In this study, we used a common procedure to assess the potential of mobilized peripheral blood (MPB) and umbilical cord blood (UCB) as sources of mesenchymal stem cells (MSCs) in comparison with bone marrow (BM). We tested three methods: plastic adhesion supplemented with 5% of BM-MSC conditioned medium, unsupplemented plastic adhesion, and selection of CD133-positive cells. MSCs derived from MPB or UCB are identified by their positive expression of mesenchymal (SH2, SH3) and negative expression of hematopoietic markers (CD14, CD34, CD45, HLA-DR). We observed that the CD133-positive cell fraction contains more MSCs with high proliferative potential. Placed in appropriate conditions, these cells proved their capacity to differentiate into adipocytes, osteocytes, chondrocytes, and neuronal/glial cells. MPB- and UCB-MSCs express Oct4, a transcriptional binding factor present in undifferentiated cells with high proliferative capacity. The selection of CD133-positive cells enabled us to obtain a homogeneous population of MSCs from UCB and MPB. These sources may have a major clinical importance thanks to their easy accessibility.
It has been shown in recent years that human bone marrow (BM) contains hematopoietic stem cells, responsible for the hematopoietic turnover, and mesenchymal stem cells (MSCs), giving to these hematopoietic stem cells an appropriate microenvironment [1, 2]. MSCs are defined as adult immature cells capable of self-renewing and of differentiating into various tissues in vivo and in vitro [3–6]. After being transplanted into a mouse model or in utero into sheep, MSCs were then engrafted in different tissues including bone, muscle, brain, lung, heart, and liver tissue, without inflammatory response, thanks to the absence of HLA class II expression. Moreover, experiments have proven that MSCs may decrease graft-versus-host disease (GVHD) by inactivating and inhibiting the proliferation of T lymphocyte [7–10]. Recent studies have also demonstrated clinical applications of MSCs to improve hematopoietic engraftment, to prevent GVHD, and to correct genetic disorders such as osteogenesis imperfecta [11, 12]. There is also a great interest in the potential use of MSCs in molecular therapy to deliver oncolytic viruses [13–15]. BM is the major source of MSCs, but this population is rare (0.001%–0.01%). Recently, MSCs have also been isolated from adipose tissues, muscle, fetal organs, brain, and teeth [16–20]. Different techniques can be used to obtain MSCs from the BM: plastic adhesion and negative (CD45, Gly-A, and RosetteSep) or positive selection (CD49-a, Stro-1, and CD133) [21–26]. However, their presence in mobilized peripheral blood (MPB) or in umbilical cord blood (UCB) still remains controversial. A few authors have demonstrated the presence of stromal or mesenchymal progenitor cells in peripheral blood or in the growth factor-mobilized apheresis products [27,28]. In their experiments, the expanded adherent cells have shown characteristics similar to those of BM-MSCs. In these studies, MSCs were isolated by the classic plastic adhesion method and subcultures. Using the same procedure, other groups have attempted to isolate MSCs from peripheral blood without success . Plastic adhesion and negative immunodepletion have been suggested to isolate MSCs, but it has been reported that only midtrimester and no full-term cord blood may contain MSCs [30–31]. In the present study, we assessed the presence of MSCs in MPB and in full-term UCB in comparison with BM. We compared the classic plastic adhesion method and subcultures with cultures containing BM-MSC conditioned medium (BM-MSC CM) and finally with CD133+ initiated cells. We observed that in these three different culture systems, MSCs can easily be isolated and expanded from MPB and UCB. The CD133-positive fraction contains more MSCs with high proliferative potential. MSCs isolated from MPB and UCB show the characteristic pattern of mesenchymal surface markers and express Oct4 (Octomer-binding Transcription Factor 4), a marker of pluripotent stem cells . We also demonstrated their potential for multilineage differentiation, because these cells can become bone, cartilage, fat tissue, and neuronal/glial cells under specific induction media.
