Mesenchymal stem cells

Abstract

About 40 years ago Friedenstein described stromal cells in the bone marrow that were spindle shaped and proliferate to form colonies. These cells attach to plastic and are able to differentiate under defined in vitro conditions into multiple cell types present in many different tissues, e.g. osteoblasts, chondroblasts, adipocytes, etc. Later on these cells, obtained from postnatal bone marrow, were called mesenchymal stem cells (MSC) or stromal stem cells. Recently the presence of somewhat similar cells has been demonstrated in many other tissues too. In spite of extensive attempts to characterize these cells we are still lacking definitive in vivo markers of MSC although retrospective functional data strongly support the existence of common adult stem cells that have the capacity to differentiate along various specific differentiation lineages. Since MSC can be rather easily isolated from the bone marrow and can also be expanded in vitro they have become a prime target for researchers of tissue regeneration. These cells have now been extensively used for transplantation experiments in animals and also for some therapeutic trials in humans. However, much new research is needed to learn enough on the molecular mechanisms of MSC differentiation to evaluate their full capacity for tissue regeneration.

• Bone
• cartilage
• differentiation
• mesenchymal stem cell
• tissue regeneration

Introduction

Most mesenchymal tissues are remodelled throughout life. Complete regeneration of damaged tissue is an extreme example of remodelling. There are remarkable differences in the regeneration capacity between different species as well as different tissues in the same individual. Notwithstanding the tissue, the remodelling process in adults requires a continuous supply of new cells suggesting the existence of cells, either local or circulating, which have retained their capacity for proliferation and differentiation. It is of interest that the German pathologist Julius Cohnheim, back in 1867 1 suggested that tissue repair in mammals was dependent on cells coming from the bloodstream. Cohnheim's hypothesis was either denied or forgotten for decades but now there are numerous in vivo and in vitro experiments demonstrating that the bone marrow (source of circulating cells), in addition to haematopoetic stem cells (HSC), contain a population (or populations) of stromal cells that under suitable conditions are able to differentiate towards a number of specific mesenchymal cell types including fibroblasts, osteoblasts, chondroblasts, adipocytes etc. 2–6. These cells are now widely called mesenchymal stem cells (MSC) and could be found even in the bone marrow of very old individuals although their number may decrease with age 7, 8.

Since the original demonstration of MSC in bone marrow, several other tissues have also been shown to contain cells that have several features in common with MSC in the bone marrow. Although we still have rather modest understanding of the details of their proliferation and differentiation, they have become an important tool of tissue regeneration studies and are presently considered as potential candidates for several clinical applications 9. Some clinical trials using MSC have already been initiated and encouraging results have been reported, for instance, in osteogenesis imperfecta 10, 11 and metachromatic leukodystrophy 12. MSC transplants have also been used to enhance engraftment of heterologous bone marrow transplants 13. In this review I will first discuss some issues concerning the biology of MSC and then evaluate their potential clinical value in regenerative medicine and in some other indications.

Key messages

• Mesenchymal stem cells (MSC) are multipotent adult stem cells present in bone marrow and having capacity to differentiate towards osteoblasts, chondroblasts, adipocytes and some other cell types.

• The potential plasticity and self‐renewal capacity of MSC offer a huge potential for clinical tissue regeneration.

• Some clinical trials using MSC have given encouraging results.

Development of MSC concept

In a series of elegant papers Friedenstein and his colleagues working in Moscow during 1960s and 70s described clonal and plastic adherent stromal cells from bone marrow 2–6, 14–17. Early studies were able to demonstrate that these cells formed fibroblast colony forming units (F‐CFU) but also had the capacity to differentiate towards osteoblasts, chondroblasts and adipocytes under defined in vitro culture conditions. Later in vitro studies in several laboratories confirmed the plasticity and genuine multipotency of bone marrow stromal cells with cultures originated from cloned single cells. It was also evident from the early experiments with MSC that only minor changes in the culture medium were needed to change their differentiation from one lineage to another supporting further the idea that these cells represent truly a plastic cell population.

