Abstract
There is currently major interest to use mesenchymal stromal cells (MSCs) for a very diverse range of therapeutic applications. This stems mainly from the immunosuppressive qualities and differentiation capacity of these cells. In this review, we focus on cell therapy applications for MSCs in hematology. In this domain, MSCs are used for the treatment or prevention of graft-versus-host disease, support of hematopoiesis, or repair of tissue toxicities after hematopoietic cell transplantation. We critically review the accumulating clinical data and elaborate on complications that might arise from treatment with MSCs. In addition, we assume that the real clinical benefit of using MSCs for these purposes can only be estimated by a better understanding of the influence of in vitro expansion on the biological properties of these cells as well as by more harmonization of the currently used expansion protocols.
Introduction
M
MSCs were further characterized in the 1980s and 1990s by Caplan and Pittenger et al. and a specific phenotype was determined: MSCs express CD73, CD90, and CD105 and are negative for hematopoietic markers, such as CD14, CD11b, CD34, CD45, CD79, or CD19. This combination of cell surface markers was retained by the International Society for Cellular Therapy (ISCT) in their standards for MSCs [8 –10].
The cultured cells do not meet all the criteria for stem cell activity as there is, for example, senescence of the cultured cells. To remove the inconsistency between nomenclature and biological properties, the ISCT proposed, in 2005, to use the name mesenchymal stromal cells and to reserve the term mesenchymal stem cell for the true mesenchymal stem cells. Both, however, are abbreviated as MSCs. Since that position statement, most scientists have adapted this terminology [11].
Rationale for Using MSCs in Clinical Hematological Practice
There is a lot of interest to use MSCs for a wide variety of clinical applications. When introducing the search term “mesenchymal stem cells” in
MSCs have been introduced in the clinic fairly quickly after their characterization in the late 1990s [12].
The interest to use MSCs in hematology arises from their three characteristics: MSCs were found to have immunosuppressive capacities, they form the stromal microenvironment in which the hematopoietic stem cells (HSCs) reside, and they can give rise to different cell types. Their multipotent differentiation capacity is probably not limited to osteocytes, chondrocytes, adipocytes, and stromal cells since a variety of groups have described differentiation into other cell types such as pancreatic cells and muscle cells [13 –19].
Clinical Applications of MSCs
Treatment/prevention of graft-versus-host disease
The most commonly known and studied application of MSCs is their use in the treatment and/or prevention of graft-versus-host disease (GVHD). GVHD remains an important cause of morbidity and mortality after allogeneic stem cell transplantation (SCT) [20]. In 2004, a landmark article was published by Le Blanc et al., describing a complete response of steroid-refractory acute GVHD in a 9-year-old boy [21]. Four years later, data of a first phase II trial within the European Group for Blood and Bone Marrow Transplantation for the treatment of steroid-refractory GVHD were published by Le Blanc et al. These data showed a complete response in 30 of a total of 55 patients and improvement of symptoms in nine of them.
This study also showed that complete responders had a lower 1-year nonrelapse mortality and higher 2-year overall survival [22]. Subsequent studies confirm that responders to MSC treatment for steroid-refractory acute GVHD fare better than nonresponders [23 –28]. A recurrent policy among these studies is also the scheduling of repeated MSC infusions [22,23,25 –27,29 –41].
However, not all data reported on MSCs for the treatment of steroid-refractory acute GVHD are uniformly positive. Osiris therapeutics developed and patented an MSC formulation for IV administration as Prochymal® in the United States and is investigating its use in a variety of domains. They started with trials in GVHD treatment, but they have widened their scope and have opened trials in Crohn's disease, diabetes, chronic obstructive pulmonary disease, and acute myocardial infarction. They reported that in a large randomized trial on GVHD treatment with Prochymal®, they could not observe an improved overall response rate.
A subgroup analysis was performed and this showed a lack of improvement in patients with skin GVHD, the most common presentation of acute GVHD; however, there was significant improvement in patients with gut and liver GVHD. These data were disclosed in a press conference and an abstract was presented at the 2010 BMT tandem meeting [33]. In perception at the time, this was considered a negative result and fueled skepticism on the utility of MSCs in GVHD treatment. Given the overall encouraging results of those early trials, studies were also set up to test the role of MSCs in the prevention of GVHD and the treatment of chronic GVHD [25,30,34 –36,42].
