Adoptive transfer of cultured bone marrow stromal cells (mesenchymal stem cells also known as MSCs) is a promising new way to aid tissue regeneration and treat a wide variety of diseases where regulation of inflammatory responses is derailed. Although significant advances have been made in the field, pinpointing important mechanistic details about how MSCs function in vitro and in vivo, there are still many unanswered questions that need to be addressed before welcoming MSCs in the therapeutic arsenal of immune mediated diseases. In this viewpoint, we highlight and discuss a few factors that we believe are critical in terms of therapeutic success employing cultured MSCs. Selecting the right donor population, choosing the best culture conditions and picking the patient population that is most likely to give a favourable therapeutic response is just as important as considering interactions between MSCs and the combination of drugs in the recipient's body. Given the complexity of MSC–host interactions, it is also imperative to develop screening tools that account for as many variables as possible and predict precisely the in vivo response rates before MSCs enter the body. To achieve this, a multidisciplinary approach is required with comprehensive knowledge of basic MSC biology, immunology, pharmacology and good clinical practice.
Cell therapy using mesenchymal stem cells (MSCs) is an exciting new tool to modulate immune responses. In an inflammatory environment, this effect will help promote tissue regeneration. Unlike drug therapy where one or more targeted medications are used, MSC treatment offers a very different kind of therapeutic approach. MSCs secrete a large array of soluble molecules (cytokines, chemokines, growth factors, eicosanoids, nitric oxide, etc.) that act in concert to elicit a biological response. Not surprisingly, the communication between the injected MSCs and the host immune cells are bilateral. Disease (host)-specific inflammatory signals trigger the release of specific substances (appropriate to the actual host pathology) [1, 2]. This makes MSC therapy a completely new, interactive (‘smarter’) therapeutic approach as opposed to conventional drug therapy. Examples of various diseases that might be treated with MSCs (based on animal models or clinical trials) can be found in Table 1 [3-7].
Table 1. A summary of diseases that have been targeted by mesenchymal stem cell (MSC) therapy in preclinical models and clinical trials
Type of disease
GVHD, graft-versus-host disease.
Myocardial infarction, acute lung injury, liver cirrhosis, chronic wounds etc.
In the viewpoint essay by Dr Khosrotehrani and the Commentary by Dr Li et al., the current use of MSCs in various skin conditions is summarized and the possible therapeutic use of cell-free MSC juice is discussed [7, 8]. In the present essay, we will highlight some important additional issues that can critically influence the success of MSC therapy in immune mediated pathologies, including dermatologic diseases.
Heterogeneity of MSCs
When we use synthetic drugs, we expect them to deliver the same (or at least very similar) therapeutic effect every time they are used. There could be slight variations due to individual differences in drug pharmacodynamics and pharmacokinetics influenced by genetics and disease. However, most of the above variables can be adjusted by slight changes in dosing or timing. When MSCs are used, the first problem arises from donor heterogeneity – some donor MSCs can behave differently than others. The donor's age, gender, race, prior and current medical conditions and medications taken may profoundly change the in vivo pool of MSCs available at the time of bone marrow (BM) harvest [9-12]. On top of this, clonally derived MSCs will give rise to a heterogeneous stromal cell population, only a subpopulation of which might have immunosuppressive properties [13-15]. In most (if not all) clinical studies, it is this mixed population that is administered to the recipients, which is likely an important culprit in the large variation in success rates of clinical trials.
Different sample collection techniques may also contribute to the problem of MSC heterogeneity. MSCs can be obtained either from BM aspirates (containing a significant amount of peripheral blood which in our own experience can inhibit proliferation of stromal cells or from crushed bone fragments where MSCs are relatively enriched. The anatomic site of sample collection also matters. There may be significant differences even between sternal or iliac crest BM. The craniofacial skeleton is developmentally different from the axial skeleton, which means that MSCs isolated from the hip may behave differently than stromal cells isolated from a wisdom tooth or the jaw [16, 17]. This difference is evident when we compare growth characteristics of cultured stromal cells from these two anatomic sites, and these differences may also be reflected in the immune functions of these cultured cells.
The variety of culture conditions
Another critically important issue is how the stromal cells are grown, once obtained from the BM. MSC cultures can be initiated and expanded in basal medium containing FBS, pooled human platelet lysate, human serum or serum-free culture medium containing recombinant growth factors [18, 19]. Either of these methods can give sufficient number of cells in a few weeks, but the cells obtained do not necessarily possess the same biological properties. Cells grown in human pooled platelet lysate (HPPL) are known to have greater proliferative potential compared with cells propagated in FBS , and their immunosuppressive phenotype might be altered as well . Also, HPPL contains heparin, a xenogenic substance that has been shown to change MSC functions . Favouring the use of heparin-free serum-converted platelet lysate over HPPL can readily address this issue .
A not so negligible factor: the recipient
Although the biological characteristics of the injected donor cells are inarguably one of the most important factors that determine the efficacy of MSCs, we should not forget about the environment where the immunomodulation is supposed to take place.