The plasticity and the importance of tissue microenvironment for MSC differentiation are also demonstrated in various in vivo animal experiments. Liechty et al. 18, for instance, injected human bone marrow derived MSC into the circulation of foetal sheep early in gestation. hMSC engrafted and persisted in multiple tissues for as long as 13 months after transplantation. In addition, transplanted human cells underwent site‐specific differentiation into chondrocytes, adipocytes, myocytes and cardiomyocytes, bone marrow stromal cells and thymic stroma in their respective tissues.

Beginning from Friedenstein's original experiments much effort has been focused on the direct identification of MSC from the bone marrow and more recently also from other tissue sources. Although there has been major progress we are, unfortunately, still lacking a defined panel of well characterized surface markers for MSC in order to identify them unambiguously in situ. Several surface markers have been studied and some of them, like STRO‐1, HOP‐26, CD49a and SB‐10/CD166 differentiate MSC from most of the other cell types 19. Thus the final definition of MSC is still based on retrospective functional experiments where certain populations of cells or cultures originated from a single stromal cell have been demonstrated to differentiate towards multiple tissue lineages.

Before unquestionable identification of MSC in situ could be done it would be very difficult to answer unambiguously the remaining open questions concerning nature, function and potential therapeutic value of MSC. Firstly, is there one and only one mesenchymal stem cell pool in an adult organism and is this pool a direct descendent of embryonal stem cells? Secondly, what is the hierarchy of MSC differentiation and are these cells able to change lineages even after a substantial differentiation along a specific lineage? Thirdly, how much does the cellular microenvironment determine the cell fate and to what extent, if any, the cell fate is already imprinted during the early development of an individual? All these and many other questions remain to be answered before the concept of MSC has been clarified completely.

Tissue distribution and isolation of MSC

In addition to bone marrow, MSC‐like cells have been shown to be present in a number of other adult and foetal tissues, including circulating blood 20, 21, cord blood 22, 23, placenta 24, amniotic fluid 25, heart 26, skeletal muscle 27, adipose tissue 28, synovial tissue 29 and pancreas 30. These are only some examples in order to demonstrate diversity of potential MSC‐containing tissues. Putting together the present data one may assume that basically all organs containing connective tissue also contain MSC.

The bone marrow is still by far the best characterized source of MSC and almost all that is known about their differentiation is based on studies with marrow‐derived MSC. Actually, in order be to critical enough, one should wait for more detailed comparison between the bone marrow and other tissue‐derived MSC before even categorizing them under the same name. This cautiousness is needed although there are already some comparative studies including several different tissues with a number of phenotypic criteria.

One of the major practical problems in MSC research is the relatively low number of self‐renewing cells in studied tissues. In the bone marrow the frequency of MSC among other cells, including haematopoietic stem cells and various differentiated stages of all other possible lineages, has been estimated to be about 10 MSC per one million of all bone marrow cells 31. Due to the lack of definitive markers for MSC it has not been possible to count exact densities of these cells among the marrow cell population.

Several different modifications from the original method of Friedenstein have been developed to enrich MSC from bone marrow cell suspension. In addition to plastic adherence, medium selection and single‐cell cloning, these include different sedimentation procedures and more recently various affinity‐based cell sorting methods. Also attempts to up‐scale the production of cells in vitro for transplantation purposes have been under extensive study 32. However, all the isolation methods used suffer from lack of absolute specificity as long as there is not a definitive phenotypic marker(s) for MSC. As mentioned earlier several tissues have been demonstrated to contain cells that are able to proliferate and differentiate into various skeletal cell phenotypes. In a recent study of Musina et al. 33 MSC were isolated from human bone marrow, adipose tissue, skin, placenta, and thymus and compared for the presence of surface markers (CD10, CD13, CD31, CD44, CD90, CD105) without observing any differences between different tissue sources. To compare the properties of human mesenchymal stem cells present in different tissues Sakaguchi et al. 34 isolated cells from bone marrow, synovium, periosteum, skeletal muscle, and adipose tissue and studied their colony forming capacity and differentiation under defined conditions. In studies of osteogenesis, the rate of alizarin red‐positive colonies was highest in bone marrow‐, synovium‐ and periosteum‐derived cells, respectively, whereas synovium‐derived cells had the greatest ability for chondrogenesis. In adipogenesis experiments, the frequency of Oil Red O‐positive colonies was highest in synovium‐ and adipose tissue‐derived cells. These and some other results suggest that MSC from different tissues may have differences in their differentiation capacity even if cultured in exactly the same microenvironment.

