Stem Cell Reviews and Reports

, Volume 7, Issue 3, pp 560–568

The Origins of Mesenchymal Stromal Cell Heterogeneity


  • Meirav Pevsner-Fischer
    • Department of Molecular Cell BiologyWeizmann Institute of Science
  • Sarit Levin
    • Department of Molecular Cell BiologyWeizmann Institute of Science
    • Department of Molecular Cell BiologyWeizmann Institute of Science

DOI: 10.1007/s12015-011-9229-7

Cite this article as:
Pevsner-Fischer, M., Levin, S. & Zipori, D. Stem Cell Rev and Rep (2011) 7: 560. doi:10.1007/s12015-011-9229-7


Cultured mesenchymal stromal cell (MSC) populations are best characterized by the capacity of some cells within this population to differentiate into mesodermal derivatives such as osteoblasts, chondrocytes and adipocytes. However, this progenitor property is not shared by all cells within the MSC population. Furthermore, MSCs exhibit variability in their phenotypes, including proliferation capacity, expression of cell surface markers and ability to secrete cytokines. These facts raise three major questions: (1) Does the in vitro observed variability reflect the existence of MSC subsets in vivo? (2) What is the molecular basis of the in vitro observed heterogeneity? and (3) What is the biological significance of this variability? This review considers the possibility that the variable nature of MSC populations contributes to the capacity of adult mammalian tissues to adapt to varying microenvironmental demands.


Mesechymal stromal cellsStem cellsHeterogeneityPlasticityHierarchal differentiationStem state


Plating bone marrow (BM) cells, in serum-containing medium, results in the formation of colonies of adherent cells with fibroblast-like characteristics. Friedenstein and colleagues found these cells capable of differentiation into chondrocytes and osteoblasts [1, 2]. Originally these cells were named as colony-forming unit fibroblasts (CFU-Fs). Based on the ability of these cells to differentiate into mesodermal cell lineages, they were then designated as mesenchymal stem cells [3].

In this review we discuss the heterogeneity observed in bone marrow derived mesenchymal cells, which are the most extensively investigated population. However, the mechanism underlying population heterogeneity will take into account information collected from MSC isolated from all tissues. We refer to cells in mesenchymal cultures as “MSCs” to indicate their stromal nature that may or may not involve the capacity to differentiate into different mesodermal lineages. As is made evident in this review, cells in such cultures differ markedly to the point that some have progenitor properties and multi-potent differentiation capacity, whereas others are devoid of this ability and are not unlike what is commonly called, fibroblasts. Thus, calling bulk mesenchymal cell cultures “multipotent stem cells” does not reflect their true nature. In this review we term low-passaged MSC as primary cultures, whereas long term cultured- or MSC passaged extensively, are referred to as cell-lines.

It is important to make the distinction between intra-population versus inter-population heterogeneities. The former relates to differences among individual isolates of MSCs, i.e., cell populations obtained either from different donors or from different mouse strains. For example, in a study which compared human iliac crest marrow MSCs from 17 healthy donors exposed to osteo-inductive media, disparity in growth rate, alkaline phosphatase activity and bone-specific gene induction were evident among the different MSC isolates [4]. Similarly, comparing mesenchymal cells of five different strains of mice indicated differences in abundance of MSCs in the BM, as well as differences in growth kinetics and alkaline phosphatase activity [5]. Inter-population heterogeneity deals with the heterogeneity within the mesenchymal cell population from an individual isolate. In this review we confine the discussion to the inter-population heterogeneity, i.e., the clonal heterogeneity. However, as will be discussed below, both intra-population and inter-population heterogeneity of mesenchymal cells may be due to similar causes.