There are a few reports suggesting that MSC treatment is more effective in paediatric patients than adults suffering from acute graft-versus-host disease (GVHD) [24, 25].Why does this discrepancy exist? What do children have that adults do not? We know that children are not small adults. Translating this to the language of immunology: children do not just have proportionally fewer immune cells, their immune system also behaves very differently when exposed to various stimuli. In comparison with adult cells, peritoneal macrophages isolated from children have been shown to produce a strikingly higher anti-inflammatory/proinflammatory cytokine ratio (IL-10/TNF-alpha) when exposed to lipopolysacharide (LPS) . It has also been reported that IL-10 producing regulatory T cells are much more abundant in the blood of infants than adults . These differences could explain why infants show better heart transplant outcomes than older recipients  and why the transplantation of cord blood cells results in lower incidence of acute GVHD after haematopoietic stem cell transplantation [29, 30]. So what about the effect of MSCs in children? Based on the above it is feasible that MSCs in an environment with more ‘regulatory capacity’ can achieve a greater immunosuppressive effect. It has now been widely accepted that injected MSCs will change the phenotype of macrophages from a pro-inflammatory (M1) to a anti-inflammatory (M2) macrophage [1, 5]. As long as the responder cells have the ability to recruit more regulatory cell types and produce more IL-10, the same number of injected MSCs may result in a more pronounced clinical effect. This also suggests that if we could prepare the recipient's immune system so that it becomes more ‘regulatory’ before infusing MSCs, we might be able to achieve a greater clinical response, especially in refractory cases. On the other hand, the need for recipient-derived regulatory cells for the effect of MSCs can make in vivo dose-response studies difficult. If the rate of generating new regulatory cell types plateaus, increasing the number of MSCs might not be able to deliver more immunosuppression.
Most patients who receive MSC therapy are already heavily medicated – mostly by immunosuppressive drugs – and they continue to receive many of these immunomodulators during and after the scheduled stromal cell infusions [31, 32]. Concomitant immunosuppressive therapy might affect in vivo MSC functions in a variety of ways. They can suppress the production of critical (recipient derived) pro-inflammatory signals that activate MSCs and kick-start the release of immunoregulatory molecules. TNF-alpha has been shown to induce production of eicosanoids in stromal cells, and inhibiting this stimulatory pathway will reduce the immunomodulatory capacity of MSCs . Commonly used TNF neutralizing drugs (like infliximab or adalimumab), as well as TNF decoy receptors (etanercept) block precisely this pathway and hence can possibly interfere with the beneficiary action of MSCs. Several other immunosuppressive drugs, although aspecifically, also reduce the concentration of TNF-alpha along with the production of other MSC-stimulating cytokines like IFN-gamma [33, 34]. In addition, circulating immunosuppressive drugs can also profoundly change the ability of MSCs to respond to stimulating signals. Several reports have shown that the release of MSC-derived prostaglandins is essential in MSC-driven immunosuppression, specifically suppression of lymphocyte proliferation and dendritic cell maturation [35, 36]. Prostaglandins released by the MSCs are also (at least partially) responsible for the generation of regulatory T cells, regulatory dendritic cells and immunosuppressive M2 monocytes/macrophages [5, 37]. Glucocorticoids – one of the most commonly used immunosuppressive drugs – are known to block arachidonic acid synthesis and thereby shut down the production of cyclooxygenase, a critical enzyme in the production of PGE2 . The same problem arises when patients are treated with non-steroidal anti-inflammatory drugs (NSAIDs) to manage pain or fever. Both non-selective COX inhibitors and COX2 selective drugs might prevent MSCs from making sufficient amounts of prostaglandins, thus decreasing their ability to interact with host immune cells. Other commonly used immunosuppressants like cyclosporine or rapamycin are also known to interfere with the synthesis of prostaglandins and nitric oxide in immune cells . This could explain why several studies have found that calcineurin inhibitors are either toxic to MSCs or simply antagonize their immunosuppressive affect. Interestingly, drugs are not always enemies of MSCs. Mycophenolate-mofetil either does not affect MSC function or it might even potentiate it [40, 41]. Antibiotics also have the potential to support MSC-driven immunomodulation by killing bacteria and providing access to toll-like receptor (TLR) ligands or other microbe-derived stimulating factors which initiate the pro-resolution programme of MSCs  (Fig. 1). Is this really so simple? Do we just have to map the effect of drugs on MSCs and assemble a list of medications that are OK to give with stromal cell therapy? Not surprisingly, the picture is more complex. The number of studies evaluating the effect of immunosuppressive drugs on MSCs is limited and mostly rely on in vitro data. Translating these results to the clinical setting is especially difficult. The complex effects of the in vivo environment on MSCs can change the way stromal cells respond to drugs, making them more or less sensitive to them. Also, MSCs in the patient are exposed to not only the originally administered drugs but also to numerous modified or conjugated drug metabolites created largely by the liver during detoxification. A problem, well-known to dermatologists who – based on in vitro assays – often struggle to identify the culprit drug behind drug hypersensitivity reactions. This complexity could explain the seemingly paradoxical observation made by a handful of laboratories that MSCs – contradicting in vitro data – work together with cyclosporine in vivo and achieve synergism in several models where immunosuppression is measured [43, 44].