Some of these tissue sources besides bone marrow, like peripheral or cord blood, adult adipose tissue and skeletal muscle may turn out to be important sources of MSC for transplantation since they are relatively easy to obtain. Although several studies have indicated that many tissues contain cells that are able to differentiate towards osteogenic and some other phenotypes, it remains to be seen if these ‘MSC’ represent similar cell populations or if they are already further differentiated than marrow derived stromal cells. So many different tissues have now been shown to contain mesenchymal cells that are able to differentiate into different mesenchymal lineages that it may well be that all mesenchymal tissues retain their ‘stem cell capacity’ through most of the individual's life span. This would be promising news for medical tissue regeneration. It remains to be seen if these ‘peripheral MSC populations’ have any tissue‐specific properties that may favour tissue‐selective use of them.

Differentiation of MSC and their niche

Embryonal stem (ES) cells from inner cell mass of early embryo are considered to be pluripotent cells and are thus able to differentiate towards all different specific cell types of an adult organism. MSC are considered to be only multipotent cells with clearly more restricted capacity for differentiation than ES cells. Most likely this is the case, but since MSC are present in several different adult tissues and in vitro studies have demonstrated that they are able to differentiate into various cell types, it may well be that the plasticity of MSC has been underestimated 35. Some cellular markers that were thought to be expressed only in ES cells have also been shown to be expressed in MSC.

A proper microenviroment for cellular differentiation in vivo most likely involves both soluble factors as well as multiple cell‐cell and cell‐matrix contacts. So how the plasticity of MSC is then modulated and what are the extra‐ and intracellular signalling pathways that are operating when MSC undergo differentiation towards specific cell lineages? This remains a major challenge of adult stem cell research since it is important to define the correct microenviroment for each lineage in order to create, for instance, proper conditions for in vitro cell expansion. Detailed molecular dissections of the signalling pathways and of transcriptional regulation determining different cellular fates are needed. Both of these are now under extensive investigation and many details have already been solved although the whole picture of MSC differentiation is still far from clear.

Effects of several different hormones, vitamins, growth factors and cytokines on MSC proliferation and differentiation have been tested applying in vitro differentiation assay, studying mainly osteogenic 36–42, chondrogenic 43–45 or adipogenic differentiation 46, 47. Self‐renewal capacity is a fundamental feature of all stem cells. It would thus be important to identify growth factors that not only promote proliferation but also retain self‐renewal capacity of MSC and to maintain their multilineage potential. On the basis of the present data it is far too early to conclude which growth factors are essential to retain the self‐renewal capacity. However, several studies seem to indicate that members of fibroblast growth factor (FGF)‐family, especially FGF‐2, play a crucial role in this. Gene transfer of growth factors applying either viral or other types of carriers has been used to create microenvironment to support differentiation to a specific lineage. Enhanced differentiation towards osteoblasts has been obtained e.g. with over‐expression of bone morphogenetic proteins. Cartilage type of differentiation was greatly enhanced by over‐expression of transforming growth factor (TGF)β in the bone marrow‐derived MSC 48.

When MSC are plated on plastic at low density, the cultures display a lag phase of about 3 days before they enter a phase of rapid exponential growth. Gregory et al. 49 found that as the cultures leave the lag phase, they secrete high levels of dickkopf‐1 (Dkk‐1), an inhibitor of the canonical Wnt signalling pathway. Biological role of the phenomenon was then confirmed by addition of recombinant Dkk‐1 into cultures toward the end of the lag period and, indeed, it increased proliferation and simultaneously decreased the concentration of β‐catenin. If the antibodies to Dkk‐1 were increased during the early log phase it decreased cellular proliferation, and also serum starvation was associated with the decreased expression of Dkk‐1 in MSC. Synthetic peptides derived from Dkk‐1 were found to enhance the proliferation of human MSC in culture. In peptide overlay assays on human MSC lysates these peptides bound a 184‐kDa protein, most likely of low density lipoprotein receptor related protein (LRP)6 50, 51. This observation opens an interesting possibility to develop Wnt antagonists to enhance MSC production both in vitro and perhaps also in vivo.