The Defining Properties of Mesenchymal Cells


MSCs display heterogeneous morphology. Different terms were used to describe their morphology: fibroblastoid cells [6], giant fat cells and blanket cells [7], spindle shaped, flattened cells [8], and very small round cells [9]. Mouse bone marrow adherent cell lines (MBA) were classified as pre-adipogenic, endothelial-adipogenic, endothelial-like, fibroblastic, osteogenic and fibro-endothelial [10] according to their morphology. Furthermore, stromal cell-lines were shown to differ in the quantity and types of extracellular matrix glycoprotein synthesis, production of myelopoietic factors and modulation of tumor cell growth [11]. Human MSCs express a variety of transcripts characteristic to mesenchymal lineages as well as transcripts typical to skeletal, adipose, muscle and endothelial tissues [12]. Clearly, the mesenchymal cell shape changes dramatically in correlation to their seeding density and is further modified on confluence. It is, to date, unclear how these morphologies relate to cell functions and what they signify.


Better understanding of MSC biology requires, among other things, the isolation of these cells to purity. A popular approach for cell isolation is the use of cell surface markers, encouraged by studies of hematopoietic stem cells (HSCs). The latter are enriched most effectively through the use of marker-combinations. For mouse HSCs, stem cell antigen (Sca)-1 and c-Kit serve for positive selection, while cells expressing lineage markers of differentiated hematopoietic cells are eliminated. The highest enrichment for HSCs was obtained by selecting cells with the highest dye efflux capacity and the phenotype CD34c-Kit+Sca-1+Lin [13]. Human HSCs are enriched within the CD34+CD38−/low cell fraction [14]. In contrast to these relatively effective means to enrich for HSCs, to date, there is not any available effective and reproducible method for purification of MSCs. This is due to the fact that MSCs do not express specific markers. The International Society for Cellular Therapy proposed three minimal criteria for cultured human MSC definition in order to minimize differences between laboratories: 1) when maintained in standard culture conditions, MSC must be plastic-adherent. 2) MSC must express CD105, CD73 and CD90, and lack expression of CD45, CD34, CD14 or CD11b, CD79 or CD19 and HLA-DR surface molecules, and 3) MSC must differentiate to osteoblasts, adipocytes and chondroblasts in vitro [15]. However, this list of proposed markers alone does not strictly characterize MSCs. There are many other proposed markers for MSCs enrichment. In mice, the phenotype PDGFRα+, Sca1+, CD45- and TER119- was suggested [16], as well as the positive selection markers stage-specific embryonic antigen 1 (SSEA-1) [17] and SSEA-4, the latter being also a marker of human MSCs [18]. For human MSCs, STRO-1 [19, 20] and CD271 were each indicated. The latter characterized the majority of highly clonogenic small round cells rapidly adhering to the plastic [21]. Within CD271+ population, Buhring et al. detected CD140b, CD340 and CD349 [22] positive cells. Sacchetti et al. proposed CD146 as a marker for bone marrow MSCs that are clonogenic and osteogenic in vivo [23]. Russel et al. claimed that CD146 expression correlates with MSC differentiation potential. No such correlation was found with the expression of CD271, CD44 and CD73 [24]. As many laboratories use different sets of antigens for MSC analysis, it is difficult to compare data from these studies. An inter-comparison among different reports demonstrated that there is no one marker or a limited set of markers specific for MSCs [10]. It seems that there is no consensus with regard to the antigenic characteristics of MSCs. A more solid manner of identifying these cells is to examine their differentiation potentials. However, the inadequacy of this approach is that it is retrospective and the nature of the cell at hand cannot be determined in real time.