The route of delivery
In the field of dermatology we often have the luxury to choose from topical, intralesional or systemic delivery of drugs . When administering MSCs to treat various skin conditions we will have the same choices. Chronic non-healing wounds could be treated topically. Discoid lupus, alopecia or morphea might be helped with intralesional injections, while treatment of widespread skin conditions could be handled by systemic infusions. The route of delivery will determine which cells MSCs will encounter first as it enters the patient's body. When deposited topically, they are exposed to keratinocytes, dermal fibroblasts and skin immune cells as well as the micro-organisms that are part of the resident skin flora. When injected into a skin lesion, MSCs will form a pocket within the dermis or in the subcutaneous tissue greeted by the local disease-specific inflammatory infiltrate. When given intravenously, MSCs will encounter and interact with all the blood components including circulating immune cells, the complement system and the elements of the coagulation cascade (Fig. 2). Clearly, the route of delivery gives access to different compartments of the patient's immune system in a special sequence, and this can influence how MSCs get activated and launch their anti-inflammatory programme.
Screening method wanted
One of the major challenges clinicians are facing is to predict the efficacy of the employed MSC therapy. A validated, reproducible and relatively easy method that correlates with clinical response rates would be invaluable in matching the right cells with the right patient population. Donor heterogeneity, recipient characteristics, accompanying drugs and the route of delivery can all affect the therapeutic outcome. The most accurate prediction is likely to be the one that accounts for all of these factors. Testing the donor MSCs against random, unrelated (not recipient derived) immune cells can give a clue of the immunomodulatory capacity of the stromal cells. Checking MSCs against the recipient's own immune system could help individualize this test and may give a better prediction. Performing this assay in whole blood, in the presence of drugs planned to be co-administered, could even better mimic the in vivo interactions that will take place after delivery of the stromal cells. What kind of assay should we choose to make these predictions, and what should be the read-out?
As skeletal stem cells make up a significant proportion of MSCs, one might be tempted to choose in vitro or in vivo stem cell assays to study differentiation potential of the chosen MSCs [46, 47]. Although these assays are considered gold standard in stem cell biology, they might be of little help when trying to estimate immunomodulatory capacity of a certain MSC batch. A relevant alternative could be a co-culture system, where donor MSCs are incubated with recipient immune cells and either inhibition of lymphocyte activation or the release of anti-inflammatory cytokines are measured. Easily accessible immune cells like PBMCs might be used for this purpose and IL-10 secretion could be measured with various methods like ELISA, ELISPOT or FACS analysis [48-50]. Why IL-10? Because it is the signature molecule of both regulatory T cell and M2 monocyte/macrophage-derived anti-inflammatory actions and it's level corresponds to an active immunomodulatory action as opposed to passive immunosuppression simply antagonizing the action of pro-inflammatory molecules like TNF-alpha [51, 52]. Using a ratio between circulating IL-10 (anti-inflammatory marker) and TNF-alpha (pro-inflammatory marker) might further increase the sensitivity of such an assay.
Conclusions and hypotheses
The lack of standardization, however, in donor recruitment, BM harvesting and culture conditions makes it very difficult (if not impossible) to interpret, compare and pool the clinical data from various trials. Interestingly, what seems to be a disadvantage can also be looked at as an exceptional opportunity. By choosing the right donor pool, selecting the right subpopulation of MSCs and applying the ideal culture methods, we might be able to create different cell collections tailored to treat specific diseases, or a specific group of patients. Reaching an international consensus and introduction of well-defined cell collection and culture methods may help standardization of cell products. Establishing an appropriate set of metrics (measuring growth factor composition of culture media, phenotypical changes in MSCs, etc.) could facilitate this process.
When haematopoietic stem cells are transplanted into a recipient, HLA-based matching of the donor and recipient cells is crucial . Ironically, when MSCs are transplanted, HLA matching is not needed, but an immunological match to predict efficacy seems to be necessary to utilize the MSCs to their full potential in this newly discovered unique cellular therapy. Experiments designed to measure the IL10/TNF-alpha secretion profile of disease-specific blood-derived immune cells against donor MSCs (in the presence of absence of certain immunosuppressive drugs) are recommended. Ultimately, these data need to be matched against clinical response rates in order to tell which patient population has the greatest chance to benefit from this new cellular treatment.
Clinical data points at the possibility of using MSCs to achieve immune-modulation in patients with a variety of immune/autoimmune problems. Picking the best cell product, matching the recipient host cells with the donor MSCs and carefully considering the actual in vivo characteristics of the patient will help us individualize MSC therapy and find the best combination for each individual. A step closer to personalized medicine.
I am thankful to Drs Eva Mezey and Jared Brown for editing the manuscript.
KN wrote the paper.
Conflict of interest
The author has declaired no conflicting interests.