Migration of Bone Marrow and Cord Blood Mesenchymal Stem Cells In Vitro Is Regulated by Stromal-Derived Factor-1-CXCR4 and Hepatocyte Growth Factor-c-met Axes and Involves Matrix Metalloproteinases



Human mesenchymal stem cells (MSCs) are increasingly being considered in cell-based therapeutic strategies for regeneration of various organs/tissues. However, the signals required for their homing and recruitment to injured sites are not yet fully understood. Because stromal-derived factor (SDF)-1 and hepatocyte growth factor (HGF) become up-regulated during tissue/organ damage, in this study we examined whether these factors chemoattract ex vivo-expanded MSCs derived from bone marrow (BM) and umbilical cord blood (CB). Specifically, we investigated the expression by MSCs of CXCR4 and c-met, the cognate receptors of SDF-1 and HGF, and their functionality after early and late passages of MSCs. We also determined whether MSCs express matrix metalloproteinases (MMPs), including membrane type 1 (MT1)-MMP, matrix-degrading enzymes that facilitate the trafficking of hematopoietic stem cells. We maintained expanded BM- or CB-derived MSCs for up to 15–18 passages with monitoring of the expression of 1) various tissue markers (cardiac and skeletal muscle, neural, liver, and endothelial cells), 2) functional CXCR4 and c-met, and 3) MMPs. We found that for up to 15–18 passages, both BM- and CB-derived MSCs 1) express mRNA for cardiac, muscle, neural, and liver markers, as well as the vascular endothelial (VE) marker VE-cadherin; 2) express CXCR4 and c-met receptors and are strongly attracted by SDF-1 and HGF gradients; 3) express MMP-2 and MT1-MMP transcripts and proteins; and 4) are chemo-invasive across the reconstituted basement membrane Matrigel. These in vitro results suggest that the SDF-1-CXCR4 and HGF-c-met axes, along with MMPs, may be involved in recruitment of expanded MSCs to damaged tissues.


Mesenchymal stem cells (MSCs) have generated a great deal of excitement as a potential source of cells for cell-based therapeutic strategies [1, 2]. Despite the absence of a single definitive marker that characterizes these cells and the lack of unequivocal evidence regarding their location and distribution in vivo, their existence is deduced from their apparent role in providing an appropriate microenvironment for self-renewal and differentiation of hematopoietic stem/progenitor cells (HSPCs) [24]. The precise identity of MSCs remains an enigma, as perhaps is indicated by the various other terminologies used to describe these adherent nonhematopoietic progenitors displaying fibro-blast-like morphology. For example, they have been referred to as bone marrow (BM) stromal (stem) cells, stromal precursor cells, colony-forming unit-fibroblasts, and multipotent adult progenitor cells [5, 6]. MSCs are called this because they contribute to the regeneration of mesenchymal tissues such as bone, cartilage, muscle, ligament, tendon, adipose tissue, and stroma; however, they are also believed to have the ability, in vitro, to differentiate into nonmesenchymal cell types [7]. Recently MSCs were reported to occur in the peripheral blood (PB) [8], and transplantation studies using BM-derived MSCs have demonstrated that these cells home to tissues injured by genetic defect [9], myeloablative therapy [10], irradiation [11], or myocardial infarction [12, 13]. Recently we and others have shown that stromal-derived factor (SDF)-1 and hepatocyte growth factor (HGF) become upregulated at sites of tissue damage [1418]. In this study we examined whether SDF-1 and HGF mediate migration in vitro of MSCs.

SDF-1 (also known as CXCL12) is an α-chemokine that strongly chemoattracts HSPCs through interaction with its unique receptor CXCR4 [19, 20]. It has become evident that the SDF-1-CXCR4 signaling axis plays an important role in the homing and engraftment of HSPCs in the BM [21, 22]. SDF-1 regulates tethering or adhesion of HSPCs to the endothelium [23, 24], the expression of the basement-membrane degrading enzymes matrix metalloproteinases (MMPs) [25], and other processes that are essential to HSPC homing and engraftment.