It has long been established that there are differences in the ability of clones of MSCs to differentiate into the three main lineages: osteocytes, adipocytes and chondrocytes. Pittenger et al. were the first to test the hypothesis that bone marrow contains individual human MSCs, which could differentiate into all three lineages; the differentiation capacity of expanded colonies derived from single cells was examined. Of six colonies, all underwent osteogenic differentiation, five underwent adipogenic differentiation, and two underwent chondrogenic differentiation, whereas one other colony displayed weak chondrocytic phenotype. Thus, the authors concluded that cells with tri-lineage differentiation ability represent human mesenchymal stem cells. Colonies which displayed limited differentiation ability either lost multi-lineage potential during in vitro culture or represent progenitor cells with limited differentiation potential [25]. Muraglia et al. broadened the analysis by examining the ability of 185 non-immortalized human bone marrow stromal cell clones, to differentiate, thereby confirming Pittenger’s study. All clones but one differentiated into the osteogenic lineage; one third of the clones differentiated into all three lineages analyzed. Most clones displayed an osteogenic-chondrogenic potential. Clones with an osteogenic-adipogenic potential, chondrogenic-adipogenic phenotype, or pure chondrogenic and adipogenic clones were not observed. It was further observed upon repeated passaging, that clones progressively lost their adipogenic and chondrogenic differentiation potential [26]. Russell et al. examined the differentiation potential of 96 clonal cultures from two human MSC donor preparations. Tri-potent MSCs accounted for nearly half of colony-forming cells. All bi-potent MSCs were observed, when for one donor the most frequent bi-potent MSCs exhibited the osteogenic-chondrogenic phenotype, for the MSC preparation from the other donor the most frequent bi-potent was osteogenic-adipogenic. Mono-potent osteoprogenitors represented 10% of the colony-forming cells in culture and mono-potent MSCs with an adipogenic phenotype were detected infrequently. MSC clones that did not exhibit the potential to differentiate into any of the three lineages were rare [24]. These results differ from Pittenger and Muraglia’s work, as all eight possible combinations of differentiation potencies were observed. Clonal differentiation heterogeneity is also observed in immortalized human MSCs. One hundred single-cell derived clones were established from parental immortalized human MSCs and the differentiation properties varied considerably: tri-, bi-, and mono-directional clones were identified. The only combination that was not observed was adipogenic-chondrogenic potential [27].

Inter-Clonal Heterogeneity

A significant observation is that single-cell derived colonies are heterogeneous, showing variability not only among colonies (or CFU-Fs), but also within the same colony. When cells from a single MSC mother colony were subjected to osteogenic induction, some of the sub-cultures efficiently differentiated into osteoblasts, while others did not. This experiment suggested that even when derived from a single cell, the progeny of MSCs can be conditioned to behave differently. The same results were noted for adipogenic differentiation [28]. As MSCs are known to lose differentiation potential following many population doublings, it is possible that some of the expanded cells lose their differentiation potential. An alternative interpretation would be that variance in potency is a result of stochastic events. In terms of morphology and commitment to differentiation, an independent study demonstrated, that cells in the inner regions differ from cells in the outer regions of an MSC colony. If the cells are re-plated at clonal densities, those differences disappear [29]. Differences in proliferation and differentiation potentials between the colony center and margins were additionally demonstrated in a two-stage colony assay [30], in which cells from primary single cell-derived colonies were detached and reseeded again, as single cells, to form new secondary colonies.

The above description of experimental data on MSC properties highlights the phenotypic and functional heterogeneity within primary cell cultures or clonal populations. The heterogeneity is reflected in cellular morphology, differential marker expression and variable differentiation potentials. This heterogeneity, observed within the MSC population, could be a result of (a) alterations induced by extensive ex vivo culturing and (b) in vivo heterogeneity and variable phenotypes that reflect the natural repertoire of MSCs. In the following section we discuss the evidence for each of these possibilities.