Recently, the treatment of steroid-refractory GVHD in pediatric patients with MSCs (Prochymal) was reimbursed in Canada and New Zealand. Table 1 gives an overview of the clinical data of trials published on the use of MSCs for the treatment or prevention of GVHD.
N, number; pt, patient; aGVHD, acute graft-versus-host disease; cGVHD, chronic graft-versus-host disease; d, day; HCT, hematopoietic cell transplantation; CR, complete response; PR, partial response; MR, minor response; NR, no response; SD, stable disease; PD, progressive disease; FU, follow-up; PTLD, post-transplant lymphoproliferative disease; TRM, transplant-related mortality; OS, overall survival; CBT, cord blood transplantation; m, month; w, week; NA, not applicable; NR, not reported.
This clinical application relies on the immunosuppressive capacities of MSCs. First, MSCs do not express major histocompatibility class II and can therefore escape immune recognition by the host. This makes it possible to use third-party MSCs from unrelated donors, irrespective of human leucocyte antigens (HLA) compatibility [43]. This is advantageous since it takes a few weeks to generate a sufficient amount of cells from MSC cultures starting from fresh donor tissue. In addition to this escape mechanism, immunosuppressive properties have been attributed to MSCs nieuwe alinea vanaf. However, the mechanisms through which MSCs exert their immunosuppressive effects are not well understood as yet. Several studies dissecting MSC immunosuppressive qualities and capacities suggest that a variety of mechanisms are implicated. MSCs can suppress Th1 and Th17 cells and can induce regulatory T cells [44,45]. Moreover, they have also been attributed immunomodulating effects on B and natural killer cells. MSCs exert immunosuppressive effects not only through cell–cell contact but also through soluble factors, such as PGE2, prostaglandin B2, and IL10 [43,46 –49].
One soluble factor that has come to attention in recent years is indoleamine 2, 3 dioxygenase (IDO). MSCs can be activated by exposure to inflammatory cytokines, most notably interferon-γ (IFN-γ) and tumor necrosis factor-α (TNF-α).
Several groups have shown that IFN-γ and TNF-α-activated MSCs exert their immunosuppressive effect through upregulation of IDO. This increase in IDO is also implicated in the differentiation of monocytes in M2 immunosuppressive macrophages, thus enhancing the MSC immunosuppressive effect. The same group also described an IDO independent pathway of immunosuppression by IFN-γ-activated MSCs through the ligands of PD1 [50 –52]. Galectins are β-galactoside-binding proteins that have also been attributed a role in MSC immunosuppression. Galectin-1 and -3 are constitutively expressed on and secreted by BM MSCs. When the gene expression of these molecules is blocked simultaneously, an almost complete abrogation of the suppressive effect of MSCs on T cells is observed.
Ungerer et al. showed that galectin-9 expression by MSCs is upregulated by exposure to IFN-γ, resulting in suppression of both B and T cells. Moreover, galectin-9 expression levels appeared to correlate with the immunosuppressive qualities of MSCs [53 –55]. A better understanding of MSC working mechanisms spurs on to efforts to improve MSC therapies. For example, murine MSCs engineered to express CCR7 show improved homing to secondary lymphoid organs and they improved survival in a GVHD model while maintaining the graft-versus-leukemia effect [56].
Ex vivo expansion of hematopoietic stem/progenitor cells
MSCs might be a useful tool to optimize ex vivo expansion of HSCs, for instance, they secrete a number of hematopoietic cytokines [16]. MSCs are also part of the hematopoietic stem cell niche. The concept of the hematopoietic stem cell niche was first introduced in 1978: specialized niche cells are in close contact with HSCs and provide specific signals that help maintain their function. The niche is formed by a network of stromal cells, including MSC, CXCL12-abundant reticular cells, adipocytes, osteoclasts, osteoblasts, osteocytes, and neuronal cells. The location of HSCs in the niche is still controversial; however, most data suggest that they are mainly located in perivascular regions and in the highly vascularized endosteal region [57 –59].
MSCs in the BM are identified in vivo as nestin+ cells and are spatially associated with HSCs. Depletion of nestin+ cells in a murine model leads to a significant reduction of BM homing of hematopoietic progenitors. These nestin+ cells also have increased expression levels of HSC maintenance genes [57]. Efforts to recreate the BM niche in vitro with MSCs have shown that MSCs seeded on three-dimensional (3D) scaffolds and cocultured with hematopoietic progenitor cells (HPCs) resulted in increased migration of HPCs into the scaffolds and HSCs were localized in clusters in contact with the MSCs. Higher levels of molecules supporting hematopoiesis were found in such 3D coculture settings, showing maintenance of a pool of quiescent HSCs [60 –62].
