During the last few years isolations of adult mesenchymal stem cells from different sources have been reported. Bone marrow derived stem cells first described by Friedenstein et al. are still the most frequently investigated cell type and often designated as the gold standard 1 . Mesenchymal stem cells derived from adipose tissue 2 , peripheral blood 3 , the lung 4 or the heart 5 however have also shown promising potential for proliferation and differentiation into different cell types. In this section we focus on the comparison of adult mesenchymal stem cells derived from bone marrow (BM), adipose tissue (AT) and peripheral blood (PB) (Figure 1).

BM-MSC are isolated from bone marrow aspirate. This invasive procedure is painful for the patient and is accompanied by a risk of infection. The commonly applied preparation method for the generation of MSC from bone marrow is density gradient centrifugation 6 . The collected fraction containing mononuclear cells (MNC) is washed and the cells are seeded on a plastic dish for proliferation. Instead of density gradient centrifugation many groups are using adherence for isolation of MSC from BM.

AT-MSC also termed as adipose-derived stem cells (ASC) are usually isolated from the biological material generated during liposuction, lipoplasty, or lipectomy procedures by enzymatic digestion with collagenase followed by centrifugation and washing 7 .

PB-MSC can be obtained from the lymphocyte separation fluid fraction of mononuclear cells after a density gradient centrifugation 3 . Another method is described by Kassis et al. in which PB-MSC are isolated from the mononuclear fraction by loading PB-MSC on fibrin microbeads followed by separation of the cell loaded beads 8 .

The amounts of mesenchymal stem cells which can be obtained by these isolations vary enormously. Pittenger et al. isolated MSC from BM by density gradient centrifugation to eliminate unwanted cell types and only 0.001 to 0.01% of the cells isolated from the density interface were mesenchymal stem cells 6 . From 1g of adipose tissue 5 × 10³ stem cells can be isolated, which is 500 times more cells than from an equivalent amount of bone marrow 2 9 . PB-MSC exhibit a colony forming efficiency (CFE) ranging from 1.2 to 13 per million mononuclear cells 10 .

Suggested minimal criteria to define human MSC were published in 2006 by Dominici et al.. These suggested criteria included positive expression of CD105 (SH2), CD73 (SH3), CD44 and CD90 and negative expression of CD45, CD34, CD14 or CD11b, CD79α or CD19 and HLA-DR surface molecules. Furthermore, mesenchymal stem cells exhibit plastic-adherence under standard culture conditions and are competent for in vitro differentiation into osteoblasts, chondroblasts and adipocytes 11 .

Several publications demonstrate the reproducible expression of important stem cell markers such as CD44, CD73 (SH3), CD90, CD105 (SH2) and CD166 and the absence of the hematopoietic markers CD14, CD34 and CD45 in BM-MSC and in Wharton's Jelly-derived MSC 6 12 . Schaffler et al. defined the surface marker set for AT-MSC (ASC) as positive CD9, CD29, CD44, CD54, CD73 (SH3), CD90, CD105 (SH2), CD106, CD146, CD166 and HLA I expression and negative CD14, CD31, CD34, CD45, CD133, CD144, HLA-DR, STRO-1 and HLA II expression 13 . De Ugarte et al. performed flow cytometry analyses of BM-MSC and AT-MSC (ASC) and found that both cell types express CD13, CD29, CD44, CD90, CD105 (SH2), CD73 (SH3), and STRO-1. But they found differences in the expression of CD49d, CD54, CD34 and CD106 between the two cell types 14 . Zuk et al. also compared the marker profiles of BM-MSC and AT-MSC (ASC) and found similar expression of CD29, CD44, CD71, CD90, CD105 (SH2) and differences in the expression of CD49d and CD106. AT-MSC (ASC) express CD49d, in contrast to BM-MSC; BM-MSC however, express CD106, which could not be detected in AT-MSC (ASC) 15 . The discrepancies in the expression of STROH-1 and CD34 of BM-MSC and AT-MSC in the different studies may be caused by different isolation methods or different media compositions used which can result in a different expression of surface molecules.

