrate proliferation without senescence and superiority in potency assays over MSCs in degree of immune suppression achievable as well as in karyotype stability.33 The Phase I trial included 36 patients, all of whom received the same, ex vivo expanded, universal-donor stromal stem cell product. The trial design included a single dose escalation of the MAP-C product administered on Day 2 after myeloablative transplantation at doses of 1 × 106, 5 × 106, and 10 × 106/kg as well as a repeat-dose escalation designed to cover the first 28 days of a transplant course. Repeat-dose escalations used doses of 1 × 106 and 5 × 106/kg MSCs administered for three full doses (over 2 weeks) and then weekly over 4 weeks with administration of five full doses. The primary goal of the study was to assess infusional and regimen related toxicities at 30 days with secondary endpoints of aGVHD incidence, overall survival and infection at 100 days. Patients (median age, 52 years) included those with acute myeloid leukemia, acute lymphocytic leukemia, chronic myeloid leukemia, and myelodysplastic syndromes. There was no infusional or drug-related toxicity reported over 30 days after treatment and no drug-related serious adverse event over 100 days. Neutrophil engraftment occurred at a median of 15 days (range, 11-25 days). Thirty of 36 patients completed the Day 100 evaluation with three patients experiencing relapse and four patients experiencing nonrelapse mortality. Both the highest single dose and the highest repeat dose were well tolerated. The overall Grade II to IV GVHD incidence was 38% and Grade III and IV was 15%. In the cohort that received 10 × 106 MAP-C/kg administered on Day 2, there was 11% total Grade II to IV with 0% Grade III and IV aGVHD. Currently, a Phase II/III aGVHD prophylaxis trial is being developed to confirm these observations. Orphan status has been granted for aGVHD prophylaxis with the MAP-C product, and FDA Fast Track status has been granted. The current state of MSC utilization for the prophylaxis and treatment of aGVHD and cGVHD remains anecdotal at best. There are other interesting observations that culture conditions may vary the response as shown in a study by Lucchini and colleagues,34 where platelet lysate was superior to xenogeneic serum to facilitate MSC growth. Arterial infusion of MSCs has appeal as it would directly target the MSC to the afflicted organ and studies of arterial infusions have demonstrated feasibility.35, 36 Clearly, the issue of passage number of the MSCs has been raised, and there are suggestions that use of early passage MSCs may have greater efficacy.37 Better understanding of the MSC secretome and transcriptome, as well as the role of exosomes, will likely influence future product selection.38, 39 Finally, it is clear that MSC therapy will likely be expensive as the cost of goods and manufacturing, particularly of donor-directed products, will be far more costly than single production of small molecule therapeutics. Cost-effective and comparative effective analyses will be needed in the long term. How would we design an aGVHD study today?40 What lessons have we learned from these recent trials? First, that aGVHD is a disease of T-effector cells and that prophylaxis, new onset, and steroid refractory actually represent a continuum of clonal expansion of end organ targeting T cells. We believe that MSCs can immunomodulate the adoptively transferred donor immune system as well as secrete factors that can facilitate tissue repair and modulate the inflammatory cytokine release of cellular constituents of the innate immune system. Many questions remain to be answered in the performance of MSCs after transplants, including optimal timing for application, optimal dose, route of delivery, and whether simultaneous cytokine support to activate or enhance the MSC clinical effect. Other outstanding questions include whether concomitant immune suppression will be required to protect the MSC product, whether frozen versus fresh MSC infusions will provide better efficacy, and how finally to better understand measures of clearance of the MSCs to best guide redosing. Most importantly, careful consideration and determination of meaningful, achievable endpoints in well-designed, controlled randomized trials are needed that can clearly isolate the effects of the MSCs from those of concomitant care. RM has received research support from Athersys, Inc.