It has been assumed recently that MSC-mediated hematopoiesis support could be improved by engineering MSCs to optimize their BM homing. This could be achieved by modifying the cell surface or by genetic engineering and using the cells as a cargo to deliver hematopoiesis-supporting cytokines in the BM stroma [63].
Ex vivo expansion of HPCs is mainly relevant in cord blood transplantation. Cord blood is a well-studied, alternative donor source for patients without a matched sibling or unrelated donor. HPCs in cord blood are more primitive and a lower cell number is required for SCT. However, the cell number of a single cord blood unit is usually not sufficient for transplantation of an adult patient. One way to obtain enough cells is to perform a double cord blood transplant. Another strategy might be ex vivo expansion of HPCs. Therefore, several strategies to expand cord blood HPCs have been tested and one of these strategies is the use of an MSC feeder layer [64].
Almost all data on this application for MSCs in hematology are preclinical, that is, either in vitro or in animal models. A concern about in vitro expansion of HPCs is the loss of long-term repopulating cells. Several studies have shown that coculture of MSCs with HPCs results in higher cell numbers and more long-term culture-initiating cells [65]. When these cells are transplanted in mouse models, enhanced engraftment is observed.
One study showed a decrease in GVHD in a double cord blood transplant setting where a group of mice was transplanted with a combination of an expanded and unexpanded cord blood unit and a control group receiving two unexpanded cord blood units. A significant survival benefit was observed in the mice receiving expanded cord blood, which was attributed to a decrease in GVHD incidence. The expanded units contained significantly more regulatory T cells compared with unexpanded cords and the authors claim that this is a result of expansion on an MSC feeder layer [66].
Robinson et al. reported in 2011 the first clinical data on cotransplantation of an unexpanded cord blood unit and another unit expanded on a layer of BM-derived MSCs from a family member [67]. Some data suggest that manipulation of HPC source through the removal of certain lymphocyte subsets improves ex vivo expansion on MSC feeder layers. It was found that removing CD3 and/or CD14-positive cells from the cord bloods before starting the expansion yielded higher numbers of total nucleated cells and CD34-positive cells [68].
Management of complications of (allogeneic) SCT
Poor graft function
As a result of the conditioning regimen and previous chemotherapy, the BM microenvironment can be damaged and this might result in a decreased support of hematopoiesis. Since MSCs can form the BM stroma and secrete hematopoietic cytokines, there is interest to use MSCs to enhance engraftment in SCT.
Graft failure can occur in any type of HCT and is associated with important morbidity and mortality [69,70]. The earliest report of a clinical application of MSCs dates back to 2000 when Koç et al. reported cotransplantation of MSCs and autologous SCT in breast cancer patients. In this study, a rapid neutrophil and platelet engraftment was observed [12]. In the setting of the autologous SCT graft, engraftment failure occurs less frequently than in the allogeneic setting. Nonetheless, it will lead to increased morbidity and mortality.
Recently, a study was published describing the treatment of 22 patients with MSCs versus MSCs+cord blood for graft failure after autologous SCT. Neutrophil engraftment was more rapid in the cord blood+MSC group (8 days vs. 18 days in the MSC group). However, this cannot be attributed to cord blood engraftment since chimerism at day +30 was 100% recipient for all subjects. The authors state that this might be due to stromal cells present in the cord blood units, enhancing the supportive effect on hematopoiesis of the BM microenvironment. Unfortunately, the study did not have a cord blood only arm, allowing examination of whether MSCs promote the effect of cord blood [71].
In allogeneic SCT grafting, engraftment failure is of concern, in particular, in cord blood transplantation and haploidentical SCT. In cord blood transplantation, cell doses are much lower than in peripheral blood SCT and greater HLA disparity is allowed; in haploidentical HCT, mega cell doses are used to overcome the greater HLA disparity. Several cotransplantation studies in these settings have been published so far. They show an improvement in engraftment and some also report a decrease of GVHD incidence [12,32,42,72 –83]. Table 2 provides an overview of the clinical data of studies using MSCs in the treatment/prevention of poor graft function.