PB-MSC express CD44, CD54, CD105 (SH2) and CD166, but not CD14, CD34, CD45, or CD31 3 . Kassis et al. isolated PB-MSC which are positive for the expression of CD90 and CD105 (SH2) and negative for CD45 and CD34 7 . Tondreau et al. performed flow cytometric analyses of BM-MSC and PB-MSC and revealed similar expression patterns, namely the presence of CD44, CD105 (SH2), and CD73 (SH3) and the absence of CD14, CD34, CD45 and HLA-II 16 .

The described findings of cell surface marker expression analyses are summarized in table 1.

In addition to distinct adult tissues (adipose tissue, bone marrow, peripheral blood), MSC can be obtained from several birth-associated tissues including placenta, amnion, umbilical cord (UC) and cord blood (CB) (Figure 1). A significant advantage of these neonatal tissues is their ready availability, thus avoiding invasive procedures and ethical problems. Moreover, birth-associated tissues harbor a variety of embryonic or premature cell populations including MSC, endothelial stem/progenitor cells (EPC, ECFC) and hematopoietic stem cells (CD34⁺, CD133⁺). It is also suggested that MSC from these neo-natal tissues may have additional capacities in comparison to MSC derived from adult sources. Indeed, several studies have reported superior cell biological properties such as improved proliferative capacity, life span and differentiation potential of MSC from birth-associated tissues over BM-MSC.

For example, MSC from the human placenta (PL-MSC) have been reported to have a higher expansion and engraftment capacity than BM-MSC 17 18 19 . In this context, it is important to note that placental tissue can be fetal or maternal in origin requiring the two types of tissue to be individually characterized with respect to MSC function. According to the first international workshop on placenta-derived stem cells, four regions of fetal placenta can be discriminated: amniotic epithelial, amniotic mesenchymal, chorionic mesenchymal, and chorionic trophoblastic tissue. Consequently, at least four different cell populations with stem or progenitor properties can be distinguished: human amniotic epithelial cells (hAEC), human amniotic mesenchymal stromal cells (hAMSC), human chorionic mesenchymal stromal cells (hCMSC), and human chorionic trophoblastic cells (hCTC) 20 . Placenta-derived MSC from fetal tissue, including amnion membrane ((AM-MSC) (or HAM; human amniotic membrane)) 21 22 23 24 , chorion membrane (CM-MSC) 25 26 and chorionic villi (CV-MSC) 27 28 29 have been have been described as having a more limited life span than MSC populations obtained from the maternal part of the extraembryonic membranes or decidua (D-MSC) 25 26 30 (Figure 1). Moreover, clonal subpopulations of D-MSC have been attributed with the potential to differentiate into tissues from all three germ layers 31 . A similarly high cellular plasticity for differentiation into the three germ layers has also been described for MSC derived from amniotic fluid (AF-MSC) 32 , amniotic epithelial cells 22 and endometrial regenerative cells 33 .

A certain heterogeneity within the stromal or stem cell population displaying mesenchymal-like characteristics such as surface marker expression, plastic adherence, self renewal and differentiation capacity has also been identified in MSC derived from the umbilical cord (UC-MSC). Separation of UC-MSC by counterflow centrifugal elutriation resulted in differentially sized subpopulations displaying altered proliferation potentials which were associated with significantly different amounts of senescent cells 34 .

With respect to MSC isolation from umbilical cord, the different parts of this tissue should also be considered individually 35 36 . MSC can be isolated from whole umbilical cord 37 , from Wharton's jelly (WJ-MSC) 38 39 40 or from umbilical cord blood (CB-MSC), 41 42 which also harbors hematopoietic stem cells, endothelial precursor cells and endothelial colony-forming cells (Figure 1). Proteome analysis of WJ-MSC revealed differences in the protein expression pattern during in vitro self-renewal 43 and other work has demonstrated that UC-MSC represent a preferred population for musculoskeletal tissue engineering 44 . Like PL-MSC and other neonatal birth-associated MSC, the UC-MSC exhibit certain cell biological properties which are different from MSC originating from adult sources (AT-MSC (ASC), BM-MSC, PB-MSC).