Perhaps the best characterized lineages presently are osteogenic and adipogenic differentiation. It is clear that osteoblasts and adipocytes originate from the same MSC through the activation of specific transcription programmes. Key transcription factors to direct MSC into osteogenic and adipogenic lineages are runt homology domain transcription factor (Runx)2 and peroxisome proliferators‐activated receptor γ (PPARγ), respectively. The activation of Runx2 and down‐regulation of PPARγ directs MSC into osteogenic lineage and reciprocal transcriptional event leads into adipogenic lineage 52. Recent study of Hong et al. 53 nicely demonstrates that lineage determination could be regulated by another transcription factor TAZ (transcription coactivator with PDZ‐binding motif ). Both of the above mentioned key transcription factors, Runx2 and PPARγ, are under the regulation of TAZ which co‐activates Runx2 dependent osteogenic gene expression and simultaneously represses PPARγ‐dependent adipogenic gene expression. In the future we will most likely see a number of these molecular rheostats that are important modulators of cellular fate and lineage differentiation. Identification of such molecules will further help to clarify signalling pathways that are needed to differentiate MSC into various specific lineages. This information is urgently needed since it could be used to create optimal microenvironment for tissue specific lineages during tissue regeneration for medical purposes.

It has been known for a long time that the number of MSC in the bone marrow declines dramatically with aging. Whether the remaining cells also lose their capacity to proliferate and differentiate and finally to repair tissue damage has not been studied well enough yet. In some studies these issues have been addressed and MSC from aging individuals seem to retain their capacity at least for bone formation 54 and tendon regeneration 55. However, recent microarray studies demonstrate differences in gene expression profiles between foetal and adult MSC providing one explanation why foetal MSC have higher proliferative capacity and are less lineage committed than adult MSC 56. Further comparative studies both at the functional as well as gene expression level are needed to find out how much aging as such affects the capacity of MSC; and especially to detect if some tissue microenvironments are better than others to protect the MSC population during aging. Baxter et al. 57 studied the effect of in vitro expansion on the rate of telomere loss and replicative capacity of MSC. They observed that even minimal expansion induced a rapid aging of MSC, with losses equivalent to about half of their total replicative lifespan. Liu et al. 58 isolated MSC from telomerase (‐/‐) mice and found that they failed to differentiate into adipocytes and chondrocytes, even at early passages, whereas wild‐type MSC under the same conditions were capable of differentiation. This suggests that telomerase activity of MSC is not required only for self‐replication but is also needed for differentiation.

Potential clinical applications of MSC

Since MSC are multipotent and their numbers can be expanded in culture, there has been much interest in their clinical potential for tissue repair and gene therapy. A number of studies have now also demonstrated the migration and multi‐organ engraftment potential of MSC in animal models and in various human organ transplantations. In addition, immunomodulatory properties of MSC and their supportive functions for haematopoietic stem cells have also opened up possibilities for interesting new avenues for their use in the treatment of autoimmune diseases and graft rejection, respectively. Although understanding of mechanisms behind MSC differentiation and cell fate determination is still incomplete and the molecular processes that drive, for instance, engraftment are complex, many clinical applications are now emerging for MSC transplantations. Some of these will be briefly discussed below. For more detailed reviews see other reviews in the present issue of Annals of Medicine.

Tissue regeneration

The potential plasticity and self‐renewal capacity of MSC offers a huge potential for clinical tissue regeneration. Although animal experiments using many different disease models have been successful and promising, the use of MSC for clinical indications is still in its infancy when compared to the common clinical use of HSC.