Materials and Methods
BM cells were obtained from sternal or iliac crest marrow aspiration of healthy donors (n = 6) after informed consent. We separated cells on Ficoll density gradient (LinfoSep; Biomedics, Madrid, Spain, http://www.biomedics.es), and the mononuclear cell (MNC) fraction was collected and washed in Hanks' balanced saline solution (HBSS; Cambrex Bio Science, Verviers, Belgium, http://www.cambrex.com). We seeded, in primoculture (PM), 5 × 104 cells per cm2 in a six-well plate in complete alpha–modified Eagle's medium (α-MEM; Cambrex Bio Science) supplemented with 10% fetal bovine serum (Biochrom, Berlin, Germany, http://www.biochrom.de), 2 mM L-glutamine (GibcoBRL, Grand Island, NY, http://www.invitrogen.com/content.cfm?pageid=9371), and 0.5% antibiotic-antimycotic solution (GibcoBRL). Every 3–4 days, we removed the medium until subconfluency was obtained, and cells were harvested with trypsin-EDTA solution (GibcoBRL) and replated at 1,000 cells per cm2.
Conditioned media were prepared after two passages (P2), when cells express more than 90% of putative mesenchymal markers (SH2 and SH3). BM-MSC supernatants were collected after 72 hours and stored at −20°C for further use.
MPB-MSC and UCB-MSC Cultures
Peripheral blood cells (n = 6) from healthy donors were collected at day 4 of G-CSF mobilization (Neupogen; Amgen, Thousand Oaks, CA, http://www.amgen.com). Informed consent was obtained from every donor. Full-term UCB samples (n = 8) were obtained with consent of the mothers within 24 hours of collection. MNCs were separated on Ficoll gradient as described for BM. We seeded 1 × 108 cells in a 75-cm2 flask in complete α-MEM supplemented, or not supplemented, with 5% CM during the first 48 hours of PM. CD133-positive cells were isolated from 1 × 108 MNCs obtained from MPB and UCB, using magnetic beads coated with anti-CD133 antibody according to the manufacturer's instructions (MACS; Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com). Purified cells were seeded at 2.5 × 104 per cm2 in complete α-MEM.
Differentiation of MSCs
Osteogenic, adipogenic, chondrogenic, and neurogenic inductions were assessed after the second passage of MPB- and UCB-MSCs. One thousand cells per ml were plated in a 24-well plate during 2–3 weeks under specific mesodermal conditions for osteogenic, adipogenic, and chondrogenic induction and during 10 days for neurogenic induction.
Medium for osteogenic induction: complete α-MEM supplemented with 60 μM ascorbic acid (Sigma, St. Louis, http://www.sigmaaldrich.com), 10 mM β-glycerophosphate (Sigma), and 0.1 μM dexamethasone (AAcidexam; Aaciphar, Brussels, Belgium). Calcium deposits were stained using the Von Kossa method (Sigma).
Adipogenic medium: complete α-MEM supplemented with 1 μM dexamethasone (Aaciphar), 60 μM indomethacin (Sigma), and 5 μg/ml insulin (Sigma). The formation of lipid vacuoles was assessed by Oil Red O staining (Sigma).
Chondrogenic medium: Dulbecco's modified Eagle's medium (DMEM; Cambrex Bio Science) supplemented with 1% fetal calf serum, 0.5 μg/ml insulin (Sigma), 50 μM ascorbic acid (Sigma), and 10 ng/ml transforming growth factor-β (SanverTech, Boechout, Belgium, http://www.tebu-bio.com). Cells were cultured in the tip of a 15-ml conical tube to allow cell culture in a micromass. Every 3–4 days, the medium was removed by low-speed centrifugation (10 minutes at 400g). After 2–3 weeks, spheroid cell mass was fixed and embedded and cuted by microtome before toluidin blue staining (Sigma).