Possible Mechanisms Accounting for MSC Heterogeneity

Heterogeneity as a Consequence of Long-Term Culture

The isolation of MSCs is based on their adherence to plastic surfaces and subsequent proliferation. The isolation process is lengthy; following massive expansion and several weeks after their initial seeding, the resulting MSCs are examined. There is a possibility, therefore, that the heterogeneity of MSCs is induced due to conditions imposed by long-term culturing. Indeed, MSCs that were cultured extensively in vitro, exhibit changed surface molecule expression profiles [31], decreased homing to the BM and spleen [32], decreased capacity to support the engraftment of NOD/SCID-repopulating HSCs [33], decreased ability to differentiate into mesodermal lineages [28, 34, 35], increased cell size [36] and increased signs of cell aging [34]. Long-term culture was further associated with continuous changes in the global gene expression profile [3638], and DNA methylation [39], accounting for the phenotypic changes occurring in these cells. Gene expression associated with cell differentiation, apoptosis and cell death was up regulated, whereas expression of genes involved in mitosis and proliferation was down regulated [37]. It is possible that the detachment of MSCs from their in vivo niche induced changes in gene expression in an attempt to adapt to the new environmental conditions. Moreover, the detachment of MSCs from their original restricting niche [40] can induce partial MSC differentiation, consequently yielding a mixture of cells at various degrees of maturation [41], while further contributing to the observed heterogeneity in MSC population within each individual culture.

In vitro culture of MSCs maintains them in a cycling state. The environment provided in the in vitro culture does not protect the cells from toxic insults such as high oxygen and other stress conditions, thus increasing the chance for accumulation of mutations. The latter may not necessarily cause transformation of the MSCs, but can generate cellular heterogeneity (summarized in [41]).

MSC heterogenic cellular morphologies can result from replicative senescence. As reviewed in [41], the number of MSC population doublings is restricted by the Hayflick’s limit, due to lack of telomerase activity in these cells [42]. However, the rate of replicative senescence can also be affected from culture conditions; either cell plating density [43] or oxygen concentrations [44, 45] can alter the number of cell doublings. These conditions can vary between cells in a single tissue-culture dish [46]. In addition, replicative senescence can be switched on by stochastic means or represent the remaining number of cell divisions that varies already at the time of MSC isolation [41]. The latter may contribute to the observed cellular heterogeneity.

It can be stated without any reservation that in vitro culturing modifies MSC populations and may contribute to their heterogeneity. Nevertheless, many studies have been performed using low passaged MSCs that have probably been modified in a limited manner. Moreover, the study of long-term cultured MSC cell strains versus cell lines indicates that such cells may maintain a stable phenotype over years of passaging, e.g., preadipocytes that support hematopoiesis [47] versus endothelial-like cells producing high titers of activin A [48]. It cannot be excluded therefore that some, if not most, of the heterogeneity of MSCs originate in vivo.

MSC Heterogeneity may Reflect Their In Vivo Repertoire

It has been suggested that the variability in cultured MSC populations represents the diverse repertoire of distinct subpopulations that exist in vivo. This suggestion was based on the analysis of expression of different classes of regulatory proteins involved in angiogenesis, hematopoiesis, cell motility and communication, immunity and defense and neural activities. It was suggested that MSCs possess functions related to the above-mentioned regulatory networks [12, 49]. For example, a subset of MSCs expressing classes of neuro–regulatory proteins may function to maintain the integrity of nervous tissue in vivo and direct nerve fiber growth into bone and marrow during tissue remodeling after injury. MSCs expressing angiogenins may induce capillary proliferation, expansion of the sinusoidal space and vessel growth as well as remodeling. Therefore the development of protocols for isolation of distinct MSC subpopulations may provide improved vectors for the treatment of specific diseases [50]. MSCs can be isolated from different tissues, including perivascular sites, exhibiting pericyte-related phenotype [10, 51, 52]. MSCs isolated from different tissues exhibit differences in their phenotypes and functions [53, 54], which might suggest multiple sources in MSC ontogeny [10, 55]. MSC development in the embryo is not well understood. MSCs were suggested to be generated through epithelial-mesenchymal transition [56, 57] or to be the progeny of an ancestral MSC [55].

If indeed MSC heterogeneity reflects their in vivo composition, then the major questions raised is: How does this heterogeneity form and what could be its biological significance?