Several theories have been presented of how bone marrow‐derived mesenchymal stem cells could induce tissue regeneration. The first possibility is that they act like genuine mesenchymal stem cells and could replace stem cell population in the specific tissue and mimic normal growth when building up new connective tissue. The second possibility is that MSC or some of their descendents are able to transdifferentiate to specific cell types in a specific microenvironment and form not only mesenchymal tissues but also ectodermal and/or endodermal components. There is also a third possibility, which could explain tissue regeneration in some specific occasions, namely formation of heterokaryons with MSC and tissue‐specific cells. In addition, it is also possible that transplanted MSC only produce paracrine factors that are able to stimulate tissue specific cells to proliferate or at least protect tissue from further damage giving better possibility for the tissue to recover. All these and perhaps also some other mechanisms should be kept in mind when MSC‐induced tissue regeneration is considered.

Up until now most of the tissue regeneration studies involving MSC (naïve or gene‐manipulated) have focused on development of new bone to heal large bone defects together with some supporting scaffolds 59–61. Obvious major limitations for the use MSC in the healing of site‐specific large bone defects are the need for the relatively high number of cells and identification of optimal supportive biomaterials to construct ‘a living implant’. Several novel approaches to improve conventional tissue culture conditions in order to get a higher number of cells and to prevent loss of their differentiation capacity have been applied. These include studies focusing on the role of serum source and serum treatment 62, 63 and use of osteogenic growth and differentiation factors, like bone morphogenetic proteins and fibroblast growth factors and some others 64–66. Over‐expression of telomerase in MSC leads also to enhanced proliferation 67. The advantage of creating a correct niche for proliferation and differentiation of MSC has stimulated the experimental use of various protein matrices to support lineage‐specific development 68. Increasing knowledge of the molecular mechanisms of stem cell differentiation may also give new pharmacological possibilities, such as intervention of mevalonate pathway by statins or mimics of Wnt‐signalling, and like inhibitors of glycogen synthetase kinase 3β, to enhance osteogenetic properties of MSC 69, 70. In spite of many promising novel strategies we are still lacking a satisfactory standard protocol for in vitro mesenchymal stem cell expansion and optimal biomaterials to prepare living implants for bone repair.

Perhaps the most exciting clinical studies concerning regeneration have been conducted in patients suffering from an inherited bone disease called osteogenesis imperfecta or brittle‐bone disease. These patients suffer multiple fractures already in utero. Direct transplantation of allogeneic MSC to a patient 11 as well as gene therapy using transfection of autologous MSC with a gene construct and application of these ‘corrected’ cells have been used with promising results 71.

In addition of bone, articular cartilage has been the target for intensive cellular therapy efforts, and local delivery of mesenchymal stem cells has been explored as a therapeutic approach in different animal models of osteoarthritis (for review see, for example, (72)). Since regeneration of articular cartilage is very poor, development of MSC‐based methods for regeneration of this tissue would be of enormous medical importance. In an elegant study of Ichinose et al. 73 cultured bone marrow cells in alginate beads with a chondrogenesis‐induction medium containing TGF‐β3. Proliferating cells first differentiated to typical matrix producing chondroblasts and gradually during the subsequent culture became hypertrophic with typical gene expression profile of hypertrophic chondrocytes. Interestingly, these cartilage nodules finally calcified and simultaneously started to express osteoblastic marker genes mimicking endochondral ossification. It remains to be seen if the authors were able to create conditions for transdifferentiation of chondrocytes to osteoblasts or if these cells differentiated from the remaining stem cells in the presence of correct microenvironment. These findings indicate that three‐dimensional culture could be a convenient in vitro model to study different steps of endochondral bone formation all the way from MSC via cartilage to bone.

Therapeutic applications of adult mesenchymal stem cells are approaching clinical use also in several fields other than bone and cartilage regeneration. Promising new tissue targets among others seem to be, for instance, tendons, ligaments, menisci, and other connective tissues 74, 75. Cartilage, bone and ligaments have also been tried to engineer using bioreactors. Human MSC, highly porous protein scaffolds and bioreactors with perfusion cartridges were used to produce specific conditions for lineage specific differentiation 76. In each case, the scaffold and bioreactor were designed to mimic some aspects of the environment present in native tissues. These and some other studies have demonstrated that also physical parameters, like loading and fluid flow, could be important to create a differentiation niche for MSC.