Neurogenic medium: neural progenitor cell basal medium (NPBM) supplemented with 5 μM isobuthylmethylxantine (IBMX), 2.5 μg/ml insulin (Sigma), 5 μM cAMP (Sigma), and 25 ng/ml nerve growth factor (Sigma). The expression of immature or mature neural/glial proteins (Nestin, Tuj-1, glial fibrillary acidic protein [GFAP], microtubule associated protein 2 [MAP-2]) was evaluated through confocal microscopy or reverse transcription (RT)–polymerase chain reaction (PCR).
Cell surface markers were evaluated through direct or indirect immunofluorescence: CD14– fluorescein isothiocyanate (FITC) (DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com), CD33-FITC (DakoCytomation), CD34-FITC (BD Biosciences, Erembodegem-Aalst, Belgium, http://www.bdbiosciences.com), CD44 (Immunosource, Halle-Zoersel, Belgium, http://www.immunosource.com), CD45-PE (Sigma), CD105 (DakoCytomation), CD133-PE (Miltenyi Biotec), HLA-DR (Immunosource), SH2 (HB-10743 cell line; American Type Culture Collection, Manassas, VA, http://www.atcc.org), SH3 (HB-10744 cell line; ATCC), βIII tubulin (Sigma), MAP-2 (Sigma), and Ig-G1-PE (antiF(ab')2 mouse-PE; Dako Cytomation). Adherent cells were detached with trypsin-EDTA solution and washed in phosphate-buffered saline. For direct labeling, cells were incubated with 1 μg/ml of antibody for 30 minutes at room temperature. For indirect fluorescence, cells were first incubated with the primary antibody and then with 1 μg/ml of the secondary antibody for 45 minutes at room temperature. Intra-cellular antigens were detected after a fixation/permeabilization step using aceton/methanol solution (v/v) for 5 minutes at −20°C. Cells were fixed with 4% paraformaldehyde and analyzed using a Coulter Epics XL flow cytometer (Beckman, Coulter, Miami, http://www.beckmancoulter.com).
Colony-Forming Unit-Fibroblast Assay
We seeded 1 × 105 MNCs and 5 × 103 cells, obtained after each passage, in 8-cm diameter Petri dishes containing complete α-MEM. After 10 days of culture, we counted fibroblastic colonies with more than 50 cells, using a May Grünwald-Giemsa coloration (Merck, Darmstadt, Germany, http://www.merck.com).
We extracted total RNA from each cell culture, using the Tripure method (Roche Diagnostics, Indianapolis, http://www.roche-diagnostics.com). Genomic DNA was degraded from RNA samples prior to applications of RT-PCR, using a DNase (RQ1 Rnase-Free Dnase; Promega, Madison, WI, http://www.promega.com). We performed the RT with 1 μg RNA, using the M-MLV RT during 45 minutes at 42°C and then 3 minutes at 99°C (Invitrogen, Merelbeek, Belgium, http://www.invitrogen.com). Briefly, we used a final volume of 20 μl containing 1 μg RNA, 4 μl first strand buffer, 10−2 M DTT, 1 mM of each dNTP, 50 U Rnase inhibitor, and 100 U M-MLV RT. Then we carried out a PCR in a final volume 50 μl containing 5 μl of cDNA diluted ×5, 5 μl PCR buffer, 30 pmoles of upstream sense and downstream sense primers, 0.5 mM dNTP, and 2.5 U Taq DNA polymerase. Cyclic parameters were denatured at 94°C for 1 minute, then annealed at 50°C for 1 minute, and finally elongated at 72°C for 1 minute. Forty cycles were performed for each primer set. We resolved the PCR products on a 2% gel agarose.
Primer sequence and length of amplified products were: Oct4 (218pb) sense: gacaacaatgaaaatcttca-3′ and antisense: ttctggcgcc-ggttacagaa-3′, GAPDH (403 pb, internal control) sense: gagctt-gacaaagtggtcgt-3′ and antisense: cccaagtccatgccatgccatcact-3′, GFAP: (203 pb) sense: gagtaccaggacctgctcaa-3′ and antisense: ttcaccacgatgttcctctt-3′ and Nestin (375 pb) sense: tctccagaaact-caagcacc-3′ and antisense: ggtcctctccaccgtatct-3′.