Heterogeneity Resulting from Hierarchical Organization of the Differentiation Cascade

A hierarchical model was proposed to explain the existence of MSC clones with divergent differentiation potential. Multi-potent mouse and rat MSCs were suggested to give rise to more restricted clones and to finally differentiate into mono-potent progenitor cells [58]. This process was suggested to be a non-random, single-step process in which multi-potential progenitors become exclusively restricted to a single lineage by particular culture conditions and inducers. In a large-scale experiment, 185 cultures of non-immortalized human BM derived MSCs were examined. No MSC clones showing differentiation potential of osteogenic-adipogenic or chondrogenic-adipogenic phenotype, nor pure chondrogenic and adipogenic cultures could be found [26]. The authors suggested that the heterogeneous composition of differentiation potentials among clones, indicated a hierarchical differentiation structure. Within this hierarchy, adipogenic ability to differentiate is the first to be eliminated. This is followed by the loss of chondrogenic differentiation ability, which results in clones capable of osteogenic differentiation only. The authors do not exclude the option that while intrinsic lineage commitment of the MSCs can be explained by this model, in vitro culture conditions may represent a microenvironment favoring osteogenesis, subsequently inducing clones to lose their adipogenic and chondrogenic differentiation potential during expansion. Concurring with this study, absence of clones with osteogenic-adipogenic or chondrogenic-adipogenic differentiation potential, in BM-MSC cultures, was demonstrated [59]. In the aforementioned study, the elimination of adipogenic differentiation potential preceded the loss of chondrogenic and osteogenic differentiation potentials [59]. To further examine the putative MSC hierarchical structure, human umbilical cord derived mesenchymal progenitors, were cloned and sub-cloned in vitro, and their differentiation potentials were examined [60]. Clones with adipogenic and chondrogenic differentiation potentials could only be found when combined with osteogenic differentiation potential. Myogenic differentiation potential was detected only in the clones which were able to differentiate into adipocytes, chondrocytes, osteocytes and fibrocytes, and was the first lost upon prolonged culture. This was interpreted as a hierarchical mechanism in MSC fate decisions [60]. In contrast to these reports [26, 59, 60], examination of foreskin derived MSCs showed no such hierarchy [61]; clones with osteogenic-adipogenic differentiation potential and clones with mono-potent adipogenic differentiation potential [61] were detected. Immortalized human MSC populations were found to contain all types of clones with the exception of chondrogenic-adipogenic ones [27]. In an additional study [24], clones with tri-potent, all combinations of bi-potent and mono-potent clones were detected, leading the authors to propose a random hierarchy model that predicts the existence of all possible combinations of bi-potent and mono-potent MSCs. This study proposed a model in which the tri-potent clones give rise to any bi-potent clone followed by loss of additional differentiation ability that finally give rise to mono-potent cells.

Detailed examination of the protocols used for clone isolation in the above section, reveals that in the studies suggesting osteogenic orientated hierarchy [26, 59, 60], clones were prepared by diluting fresh BM or non-passaged BM adherent cultures. In contrast, in the papers showing human BM-MSC clones with adipogenic-osteogenic and all mono-potent differentiation ability, the clones examined were prepared using BM adherent cultures of passage-2 or more [24, 27, 61]. It is possible that culture conditions influence selection or loss of subpopulations of multi-potent progenitors. Therefore, the in vitro hierarchical structure may not necessarily reflect the differentiation sequence of MSCs in vivo.

It was indeed proposed that the in vitro observed tri-lineage differentiation of MSCs is not valid. Instead, in vivo differentiation, upon heterotopic transplantation, should be used for MSC characterization as it probes the native potency of MSCs, in the absence of artificial cues [62]. The widespread use of in vitro assays is a major source of confusion, controversy and disagreement that may stem from the fact that the properties of cultured MSCs and their regulatory networks are rather different from those of their in vivo counterparts [62]. It should be noted that heterogeneity is also observed when osteogenic differentiation is tested in vivo. Analysis of bone formation in vivo by single-colony derived cells of human MSCs was assayed [8]. Heterogeneity was demonstrated by the ability of MSC clones to form bone and to support hematopoietic tissue formation when transplanted into immunodeficient mice. Whereas multi-colony strains formed bone, only 58.8% (20 out of 34) single colony-derived clones generated bone. Extensive bone formation, accompanied by hematopoietic tissue within the bony structures, was observed in 8 out of the 20 single colony-derived cells positive for bone formation [8].