N, number; pt, patient; aGVHD, acute graft-versus-host disease; cGVHD, chronic graft-versus-host disease; BM, bone marrow; CB, cord blood; AT, adipose tissue; PGF, poor graft function; d, day; h, hour; HCT, hematopoietic cell transplantation; MRD, matched related donor; (M)MUD, (mis)matched unrelated donor; haplo, haploidentical; HLA id, HLA identical; sib, sibling; PBSCs, peripheral blood stem cells; CBT, cord blood transplantation; PMN, neutrophil; plt, platelet; CR, complete response; PR, partial response; OR, overall response.
Tissue damage after HCT
The main area of interest to use MSCs in hematology remains HCT and its complications. In addition to GVHD and poor graft function, MSCs might also be of use in the treatment of tissue damage after SCT given their multipotent differentiation capacity. Studies have shown the benefit of MSCs in repair of cartilage or bone defects or damaged myocardium [84,85]. The tissue damage after SCT can be due to the toxicity of the conditioning regimen, GVHD, infections, etc.
Interest in this application for MSCs arose when some patients included in studies for GVHD treatment with MSCs showed remarkable repair of damaged tissues combined with several studies showing that MSCs preferably home to sites of tissue injury [8,21,86 –88]. In 2007, Ringdén et al. reported the results of a pilot study of 10 patients treated with MSCs for hemorrhagic cystitis, pneumomediastinum, or colon perforation as complications of allogeneic SCT. These data were promising and, moreover, MSC DNA could be detected in the bladder of one of the patients [89], indeed suggesting homing of MSCs to the injured tissues. However, additional and larger trials are needed to corroborate these findings.
Caveats in Using MSCs for Clinical Purposes
Suppression of the graft-versus-tumor effect?
Allogeneic SCT relies on two pillars for its therapeutic effect: disease eradication by pre-SCT treatment and the conditioning regimen (on the one hand) and alloimmune activity called the graft-versus-tumor effect (on the other hand). Since potent immunosuppressive capacities have been ascribed to MSCs, there is some concern in the use of these cells for the treatment of complications of allogeneic SCT. In 2008, Ning et al. [90] reported results of an open label, randomized clinical trial of HPC or HPC+MSC transplantation in patients with hematological malignancies.
The authors saw a significantly lower incidence of GVHD, but this came at a cost of significantly higher relapse rates in the cotransplantation group (60% vs. 20%), resulting in a significantly lower disease-free survival in this group (60% vs. 66.7%) [90]. Many—if not all—studies on the use of MSCs for treatment of late GVHD have also included relapse incidence as an endpoint, and so far in these studies, no significant increases in relapse were reported. However, to be conclusive, a large randomized controlled trial would be required.
Increased risk of infections
Since MSCs suppress immune responses, one might expect that treatment with MSCs increases infectious complications [31,91]. In most recent studies on therapeutic applications of MSCs, infectious complications are reported. However, in the setting of GVHD treatment or poor graft function, there is already an increased incidence of infectious complications due to prolonged neutropenia, impaired immune recovery, and prolonged immunosuppressive treatment. This makes it difficult to determine the impact of MSC treatment on the incidence of infections.
There has been concern that MSC treatment leads to increased cytomegalovirus (CMV) infections. Lucchini et al. addressed this problem and retrospectively analyzed frequency and severity in 24 patients treated with MSCs in their center. They could not find an increase in viral infections or in mortality compared with a historical control group [92]. Earlier this year, Calkoen et al. [93] reported on viral complications in a pediatric patient population treated with corticosteroids, MSCs, or another second-line immunosuppressive therapy. They could not find an increased incidence of CMV, Epstein-Barr virus (EBV), or adenovirus infections in the MSC-treated population; however, MSC treatment correlated with worse survival in adenovirus infection [93].
Tumor formation/malignant transformation of MSCs
When pluripotent stem cells, such as embryonic stem cells or induced pluripotent stem cells, are injected in mice they will form teratomas [94]. Of course formation of ectopic tissue after injection of MSCs has also been a concern. MSCs might not only enhance tumor growth or disease progression in a subject but the MSCs itself might transform into a malignant cell as well. This has been shown in murine MSCs both in vitro and in vivo [95 –98]. There have been reports on malignant transformation of human MSCs in vitro; however, they were retracted several years later and attributed to contamination with tumor cell lines.