Characterization of MPB- and UCB-MSCs
The first method described to isolate BM-MSCs was based on their strong capacity to adhere to plastic culture flasks . Cells obtained in this condition enable us after successive passages to generate a homogeneous population of MSCs capable of differentiating in several tissues (e.g., adipocytes, osteocytes, chondrocytes, neurons, and myocytes). In this study, we assessed the presence of MSCs in MPB and in UCB. In spite of many studies, their presence in both these sources is still controversial. We tested different culture procedures to evaluate the presence of MSCs in MPB and UCB in comparison with BM. MPB- and UCB-MSCs were isolated either by classic adhesion method or after CD133 cell selection. Because MSCs are capable of secreting cytokines and growth factors essential for their proliferation and growth and because they are sparse in UCB and MPB, we postulated that BM-MSC CM could be helpful to accelerate the proliferation of MSCs in these sources. During the first 48 hours of culture, we incubated MPB or UCB mononucleated cells in the presence of 5% of BM-MSC CM.
After CD133 selection, the cell fraction seeded in PM was composed of 90.4% ± 1.03% of CD133+ cells (data not shown). More than 89.5% ± 2.4% of CD133+ cells coexpressed the CD34 antigen, a hematopoietic progenitor marker (Fig. 1).
During a 15-day PM, the majority of adherent cells observed in MPB or UCB, whatever the method used, were comprised of hematopoietic cells, monocytes/macrophages, endothelial cells, and osteoclasts. The addition of CM during the first 48 hours of PM generated more nucleated cells, but similar cell composition was observed. After the second passage, adherent cells were constituted by a homogeneous population of MSCs. Through fluorescence-activated cell sorting analysis, we demonstrated that the profile of expression of UCB- and MPB-MSCs was similar to BM-MSCs. Cells were positive for SH2, SH3, CD105, and CD44, but negative for CD14, CD34, CD45, and HLA-DR (Fig. 2).
UCB- and MPB-MSC Plasticity
Another characteristic of MSCs is their capacity to differentiate in vitro into different tissues after specific induction. In this study, we evaluated the potential of MSCs derived from MPB or UCB to differentiate into adipocytes, osteocytes, chondrocytes, and neuronal cells. Two weeks after mesodermal induction (adipogenic, osteogenic, and chondrogenic media), we assessed lipid vacuoles, calcium deposits, and chondrogenic matrix using Oil Red O, Von Kossa, and Toluidin blue, respectively (Figs. 3A–3D). Ten days after neurogenic induction, cells displayed typical neuron-like morphology and were positive for Nestin, Tuj-1, MAP-2, and GFAP (Figs. 3E–3G).
Colony-Forming Unit-Fibroblast Potential Study
MSCs are characterized by their ability to form colonies comprising nonrefringent spindle-shaped cells deriving from a single cell (colony-forming unit-fibroblast [CFU-F] assay). This assay was used to evaluate the cloning efficiency (number of CFU-F/5,000 seeded cells) of MSCs obtained from BM, MPB, and UCB. No colony was observed in MPB- and in UCB-MNCs. However, we obtained few fibroblastic colonies in the case of BM MNCs (3.5 ± 1.5 CFU-F). This last result confirms the very low frequency of MSCs in MPB and UCB. After the PM of MPB and UCB, CFU-F colonies appeared and we counted them. Interestingly, we observed that a great number of CFU-F can be obtained when CD133+ cell selection was performed, demonstrating that CD133-positive cell fraction is rich in MSC clonogenic precursors. When CM was added during the first 48 hours of culture, we also observed a slight increase in the number of CFU-F. After two passages (P2), whatever the method used to isolate MSCs from MPB and UCB, a similar level of CFU-F was observed (Fig. 4).