One of the most exciting potential applications for the clinical use of MSC is regeneration of the heart muscle after myocardial infarct. It has been shown that multipotent mesenchymal stem cells are present within the connective tissues of the heart and may thus be useful for local tissue regeneration if somehow inducible 77. However, more interest is now focused on either systemic or local implantation of bone marrow‐derived allogeneic or autologous MSC into damaged myocardium 78. Experiments using several different animal models have thus clearly shown that engraftment of MSC induces at least transient progress of the healing process in the damaged myocardium. However, there is much controversy concerning the mechanism(s). First of all, it remains to be seen if bone marrow‐derived stem cells could differentiate autonomously into muscle cells or if tissue regeneration is reached via cellular fusion. Second, what is the exact origin of cardiomyocyte progenitors in the bone marrow? Are they a subpopulation of MSC or HSC and does actual transdifferentiation take place? Earlier studies suggested that HSC can transdifferentiate into cardiomyocytes, however, more recent studies have challenged the origin of myocardium‐engrafted cardiomyocytes being HSC 79. Thirdly, more recent studies have suggested that the protective effect of MSC in myocardial infarction is mainly due to their paracrine effects, meaning that transplanted cells secrete soluble factors that are able to protect cells from tissue damage 80, 81. This effect does not need an actual replacement of cardiomyoblasts with MSC‐derived cells. In conclusion, several in vivo studies indicate that MSC transplantation improves cardiac function in animal models of cardiomyopathy and ischaemic damage. It remains to be seen if this is direct tissue regeneration via injected MSC or if it is mainly through induction of myogenesis and angiogenesis from local progenitors or mainly by inhibition of myocardial fibrosis by paracrinic factors produced by cardiac progenitors 82. It is interesting to note here that a recent study of Ma et al. 83 also suggested a paracrinic role for transplanted MSC in the healing of chemically damaged cornea in rats. In addition to the studies mentioned, MSC have been applied in a number of other experimental tissue regeneration studies. Thus there are good reasons to believe that at least some of these experimental models will produce new potential applications for further clinical practice.

MSC as targeting vehicles

Animal studies have shown the engraftment of MSC to different tissues and further suggested that it may be enhanced by local tissue damage and stress. This has opened up a possibility that in addition to direct tissue regeneration, MSC could also be used as targeting vehicles in gene therapy with various therapeutic molecules. At present there are experimental data and animal models to test MSC‐based gene therapy for both malignant and non‐malignant indications.

It has been shown that intra‐tumour injection of MSC into malignant gliomas caused significant inhibition of tumour growth and increased the survival of glioma‐bearing rats. Gene‐modification of MSC by infection with an adenoviral vector encoding human interleukin‐2 (IL‐2) clearly enhanced the anti‐tumour effect and prolonged the survival of tumour‐bearing animals 84. Injection of MSC over‐expressing interferon β suppressed the growth of pulmonary metastases, presumably through the local production of interferon β in the tumour microenvironment 85. These and other related studies demonstrate that gene therapy employing MSC as a targeting vehicle would be a promising therapeutic approach for some treatment‐resistant tumours. In addition, they may also have a direct capacity to restrict malignant growth.

Transfection of MSC with potentially therapeutic genes has been tried also in many other disease models. They have turned out to be potentially useful gene delivery vehicles of gene therapy e.g. for Parkinson's disease 86. Genetically modified mesenchymal stem cells have also been engrafted into the heart to develop a biological pacemaker 87. Kurozumi et al. 88 transfected telomerized human MSC with the brain‐derived neurotrophic factor (BDNF) gene using a fibre‐mutant adenovirus vector and reported that such treatment improved recovery in a rat transient middle cerebral artery occlusion model. In conclusion, several studies have demonstrated that MSC could be transfected with therapeutic molecules and these gene‐modified cells retain their capacity for engraftment into damaged tissues or tumours.