DNA fingerprinting of the cells believed to be transformed MSCs showed that they did not have the profile of the original MSCs, but their profiles were comparable with a fibrosarcoma and osteosarcoma cell line in one laboratory and human glioma cell lines in the other laboratory, proving cross-contamination of these MSC cultures with tumor cell lines and disproving spontaneous transformation of MSCs in culture [99 –102].
Because of these reports, several groups have addressed this issue. Since MSCs require ex vivo expansion to obtain sufficient cell numbers for therapeutic applications, one could suppose that under the proliferative strain of the culture conditions, MSCs might acquire cytogenetic abnormalities predisposing to malignant transformation of the cells. Therefore, most groups concentrated at first on cytogenetic stability of MSCs in culture. Indeed, several groups found already at quite early stages of the cultures (passage 3/4 cells) as well as in later passage MSCs that there were cytogenetic abnormalities. However, in in vivo models, these cells did not form tumors, nor could malignant transformation of these cells in vitro be observed. As expected, MSC cultures showed evidence of senescence, regardless of cytogenetics, while telomere length shortened and no evidence of increased telomerase activity could be found in these MSCs [103 –105].
The most recent studies using MSCs for therapeutic applications often include evidence for malignant transformation as an endpoint. So far, in the results of clinical trials in hematology that have been reported, there is no incidence of malignant transformation of MSCs in patients treated with MSCs. In 2011, Centeno et al. published an update on safety and complications of MSC implantation for orthopedic interventions. They followed 339 patients with magnetic resonance imaging and could not demonstrate evidence of malignant transformation [106]. von Bahr et al. examined tissues from patients treated with MSCs at autopsy. They could not find any ectopic tissue formation. They attribute this to the fact that there does not appear to be a sustained engraftment of MSCs since levels of MSC donor DNA were negatively correlated with time since MSC infusion [107].
As discussed above, there is as yet no evidence of malignant transformation of human MSCs. Malignant transformation appears to be restricted to murine MSCs and occurs after prolonged expansion. In our study, with murine MSCs, we observed evidence for transformation only after multiple (8–9) passages [95]. However, caution with human MSCs is warranted and malignant transformation should remain a point of attention in all clinical trials with MSCs.
Fetal bovine serum
To expand MSCs in culture, a growth supplement has to be added to a basal medium. The first groups describing MSC expansion used fetal bovine serum (FBS) in varying concentrations for this purpose [1,8]. However, this poses some challenges: there is of course the danger of transmission of pathogens (bacterial, viral, mycoplasma, prion); additionally, the components are not defined and will vary with different batches, resulting in variable efficacy in cultures. FBS proteins can also give rise to xenogeneic reactions, resulting in allergic infusion reactions in case of repeated infusions, and the presence of antibodies to FBS can also threaten viability of MSCs after administration [108 –110].
Efforts are ongoing to find alternatives: the most promising candidate is the replacement of FBS with platelet lysate (PL), thus humanizing the culture medium [110 –112]. In recent years, an increasing number of clinical studies with MSCs expanded in a culture medium supplemented with PL are performed [28,35,41,113]. However, PL still does not resolve all the issues with FBS: there is still a risk of pathogen transmission and PL batches also have a variable quality. A satisfactory serum-free medium, however, has not been developed yet.
Caveats in the Interpretation of MSC Clinical Trials
The use of MSCs in a wide variety of clinical domains is investigated by a great number of research groups. In this review, we focused on the application of MSCs in the field of hematology. Since they were first derived from BM, a logical first step was to investigate their potential use in this particular domain. They have certain characteristics that make them promising tools for regenerative medicine (restoration of the BM microenvironment) and immunotherapy. In vitro data confirm these expectations; however, clinical trials are not uniformly positive [26,33].
It should be noted that MSCs have been moved quite rapidly from the bench to the bedside. They were already described in the 1970s by Friedenstein, but it was just in the early 1990s that Caplan and Pittenger et al. published further characterization of these cells and described the mesengenic process. Already in the late 1990s, patients were treated with MSCs [12]. Working mechanisms of MSCs were and are not fully understood, although we are gaining more and more insight. The evidence is accumulating in in vitro and in vivo studies that MSCs can be a valid treatment option in a number of hematological conditions [44,114 –117].