The proliferation capacity and senescence of these cells have been analyzed by many scientists over the last few years. The proliferation capacity of cells is important with regard to their application in cell therapy and tissue engineering. Baksh et al compared umbilical cord perivascular cells (UCPVC) to BM-MSC and determined that the UCPVCs also have a higher proliferation capacity than the BM-MSC 45 . Moreover, various papers have been published demonstrating that UC-MSC exhibit a higher proliferation capacity than BM-MSC 46 47 48 49 . Lu et al. performed proliferation studies with BM-MSC and UC-MSC which revealed that BM-MSC showed significantly slower population doubling times. The mean doubling time of the UC-MSC in passage 1 (P1) was about 24h and remained almost constant up to P10. In contrast the mean doubling time of BM-MSC was 40h and increased considerably after P6 49 . They determined the population doublings over 20 days and observed that after three days both cell types showed similar population doubling times, but that from day seven on the population doubling time of the UCPVCs was significantly increased 46 . Additionally, they found that the UCPVCs continued to grow by multi-layering, in contrast to the proliferation of BM-MSC that was inhibited due to contact inhibition. AT-MSC (ASC) have also been shown to have higher proliferation capacities than BM-MSC 50 . Peng et al. described population doubling times of 45.2 h for AT-MSC (ASC) and 61.2 h for BM-MSC. Moreover, they revealed that the BM-MSC were morphologically larger as compared to AT-MSC (ASC) 51 . It should however be noted that differences in the doubling times of AT-MSC (ASC) originating from different regions of the body have been reported 52 53 . Van Harmelen et al. published that AT-MSC (ASC) from the subcutaneous adipose tissue region proliferated faster (doubling time, 4 +/- 1 days) than those from the omental region (doubling time, 5 +/- 1 days) 54 . In addition to the origin of the cells, the cultivation conditions and various medium supplements may have an effect on doubling times of the AT-MSC (ASC). Own experiments revealed shorter doubling times for AT-MSC (ASC) cultured in human serum instead of fetal calf serum (unpublished data).

Besides the higher proliferative activity of UC-MSC the cells show no sign of senescence over several passages 55 56 . Conconi et al. cultured UC-MSC over 16 serial passages and found no variation in cell morphology or senescence 57 . Mitchell et al. cultured porcine UC-MSC for more than 80 doubling times with no decrease of proliferative capacity 58 . Kern et al. investigated the senescence ratio of AT-MSC (ASC) in comparison to BM-MSC. AT-MSC (ASC) could be cultivated up to passage number 8 without any sign of senescence whereas in BM-MSC senescence was demonstrated already in cells from passage number 7 50 .

The differentiation of UC-MSC and AT-MSC (ASC) along the adipogenic, chondrogenic and osteogenic lineages has been investigated by many working groups. Furthermore, in vitro differentiation into cardiomyocytes 40 59 , endothelial cells 48 60 or neuronal cells 61 62 has been reported.

Differences in cell functions between MSC populations derived from adult or neonatal tissues are also influenced by the microenvironment. Within the appropriate tissues in vivo, stem cells like MSC are usually present in stem cell niches under hypoxic conditions. Therefore, in vitro primary culture in a normoxic atmosphere (21% O₂) can be considered as an exposure to enhanced oxidative stress and promotes the generation of metabolic radicals or reactive oxygen species (ROS). The intracellular accumulation of ROS can cause protein and DNA damage if these compounds are insufficiently metabolized by an appropriate anti-oxidative defense system. Consequently, ROS accumulation at high oxygen levels induces elevated apoptosis and premature aging by STASIS (stress or aberrant signaling-inducing senescence) 80 . Indeed, MSC cultured under normoxic conditions exhibit premature senescence and a reduction in population doublings in comparison to cells cultured under hypoxia 81 82 and may also show restricted cell division due to telomere shortening and replicative senescence 80 83 . The migratory capability of MSC cultured under hypoxic conditions is also enhanced in contrast to that seen in normoxia 84 . Hypoxic conditions therefore influence proliferation and cell fate commitment, meaning that gradients of oxygen tensions influence the prolonged maintenance of a stem cell phenotype and pluripotency 85 . Moreover, serum starvation and deprivation of growth factors can promote premature aging in MSC (Figure 2) and studies of MSC in a hypoxic environment show that serum starvation can be associated with massive cell death 86 .