HGF was originally isolated as a mitogen for adult hepatocytes and has been shown to be a potent regulator of HSPC proliferation and differentiation [26], as well as a powerful stimulator of angiogenesis [27]. The activity of HGF is mediated primarily through binding to its receptor, a transmembrane tyrosine kinase encoded by the MET proto-oncogene (c-met). We have shown that the HGF-c-met axis, like the SDF-1-CXCR4 axis, plays a role in directing metastasis of rhabdomyosarcoma cells into the BM [16], and recently it was demonstrated that HGF exerts a strong chemotactic effect on MSCs in a wound-healing model [28].

In this work, we examined the expression of the CXCR4 and c-met receptors in MSCs from cultures established from BM and umbilical cord blood (CB) and determined whether they remain functional during MSC expansion for up to 18 passages. We also examined whether MSCs express soluble MMPs (MMP-2, MMP-9) and membrane type 1 (MT1)-MMP. This latter enzyme plays a critical role in proMMP-2 activation. We also monitored for up to 18 passages the expression of markers for cardiac and skeletal muscle, neural, liver, and endothelial cells by MSCs to determine whether their potential to contribute to tissue regeneration persists throughout the expansion period. We show that the SDF-1-CXCR4 and HGF-c-met axes regulate trafficking of BM- or CB-derived MSCs in vitro, suggesting their involvement in recruitment of MSCs to damaged tissues.

Materials and Methods

Cells and Cultures

CB cells were obtained with the mothers' informed consent after delivery, and BM cells were obtained from unrelated donors with their informed consent, both in accordance with the guidelines approved by the University of Alberta Ethics Committee. In all cases samples were processed within 24 hours of collection. Light-density mononuclear cells (MNCs) were separated by centrifugation using Percoll density gradient (1.077 g/ml; Amersham Biosciences, Uppsala, Sweden, http://www.amersham.com). The MNCs were washed, suspended in Iscove's modified Dulbecco's medium (IMDM; Invitrogen, Grand Island, NY, http://www.invitrogen.com) supplemented with 20% fetal bovine serum (FBS; Invitrogen) and seeded at a concentration of 0.4 × 106 cells per cm2 for BM MNCs and 1 × 106 to 5 × 106 cells per cm2 for CB MNCs. Cultures were maintained at 37°C in a humidified atmosphere containing 5% CO2; after 72 hours for BM MNCs and 24 hours for CB MNCs, nonadherent cells were removed, and the complete medium was replaced. Once confluence was achieved, the cells were passaged and maintained in IMDM supplemented with 10% FBS; medium changes were performed weekly thereafter. Passaging was carried out by trypsinizing the cells (0.05% trypsin-ethylenediaminetetraacetic acid [EDTA]; Invitrogen), washing, and subculturing at a density of 2 × 104 cells per cm2 without the addition of any other growth factors or exogenous recombinant cytokines.

To confirm the identity of the BM and CB MSCs, the cells were grown in medium that is conducive to differentiation into bone (osteoclasts) and fat (adipocytes). For osteogenic differentiation CB and BM-derived MSCs were grown to 90% confluency in six-well plates and then incubated in osteogenic medium (10−8 M dexamethasone, 0.2 mM ascorbic acid, 10 mM β-glycerophosphate; Stem Cell Technologies, Vancouver, BC, Canada, http://www.stemcell.com). The media were replaced every 3 days, and deposition of bone mineral was observed after 2 weeks. The cultures were washed with phosphate-buffered saline (PBS) and fixed in ice-cold 70% ethanol for 1 hour, rinsed with water, and stained for 10 minutes with 1 ml of 40 mM Alizarin red (pH 4.1; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), before being rinsed three times with PBS and photographed. For adipogenic differentiation, MSCs in 90% confluent cultures were grown in six-well plates in basal medium for human MSCs with adipogenic stimulatory supplements (Stem Cell Technologies), which was replaced every 3 days. Cells containing lipid droplets were observed after 2 weeks. Cells were then washed with PBS, fixed in 10% formalin for 10 minutes, and then stained for 10 minutes with Oil Red O solution (Sigma-Aldrich) before being photographed. All images were taken under ×20 magnification using a Camedia C-3040 digital camera attached to an Olympus CK 40 inverted microscope (Melville, NY).