The number of clinical trials with MSCs increases further. In hematology, the majority of clinical trials with MSCs focus on prevention or treatment of GVHD and poor graft function.
However, it is difficult, if not impossible, to compare the data from these different trials since many parameters are highly variable. Tables 1 and 3 give an overview of the clinical variables and outcome of trials with MSCs in GVHD and support of engraftment. Tables 3 and 4 list the variable parameters of the cellular products used in these trials. These vary not only between different trials but often also within one trial.
P, passage number; BM, bone marrow; CB, cord blood; AT, adipose tissue; HLA id sib, HLA identical sibling; FBS, fetal bovine serum; PL, platelet lysate; MRD, matched related donor; MMRD, mismatched related donor; NR, not reported; MSCs, mesenchymal stromal cells.
P, passage number; BM, bone marrow; CB, cord blood; AT, adipose tissue; FBS, fetal bovine serum; PL, platelet lysate; d, day; HSC, hematopoietic stem cell; haplo, haploidentical; HLA id, HLA identical; sib, sibling; NR, not reported.
To obtain a sufficient amount of MSCs for clinical applications, MSCs have to expand ex vivo. The source of MSCs differs as most groups start MSC cultures from BM or cord blood, while a few groups use adipose tissue-derived MSCs. In some studies, even though MSCs are all derived from BM, the source of BM is different between patients. In the pivotal trial of Le Blanc et al., for example, MSCs are derived from BM from the HLA-identical sibling, a haploidentical donor, or a third-party (unrelated) donor [22].
Although MSCs can escape immune recognition, there could be a difference in clinical efficacy and MSC survival after systemic administration when the MSC and HPC donor is the same. The group from Karolinska Institutet in Sweden has addressed this in a follow-up report of patients treated in their center with MSCs for GVHD or hemorrhagic cystitis and could not correlate the MSC source (related or unrelated) with response [118].
For the ex vivo expansion of MSCs, variations of the protocol originally described are used. The most obvious difference is the replacement of FBS most commonly by PL. One group expanded MSCs in serum of the donor [119]. The number of cells seeded to start a culture and after passaging is highly variable. Again, in most of the trials, patients included in the same trial receive MSCs cultured for a different length of time and thus after different passage numbers. Often duration of expansion is a function of the cell number needed and obtained.
Although all these culture conditions render cells with the MSC characteristics, spindle-shaped morphology, the combination of phenotypic markers, and at least trilineage (adipocytic, osteocytic, chondrocytic) differentiation potential, it is becoming increasingly clear that biological characteristics are influenced by the culture conditions [120 –123]. In 2012, the group of Le Blanc analyzed the role of passaging on the clinical outcome in patients treated with MSCs in their institution between 2002 and 2007 and found that patients treated with early passage MSCs (1–2) had a significantly better response rate and overall survival than those treated with MSCs cultured for three or more passages [118]. What appear to be small details therefore might have a significant impact on the effector functions of the cells.
In addition to the differences in expansion protocols of MSCs, published studies also differ in other aspects. The infused number is highly variable, with cell numbers varying between studies from 0.3 to 10 million MSCs/kg recipient weight MSCs [42,77]. However, even in the same trial, the infused cell number can differ up to 10-fold among the different patients [23]. In most studies, patients receive 1–2 million MSCs/kg recipient weight; however, this appears to be a rather arbitrary number, and again in a number of studies, the infused cell number depends on the number of cells obtained after expansion.
There are almost no data in literature on dose finding for MSCs in the treatment of poor graft function or tissue repair after allogeneic SCT or to obtain a maximal effect on GVHD. One study with Prochymal® compared two dose levels: 2 and 8 million/kg. Since there was no difference observed, the remaining patients were treated with 2 million/kg. The total dose of MSCs used also depends on the number of infusions planned. In some trials, repeated infusions, for example, biweekly, are administered irrespective of the response, and in other trials, repeated infusions are planned as a function of response [22,25 –41,47,113].
The timing of infusion is quite constant in trials for the support of engraftment. MSCs are usually infused 4–6 h before HCT. However, in GVHD trials, not only the infusion time points of the first MSC infusion but also between repeated infusions are highly variable, again not only between different studies but also in the same trial (Table 1). The variability between repeated infusion stems from the fact that they are often added to improve response or in case of relapse of GVHD. Logically, patients might benefit from early treatment with MSCs; however, given the rather small patient numbers in trials reported as yet and the variety in number of treatment lines [1 –8] between patients included in these studies, this cannot be assessed.