To evaluate the expansion of MSCs derived from UCB and MPB, we calculated the number of cells seeded in primary culture to 1 × 105 cells per cm2 for each cell culture condition. Generally, MSC cultures were initiated from 20 ml BM aspirates and, in the case of UCB, from samples of more than 50 ml of blood. A large volume of blood and a great quantity of MNCs can be collected from MPB (Table 1). In these sources, MSCs are sparse, and to increase the probability of obtaining MSCs, more cells were seeded in PM. Indeed, 1 × 106 MNCs per cm2 were required to initiate the culture from UCB and MPB, versus 1 × 104 cells per cm2 for BM. The growth curves of MSCs from the three sources can be observed in Figure 5. For each passage, we seeded 100 cells per cm2 and made comparisons with the starting fraction. We observed a great variability between BM samples (BM1 versus BM2). BM1 contained more MSCs than BM2, but BM1 corresponds to particulate fraction removed by nylon filtration prior to transplantation, which is rich in hematopoietic and stromal cells . BM2 is obtained though classic sternal aspiration. The culture of CD133+ cells from UCB and MPB enabled us to obtain a great quantity of MSCs in comparison with the culture of MNCs in complete α-MEM supplemented, or not supplemented, with 5% CM (Figs. 5A, 5B). After fourth passages (P4), 6.8 × 108 cells and 3.3 × 107 cells were obtained from UCB and MPB, respectively, from the CD133+ cell fraction. Using plastic adhesion to expand BM MSCs, 3.8 × 106 to 7.9 × 109 cells were obtained from BM2 and BM1. These results demonstrated that the CD133+ cell fraction contains more MSCs with high proliferative potential. The decrease of cell number observed between PM and P1 culture is associated with the elimination of hematopoietic cells from the culture.
Table Table 1.. Characteristics of samples
aThe volume corresponds to the sample received for experimental research.
Oct4, the octomer-binding transcription factor 4, is an important binding transcription factor present in undifferentiated embryonic stem cells with a high proliferative capacity, but also detected at a lower level in multipotent adult progenitor cells (MAPCs) . Pochampally et al.  recently selected a subpopulation of early progenitor cells in a serum-deprived culture of human marrow stromal cells with increased expression of Oct4. In this study, we evaluated Oct4 expression in MNCs and MSCs obtained from BM, MPB, and UCB (Fig. 6). All MSCs isolated by plastic adhesion method were assessed after the second passage, when more than 90% of cells express mesenchymal markers (SH2, SH3) but not hematopoietic markers (CD14, CD34, CD45), and these MSCs are multipotent. We observed that MNCs have no transcript for Oct4, whereas MSCs derived from all three sources (BM, MPB, and UCB) express Oct4. This result confirms the presence of MSCs with high proliferative capacity in MPB and in full-term UCB.
BM represents the main source of MSCs with a low frequency of 0.001%–0.01%, depending on the age of donors [35, 36]. These elongated fibroblast-like cells have the capacity to self-renew and differentiate into multiple lineages of mesodermal tissues such as cartilage, bone, fat, tendon, muscle, myocardium, and stromal cells. Endodermal (endothelial cells) and ectodermal (neural cells) lineages have also been obtained in vitro after specific induction [5, 33, 37]. Because of the high potential of these cells to be subcultured and differentiated in vitro, they have aroused a great interest in their use for cell and gene therapy. In the present study, we evaluated two other potential sources of mesenchymal cells: MPB and full-term UCB.