The prevailing point of view is that cell populations originate from a mother cell by defined hierarchy such as suggested for HSC (reviewed in [8, 63]). The proposed models for MSC differentiation, mentioned above, are primarily deductions from the structure of the hematopoietic system, whereas solid evidences for hierarchical organization of mesenchymal populations are not available. Firstly, intermediate steps in the differentiation cascade were not described [55]. Secondly, a hierarchical paradigm does not take into account possible MSC plasticity and the functional overlap that exists between lineages, such as adipocytic and/or osteoblastic and stromal lineages. As opposed to the paradigm in which a stem cell is a fixed entity that assumes a deterministic and irreversible differentiation scenario [64], a new concept has been suggested: stemness is a transient and reversible cell state rather than a fixed property that develops following interaction with the environment (reviewed in [10]). The transitions between the differentiated and more earlier stem state states, as well as the transitions between various differentiated cells, are termed cellular plasticity. These have been suggested as being a fundamental property of MSCs [10, 55, 65, 66].

Plasticity of the Mesenchymal Phenotype

Plasticity relates to a whole range of cellular functions, including switches in lineage and cell fate choices during the cell’s life span [10, 55, 65, 66]. MSCs were found to possess plastic properties: the study of bone marrow derived mesenchymal cell lines, showed the morphologic phenotype to be very flexible [47, 67, 68]; MSCs show plastic behavior, [10, 69] they can dedifferentiate [70, 71] and trans-differentiate either between mesodermal lineages [72, 73] and non mesodermal germ layers [7478]. All of these can contribute to the phenotypic and functional heterogeneity of MSC cultures. Hematopoietic cell lines maintain a relatively fixed phenotype on in vitro derivation. The study of cultured mesenchymal cell lines leads to the conclusion that, on culture, these cells continue to change fate, and are therefore plastic [65]. The molecular mechanisms leading to plasticity and to the plastic nature of the MSC, and stem state in general, were suggested to involve stochastic changes in MSCs [7981], and are reviewed in the next section.

Stochastic Changes in the Cellular Phenotype of MSCs

Cellular heterogeneity within a clonal population was proposed to be the outcome of persistent cell individuality resulting from fluctuations of protein levels in mammalian cells, inducing ‘non-genetic cell individuality’ [81]. These fluctuations within clonal populations were suggested to be a manifestation either of ‘gene expression noise’ [8284], or stable phenotypic variants [85]. As for progenitor cells, the degree of multi-potency is in general established by intracellular states, including that of gene expression patterns, protein concentrations and epigenetic factors [86]. The fluctuations in these intracellular events may be the basis of the heterogeneity observed in progenitors in general, and in the MSC population, in particular.

Processes of molecular fluctuation within the individual cell have been proposed to be a main factor in cell differentiation. As a cell becomes more lineage-restricted, deterministic processes appear to be more relevant. For example, it was shown that high oxygen tension reduces stochastic state fluctuations (noise) of differentiated states, thereby partly inducing differentiation [87]. A stochastic repression/induction model was proposed accordingly [79]. In this model, MSC differentiation is defined not as a part of the deterministic dogma of hierarchy, but as the loss of stem cell properties. Cell differentiation is quantified as being a state of the cell that could shift in any given moment as a result of stochastic changes. The authors assume that an important environmental factor, affecting MSC expansion, is oxygen tension that is able to control state fluctuations. Decreasing oxygen tension increased the amplitude of fluctuations in differentiated states, thereby changing the composition of cells along the axis of differentiation states, thus increasing heterogeneity in a shorter time needed for high pressure oxygen to reach the same state. Similarly to hematopoietic cell differentiation [81], the authors suggest that there is no simple set of individual master genes for each mesenchymal lineage that alone can account for the change in differentiation states along all mesenchymal pathways. Rather, MSC differentiation state is under the influence of multiple inductive and repressive factors.