The majority of studies in hematology with MSCs investigate their role in the treatment of GVHD based on their immunosuppressive qualities. There are commonly used release criteria for MSCs: phenotype, differentiation capacity, and microbial screens. However, there is no standard or quality control for the immunosuppressive capacities of culture-expanded MSCs. Different batches of MSCs are not only used in different studies but also in the same trial, and MSCs cultured with different batches of FBS or PL might have different immunosuppressive qualities. In recent years, researchers are trying to develop and define such a quality control system. Nazarov et al., for example, devised an in vitro assay to evaluate and quantify immunosuppression of expanded MSCs by using murine clonal T cells, determining proliferation rate, cell surface marker expression, expression of mRNA of transcription factors, and cytokine secretion expression [124].
Issues have been raised concerning cryopreservation of MSCs. This procedure is widely used because in vitro expanded MSCs can be stored in cell banks, allowing to offer cryopreserved MSCs as an off-the-shelf product. Moreover, MSCs have to be expanded in vitro over several weeks to obtain a sufficient cell number, whereas treatment is often required as soon as possible in these clinical settings. Some authors have suggested that cryopreservation hampers the immunosuppressive qualities of MSCs [125], while others claim that cryopreservation does not interfere with MSC function [126 –128].
Finally, there are some general remarks on the studies published. Often, the number of patients included in the different studies is small; this renders the interpretation of the results even more difficult given the variability of parameters between patients included. There is also a lack of randomized trials. One large placebo-controlled trial has been performed by Osiris with Prochymal®. Results have only been published as an abstract since there was no difference in overall outcome [33]. Most groups use a historic control as reference.
Since MSCs in hematology are mainly studied in the niche, that is, hematopoietic cell transplantation, it follows naturally that relatively small patient populations will be eligible for trials on the use of MSCs. As shown in Tables 1 –4 and discussed above, there is a great variability between different trials, rendering interpretation of data difficult. Therefore, a more rigorous study design for future studies is necessary to be able to examine the therapeutic potential in well-defined settings.
In the development of new drugs, these compounds have to prove themselves in trials against a standard or placebo, MSCs should also be treated as such. Already, in in vitro settings, MSC efficacy can be compared with placebo or negative controls. It is only logical to then go on and use these cellular products that have proven effect in vitro in large, multicenter randomized trials. Such a trial will be easier to set up in cotransplantation studies in cord blood transplantation or haploidentical HCT for support of engraftment. In GVHD, where there is no golden standard treatment for steroid-refractory disease, such a trial might be more difficult to design, but efforts should also be made in these trials.
The scientific community working with MSCs is well aware of these issues and efforts are being made to obtain more uniform expansion protocols, as is proven by recent articles describing standardized, good manufacturing practice (GMP)-approved expansion protocols and the development of appropriate quality controls [124,129 –135].
Conclusion
In conclusion, we can state that considerable progress has been made over the last 20 years since MSCs were first introduced in clinical applications. Evidence from both in vitro and in vivo data is accumulating regarding MSCs as a valid treatment option in a number of hematological conditions.
MSCs can escape immune response and they can be banked and are thus easily accessible as an off-the-shelf product. Several questions regarding the working mechanisms of MSCs remain unanswered, but we are gaining more and more insight. With this growing knowledge about the biology of MSCs, efforts can be made to optimize the cells for therapeutic purposes or their use as targeted therapeutics.
MSCs have been rushed from the bench to the bedside, but it would be wise to take a step back to the bench again. Clinical trials with MSCs can only benefit from a better understanding of MSC working mechanisms and definition of optimal and standardized culture conditions. Moving forward, from the bench to the bedside, clinical trials should be designed rationally and rigorously both in MSC-specific parameters as in treatment schedules. With rigorous study design, patient groups most likely to benefit from MSC therapy can be identified and an optimal treatment schedule can be defined, allowing MSCs to find their way to daily clinical practice in hematology.
Footnotes
Ackowledgments
The authors would like to thank Miss Nicole Arras for assistance in preparing the manuscript. A.D.B. received a PhD fellowship from the Vrije Universiteit Brussel (Horizontale onderzoeksaktie).
Author Disclosure Statement
No competing financial interests exist.