UCB and MPB are currently used as a source of hematopoietic cells and progenitor cells after chemotherapy, and their application as an MSC source is still debated. Previous studies have demonstrated that preterm UCB contains more adherent fibroblastoid cells than full-term UCB but that only 25% of UCB harvested gave rise to mesenchymal-like cells [31, 38]. MSCs can also be isolated from the subendothelial layer of the umbilical cord vein but not in full-term UCB [39, 40]. Recently, Lee et al. obtained homogeneous plastic adherent cells from the MNC fractions of cryopreserved UCB . These cells exhibited fibroblast-like morphology and typical mesenchymal phenotype. The cause of these divergent results could be related to differences in culture conditions required for the growth of UCB-MSCs. Recently, critical parameters for the isolation of MSCs from UCB have been described. The time between collection and isolation (<15 hours), the volume of samples, and the count of MNCs seem to be crucial for obtaining MSCs from full-term UCB . In MPB, similar controversial results have been reported about the effective presence of MSCs. MPB-MSCs have been described in blood collections from breast cancer patients and in normal volunteer blood. In these studies, MSCs were also isolated though the plastic adhesion method. After several passages, the adherent cells exhibited the same morphology and phenotype as BM-MSCs. However, using the same procedure, no MSCs have been observed by other groups [12, 27–29]. In our study, we developed different methods for obtaining MSCs from full-term UCB or MPB. In the first, we used the classic plastic adhesion and subcultures. In the second, we added 5% of CM during the first 48 hours of PM. CM was obtained from BM-MSC cultures containing a homogeneous population with more than 90% of cells expressing the putative mesenchymal markers (SH2, SH3). Indeed, MSCs produce their own cytokines and growth factors essential to their development. These findings prompted us to evaluate if the addition of BM-MSC CM can modulate the growth of MSCs in UCB and MPB samples. The third method was the immunoselection of CD133+ cells directly from fresh samples. CD133+ cells are considered as a population of noncommitted early progenitors capable of self-renewing and differentiating into blood cells and other cell types. Thus, CD133 seems to be a marker associated with more primitive stem cell phenotype than CD34 . Using the purified CD133-positive cell fraction, we obtained a large quantity of MSCs after the PM. After four passages, the number of cells harvested was similar to BM-MSCs. These results suggest that the CD133-positive cell fraction contains both hematopoietic and mesenchymal stem cells. Indeed, more than 85% of CD133-positive cells were CD34 positive. The absence of CFU-F in MPB and UCB prior to expansion or selection could have diverse explanations. The MSC frequency in adult BM is one in 3 to 4 × 104 cells, using the CFU-F assay . In similar conditions, UCB and MPB do not produce CFU-F. The frequency of MSCs in UCB and MPB is thus so low that their survival or growth can be affected by culture conditions (serum, pH of medium). The MNC fraction obtained from UCB and MPB yields an adherent layer of heterogeneous cells in the primary culture. We cannot exclude the possibility that accessory cells such as macrophages, lymphocytes, or other hematopoietic cells could inhibit the proliferation of CFU-F. To determine whether MSCs isolated from UCB and MPB, like BM-MSCs, are able to differentiate into multiple cell types, cells were plated into specific induction media for the generation of adipocytes, osteocytes, chondrocytes, and neuronal/glial cells. After 2–3 weeks of culture under these conditions, lipid vacuoles, calcium deposits, chondrogenic matrix, and neuronal/glial cells were observed, which demonstrated the multipotentiality of MSCs. Finally, we observed that MSCs, isolated from the MPB or UCB as embryonic stem cells, express Oct4, a transcriptional factor present in undifferentiated cells with high proliferative potential.
MPB and UCB are important sources of hematopoietic stem cells already used in cell therapy after chemotherapy. Due to the large volume and accessibility of these sources, it was important to find a common, easily reproducible method for purifying MSCs. In this study, we demonstrated that MSCs can be easily isolated from MPB or UCB, using a CD133-positive selection. This method of selection makes it possible after four passages to rapidly obtain a large quantity of MSCs, as described for BM. These MPB- and UCB-MSCs have a potential similar to MSCs derived from BM, such as differentiation into different tissues (fat, bone, cartilage, and neural cells) and proliferative capacity (CFU-F assay, cumulative cell number, and Oct4 expression). This interesting finding demonstrates the presence of “circulating” MSCs in MPB and in UCB and the possibility of isolating them with a common method based on the CD133-positive selection. We therefore suggest that UCB and MPB could be considered not only as a source of hematopoietic cells but also as a source of MSCs for cellular therapy.
This work was supported by a grant from the Fonds National de la Recherche Scientifique (FNRS) of Belgium.