Previous work has demonstrated that the culture of MSC under hypoxic conditions is accompanied by increased Oct4 expression and telomerase activity 81 which are involved in the maintenance of stemness. Other studies have identified changes in the transcriptome of hypoxic stem cells that seem to indicated that hypoxic conditions rather reflect a physiologically normoxic environment for the cells 87 .

Hypoxic conditions induce the transcription factor hypoxia-inducing factor-α which can promote certain differentiation phenotypes in MSC. Thus, chondrogenic differentiation of AT-MSC (ASC) has been observed at enhanced levels under hypoxic conditions where osteogenesis is inhibited. In contrast, enhanced osteogenic differentiation of AT-MSC (ASC) can be induced under normoxia 88 89 . Moreover, other MSC populations including UC-MSC exhibit differences in energy turnover and the expression of energy metabolism-associated genes at different hypoxic conditions 90 . Functional changes of MSC under hypoxia also include increased secretory activity, i.e. of vascular endothelial growth factor and interleukin-6 as well as mobilization and homing by the induction of stromal cell-derived factor-1 expression and the corresponding receptor CXCR4 91 . In this context, MSC subpopulations displaying a high aldehyde-dehydrogenase activity have been reported with increased responsiveness to hypoxia, including an upregulation of Flt-1, CXCR4 and angiopoietin-2 92 .

Together, these findings further substantiate that oxygen tensions contribute to the regulation of MSC function and fate. Whereas MSC can display and maintain a hypoxic microenvironment within the appropriate tissues in vivo, these conditions may also contribute to the control of other important cellular functions of MSC including their immune-modulatory properties.

Two outstanding features of MSC are relevant to immunity: 1) immunosuppression and 2) the so called immunoprivilege. What do these terms mean? MSC-mediated immunosuppression describes the fact that MSC are able to suppress several functions exerted by diverse immunocytes such as T-, B-, and NK cells. The affected functions comprise proliferation, production of soluble factors (e.g. cytokines), and cellular cytotoxicity (Figure 3). Immunoprivilege means that MSC themselves are somehow protected from immunological defence mechanisms. Undoubtedly, there is much truth in the reports of MSC-mediated immune effects. Nevertheless, there also seems to be some conflicting data. The inconsistencies between some reports may, however, be due to the population diversity of the primary cultures and to the tissue- and species-origin of the MSC tested. As mentioned above, MSC are currently characterized using a minimum of surface markers, which might not be sufficient for their precise definition. A further cause for conflicting data could be the source (adipose tissue, blood, bone marrow, umbilical cord, umbilical cord blood) for isolation of the MSC. In order to review the immunological findings in MSC research we screened the literature and compared the data on immunological interactions between MSC and immune cells.

MSC within a primary culture can also exhibit different states of activation which can be related to the expression levels of certain micro RNAs (miR) including miR335 132 . miR are small non-coding RNAs of about 20 to 22 nucleotides, which, upon sequence-specific binding to mRNAs, repress the translation of the corresponding proteins or induce a subsequent degradation of the miR/mRNA complexes.

A variety of different miR play an important role in regulating differentiation pathways and cell fate in MSC which recently has been reviewed by Guo et al. 133 . For example, osteogenic differentiation of MSC can be blocked by miR-125b, miR133, miR135 and miR206 which attenuate the expression of ERBB2, RUNX2, Smad5 and connexin-43, respectively. Likewise, expression of further specific miR are involved in the regulation of adipogenic and chondrogenic differentiation and pathways beyond the mesodermal lineage 133 . Moreover, miR are also involved in the regulation of replicative senescence and wound healing of MSC. Thus, miR which target distinct DNA-methyl transferases can promote senescence of MSC 133 . Although the molecular mechanisms of MSC senescence after a limited number of cell divisions are still poorly understood, cell fusion processes which are known for MSC or asymmetric cell divisions may also contribute to this phenomenon which enables the segregation of daughter cells committed to either senescence or retaining reproductive capacity in correspondence to the parental cells 134 .