Fluorescence-Activated Cell Sorting Analysis

The expression of CD45, STRO-1, c-met, and CXCR4 in BM-and CB-derived MSCs was evaluated by fluorescence-activated cell sorting (FACS) analysis as previously described [16]. The CD45 antigen was detected using monoclonal anti-CD45-phycoerythrin (PE) conjugate, clone BRA-55 (Sigma-Aldrich), CXCR4 antigen with PE-anti-CXCR4 monoclonal antibody clone 12.5 (R&D Systems Inc., Minneapolis, http://www.rndsystems.com), c-met antigen with anti-c-met monoclonal antibody clone DO-24 (UPS Biotechnology, Lake Placid, NY), and STRO-1 with fluorescein isothiocyanate (FITC)-anti-STRO-1 monoclonal antibody (Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, IA, http://www.uiowa.edu/∼dshbwww). Briefly, the cells were stained in PBS (Ca-and Mg-free) supplemented with 5% FBS (Hyclone, Logan, UT, http://www.hyclone.com). After the final wash, cells were fixed in 1% paraformaldehyde prior to FACS analysis, which was performed by FACscan (Becton, Dickinson and Company, San Jose, CA, http://www.bd.com) using FITC- or PE-goat-anti-mouse immunoglobulin (IgG) as the isotype control. To eliminate any nonspecific binding, the same ratio of fluorochrome/protein for the isotype control and specific antibody was used.

Gel-Based Reverse Transcription-Polymerase Chain Reaction

MSCs from 1 up to 18 passages were harvested at near confluence, and total RNA was extracted using Trizol reagent and following the manufacturer's instructions (Invitrogen). Reverse transcription was carried out using Moloney murine leukemia virus reverse transcriptase, and the resulting cDNA fragments were amplified using Taq polymerase (both from Invitrogen). Primer sequences were obtained from GenBank (Los Alamos, NM) and are listed in Table 1. Thermocycling was performed with an Eppendorf Mastercycler personal thermocycler (Westbury, NY), and the PCR products were electrophoresed on a 2% agarose (Invitrogen) gel containing ethidium bromide (Sigma-Aldrich). Gels were visualized under UV light and photographed using a Kodak DC120 digital camera (Eastman Kodak, Rochester, NY).

Real-time Reverse Transcription-Polymerase Chain Reaction

Real-time reverse transcription-polymerase chain reaction (RT-PCR) was performed as previously described [17]. For analysis of Myf5, MyoD, myogenin, glial fibrillary acidic protein (GFAP), nestin, and α-fetoprotein mRNA levels, total mRNA was isolated from cells with the RNeasy Mini Kit (Qiagen, Valencia, CA, http://www1.qiagen.com). mRNA was reverse-transcribed with TaqMan reverse transcription reagents (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Quantitative assessment of Myf5, MyoD, myogenin, GFAP, nestin, α-fetoprotein, and β2 microglobulin mRNA levels was performed by real-time RT-PCR using an ABI PRISM 7000 Sequence Detection System (Applied Biosystems). Primers were designed with Primer Express software (Applied Biosystems), and sequences were as described in detail previously [17]. A 25-μl reaction mixture containing 12.5 μl of SYBR Green PCR master mix and 10 ng of cDNA template and forward and reverse primers was used. The threshold cycle (Ct) (i.e., the cycle number at which the amount of amplified gene of interest reached a fixed threshold) was subsequently determined. Relative quantitation of Myf5, MyoD, myogenin, GFAP, nestin, and α-fetoprotein mRNA expression was calculated using the comparative Ct method. The relative quantitation value of target, normalized to an endogenous control β2 microglobulin gene, and relative to a calibrator, is expressed as 2−ΔΔCt (fold difference), where ΔCt = (Ct of target genes [Myf5, MyoD, myogenin, GFAP, nestin, and α-fetoprotein]) − (Ct of endogenous control gene [β2 microglobulin]), and ΔΔCt = (ΔCt of samples for target gene) − (ΔCt of calibrator for the target gene). To avoid the possibility of amplifying contaminating DNA, 1) all of the primers for real-time RT-PCR were designed with an intron sequence inside the cDNA to be amplified; 2) reactions were performed with appropriate negative controls (template-free controls); 3) a uniform amplification of the products was rechecked by analyzing the melting curves of the amplified products (dissociation graphs); 4) the melting temperature (Tm) was 57°–60°C (the probe Tm was at least 10°C higher than the primer Tm); and 5) gel electrophoresis was performed to confirm the correct size of the amplification and the absence of nonspecific bands.