Immunomodulatory role of MSC

A number of different studies have demonstrated that MSC avoid allogeneic rejection in humans and in different animal models 89, 90. This opens up a much broader use of MSC for tissue transplantation purposes than use of autologous cell sources only. Reasons for this are not totally clear but it has been shown that MSC often lack major histocompatibility complex (MHC)‐II and co‐stimulatory molecules 91. Mechanisms may well be similar to that behind maternal tolerance of the foetal allograft. MSC have been shown to have a direct immunosuppressive effect on T‐cell proliferation in vitro. It has also been shown that MSC induced apoptosis in activated T‐cells but had no effect on resting T‐cells 92. In addition, MSC may also have some capacity to modulate the differentiation, maturation and function of dendritic cells 93. This may be partially due to the fact that MSC decreased their capacity to produce cytokines, for instance interleukin (IL)‐12 94. These immunosuppressive properties of MSC have also been tested in various in vivo models and clinical situations. Co‐injection of MSC has clearly prolonged the rejection of mismatched skin grafts in animals 95 and it has been shown that MSC were homing to several damaged tissues after radiation‐induced multi‐organ failure 96. Treatment of severe acute graft‐versus‐host diseases have been tried with transplantation of MSC with striking results suggesting that immunosuppressive properties of MSC may be of interest when developing new strategies to treat these complicated and difficult conditions 97. Some of the recent experimental 98 and clinical reports 99 are also encouraging and suggest that transplantation of MSC may have a role even in the treatment of autoimmune diseases. However, immunosuppressive effects may also have unwanted consequences since it has been shown, for instance that co‐injection of MSC with B16 melanoma cells favoured the tumour growth in allogeneic recipient mice 100. Although co‐transplantation of MSC with HSC has given very promising results we obviously need more experimental and clinical data in order to determine the value of co‐transplantation of MSC in different clinical conditions.

Safety of MSC transplantation and malignant growth

Stem cells' ability for self‐renewal includes a potential possibility for unlimited proliferation and growth. This immediately prompts two questions: First, can transplanted MSC become malignant raising thus a major concern about the safety of adult stem cell therapies? Second, are tumours as such derived from either local or bone marrow‐derived stem cells?

Until recently MSC transplantations were thought to be without major risk of malignant transformation. However, several lines of evidence now suggest that in vitro expanded cells may be the target of neoplastic transformation. Serakinci et al. 101 transduced human mesenchymal stem cells with the telomerase gene to investigate the neoplastic potential of adult stem cells. The human telomerase reverse transcriptase (hTERT)‐transduced line developed loss of contact inhibition, anchorage independence and formed tumours in mice. In the following experiments hTERT‐transduced cell lines showed highly variable tumorigenecity. Han et al. 102 was also able to establish a tumour cell line from mutated human embryonic bone marrow mesenchymal stem cells (MSC). They further showed that nucleostemin may play an important role in both tumorigenesis and transforming human embryonic bone marrow mesenchymal stem cells into tumour cells. Finally, Rubio et al. 103 demonstrated that although MSC can be managed safely during the standard ex vivo expansion period (6–8 weeks), they can undergo spontaneous transformation during long‐term in vitro culture (4–5 months). On the other hand, several clinical and experimental trials have been performed without any signs of transformation. However, the above mentioned studies clearly arouse serious concerns on the use of in vitro expanded MSC.

The role bone marrow MSC in the progression of cancer metastases deserves some attention, too. It is well known that especially breast cancer metastases in bone marrow are associated with increased bone resorption and enhancement of osteoclast formation from HSC population. Previously it was thought that this was mainly due to the fact that breast cancer cells secrete osteoclast stimulating factors. However, more recent data suggest also that MSC in bone marrow may be the primary target of cancer cells‐produced cytokines and mediate the stimulatory effect to osteoclast progenitors. In the case of prostate cancer metastasis it is even more obvious that their metastases directly stimulate MSC to differentiate into osteoblasts. Thus different tumour cells may induce specific changes in the bone marrow microenvironment and modify stem cell plasticity accordingly.

According to cancer stem cell theory epithelial tumours are thought to originate from transformation of tissue stem cells. If cancer stem cells could in fact be directly recruited from bone marrow by some non‐specific stimuli, such as chronic inflammation, this would be an interesting possibility. According to results of Houghton et al. 104 this could the case. They showed that chronic infection of C57BL/6 mice with helicobacter induces repopulation of the stomach with bone marrow derived stem cells. These cells then progressed through metaplasia and dysplasia to cancer.