Furthermore, MSC can secrete micro vesicles which contain certain pre-microRNAs 135 . The released exosomes facilitate cell-to-cell communications and thus, can alter cell activities in target cells.

A proposed MSC model suggested that high miR-335 expression contributes to a potential non-activated (silenced) MSC auto-maintenance state, in contrast to low levels of miR-335 which produce an activated state leading to proliferation, migration and differentiation in MSC 132 . Of interest, a functional role in the regulation of epithelial-to-mesenchymal transition and a neoplastic development in breast tissue has also been attributed to miR-335 136 . Moreover, MSC display some similarities to normal and tumorigenic human breast epithelial cells with respect to the gene expression pattern 137 and some surface receptor levels 138 139 . Whereas the location of MSC in the adipose tissue of the breast adjacent to mammary epithelial cells enables interactions by stimulatory cytokines and/or miR-containing micro vesicles, these stimulatory effects suggest a close functional relationship between these cell types 140 . Indeed, previous work has demonstrated that although MSC themselves do not develop teratoma even when derived from a teratoma-forming human embryonic stem cell line 141 , a close vicinity to neoplastic breast epithelial cells within the tissue microenvironment can stimulate growth and metastasis of breast cancer cells by cytokines including CCL5 (Rantes) 142 and may most probably also influence the exchange of miR-containing micro vesicles. Thus, synergistic effects of MSC in cooperation with other cell types, e.g. tumor cells must be considered and require further elucidation.

MSC represent an important stem cell population with multipotent capabilities which are extremely useful for clinical applications. Although certain discrepancies within the MSC literature result in differing descriptions of the biological properties of MSC, these effects may be explainable in part by the existence of distinct subpopulations within a tissue-derived primary culture that exhibit some variation in function 33 34 92 . Moreover, different isolation methods of MSC, particularly the use of proteases to digest the extracellular matrix for an enrichment of the stem cells, may alter MSC functions by non-specific degradation e.g. of certain surface receptors, whereas the explant culture of MSC from tissue pieces avoids such potential artifacts.

In summary, MSC can self renew to a certain extend and differentiate (Figure 4). Moreover, they can display a variety of important cell functions in the organism including migration and transport functions to sites of local injuries or tissue damage to support appropriate cell and tissue renewal to replace the damaged areas (Figure 4). Concomitantly, MSC are non-immunogenic due to their immune-modulatory capabilities and no teratoma formation of MSC after allogenic human transplantations has been observed to date (Figure 4) which indicates an enormous potential for the clinical use of these cells, particularly in regenerative medicine.

Whereas a variety of different tissue sources for MSC have been described, MSC from birth-associated tissues, preferably parts of the placenta (i.e. D-MSC) and the umbilical cord/Wharton's jelly (UC- and WJ-MSC) may offer certain advantages. These include their non-invasive and ethically non-problematic availability. More importantly, MSC from these neonatal tissues possess increased proliferative capacity in vitro, especially under hypoxic conditions, in comparison to some MSC populations obtained from adult tissues. Quiescent stem cells within their niches of various tissues can be activated if required, however, even a reprogramming of cells via a retrodifferentiation program 143 or further processes to rejuvenate cells to a more juvenile and undifferentiated phenotype 134 are often not sufficient to cope with tissue requirements after injury or disease-associated tissue damage and degeneration. Therefore, in contrast to the limitations of bone marrow or adipose tissue, MSC from birth-associated tissues can be obtained in large quantities, and the required numbers of these stem cells can be transplanted in therapeutic approaches for tissue replacement.

Taken together, multifunctional MSC from parts of the placenta and the umbilical cord may represent a very promising stem cell population in regenerative medicine.

The authors declare that they have no competing interests.

CK and SB contributed the adult MSC part including Figure 2. RJ focused on the immune functions of MSC and provided Figure 3. RH contributed the other parts including the focus on embryonic/neonatal tissue-associated MSC, hypoxia and miRNA. Furthermore, RH provided Figures.1 and 4 and drafted the manuscript. All authors have read and approved the final version of the manuscript.