An example of stochastic changes in progenitor cell properties, comes from the examination of expression levels of the stem cell marker Sca-1 [81]. In clonal populations of mouse hematopoietic progenitor cells, spontaneous ‘outlier’ cells with either extremely high, or conversely low expression levels of Sca-1, reconstitute the parental distribution of Sca-1. The authors suggest that reconstitution of the original Sca-1 expression can be described by a model incorporating noise-driven transitions between discrete subpopulations. However, clonal heterogeneity of gene expression does not simply result from random fluctuations in the expression of a single gene, but reflects fluctuating transcriptome-wide noise that has significant biological functionality in cell fate commitment. This study therefore proposed that heterogeneity may account for the stochasticity of cell fate decisions in stem cells [81]. MSCs were shown to be heterogeneous in expression of multiple cell markers within a single culture or clone [1723]. The influence of intrinsic stochastic gene expression changes in combination with culture conditions, might account for MSC heterogeneity either in marker expression or other functions, accounting for the heterogenic phenotypes described in the first part of this review.


MSCs display heterogeneity in phenotypes and functions. The biological benefits of such biochemical heterogeneity may be related to the proposed functions of these cells. MSCs may be involved in tissue repair and inflammation. Both conditions inflict stress and endanger the integrity of the tissue. The heterogeneity of MSCs may allow selection of the appropriate cell type for each extreme condition. By contrast, homogenous and rigid populations could be counterproductive under strong demand for tissue repair and immunomodulation. It was proposed that the heterogeneous populations reflect the complexity of the stromal system in the bone marrow. The varied functions it performs to regulate tissue homeostasis, contributes to the demonstrated functionality of MSCs in vivo [50]. A role for plasticity, leading to heterogeneity in MSCs, was suggested to be similar to that of embryonic mesenchyme. In the embryo tissue and organs, constructions are dramatic and robust processes. They entail very extensive plastic cell behavior. In contrast, the adult organism requires relatively few changes, most of which are related to tissue maintenance in replacing damaged or aged cells. These processes occur at a low incidence and when compared to those in the embryo, take place rarely. Therefore, the plastic nature of adult MSCs is suppressed and is only revealed on their in vitro culture in the absence of tissue constrains [65], or under extreme conditions of tissue destruction in vivo. The plastic nature of MSCs is revealed by the expression of a variety of gene families that characterize their differentiated progeny. In this respect, MSCs are in a standby state, being prepared to take a dramatic leap into the required direction [55, 65, 66]. It seems that long-term culture of MSCs leads to the selection of specific clones overtaking the culture. Since MSCs may be a priori heterogeneous, functionally divergent MSC cultures could simply be an outcome of specific culture conditions that select particular type of MSCs. To date, it is unclear which cell(s) type is the in vivo precursor of cells that are grown in culture as MSCs. Many contenders have been suggested; it cannot be excluded that more than one cell type grown in culture develops an MSC phenotype. Further research should therefore focus on the identification of precursors of MSCs in vivo. Such information may allow purification of MSCs without their culture and would provide an answer to the question of whether MSC heterogeneity is a defining property of these progenitors in vivo.


The authors are indebted to the Helen and Martin Kimmel Institute for Stem Cell Research and the M.D. Moross Institute for Cancer Research, at the Weizmann Institute, the Gabrielle Rich Center for Transplantation Biology and the support of the Legacy Heritage Fund of New York. DZ is an incumbent of the Joe and Celia Weinstein Professorial Chair at the Weizmann Institute of Science.

Conflicts of Interest

The authors declare no potential conflicts of interest.

Copyright information

© Springer Science+Business Media, LLC 2011