Zymography and Western Blotting

MMP-2 and MMP-9 protein activities were evaluated by zymography as previously described [29] and MTI-MMP was examined using Western blot. Briefly, MSCs (1 × 106 cells per ml) were preincubated in serum-free medium for 24 hours (at 37°C, 5% CO2), and the cell-conditioned media and cell lysates were collected and analyzed. For MT1-MMP immunoblotting cell pellets were sonicated in lysis buffer (1% Triton, 10 mM Tris, 150 mM NaCl, 1 mM EDTA, and 1 mM EGTA) containing protease inhibitor cocktail (10 μg/ml leupeptin, 10 μg/ml aprotinin, 1 μg/ml pepstatin A, 1 mM phenylmethylsulfonyl fluoride, and 2 mM Na3VO4). Cell lysates were clarified by centrifugation at 14,000 rpm for 10 minutes at 4°C, and the protein concentrations were determined using the Bradford protein assay (Bio-Rad, Hercules, CA, http://www.bio-rad.com). Samples were resolved by a 10% polyacrylamide gel under reducing conditions and transferred onto a polyvinylidene fluoride membrane. Following blockage overnight at 4°C with 5% fat-free dried milk in Tris-buffered saline and 0.05% Tween 20, the membrane was probed with a specific monoclonal antibody directed against the catalytic region of human MT1-MMP (mouse anti-human MT1-MMP, clone 114–6G6; Chemicon, Temecula, CA, http://www.chemicon.com) for 2 hours at room temperature. The membrane was further probed with a secondary antibody (Immunopure goat anti-mouse, peroxidase-conjugated IgG; Pierce, Rockford, IL, http://www.piercenet.com) to visualize the bands. Chemiluminescence detection was performed using the Supersignal West Pico system (Pierce) and the Fluor-S MAX2 Multiimager and the Quantity One version 4.3.1 software (Bio-Rad).

Trans-Matrigel Chemoinvasion Assay

We used a chemoinvasion assay [25] to evaluate the ability of MSCs to cross the reconstituted basement membrane Matrigel. Polycarbonate filters (13-mm diameter, 8-μm pore size; Nucleopore, Toronto, http://www.sterlitech.com) were coated with 25 μg of Matrigel (Collaborative Biomedical Products, Bedford, MA, http://www.bioscience.org). The lower compartments of modified Boyden chambers (Neuro Probe Inc., Gaithersburg, MD, http://www.neuroprobe.com) contained IMDM supplemented with 0.5% bovine serum albumin (BSA; Sigma-Aldrich) only (control) or with HGF (40 μg/ml; R&D Systems) or SDF-1 (100 ng/ml; Biomedical Research Centre, University of British Columbia, Vancouver, BC, Canada, http://www.brc.ubc.ca). Cells that had been preincubated in IMDM-0.5% BSA for 30 minutes were loaded onto the upper compartments (0.2 × 106 cells per chamber) and incubated at 37°C, 5% CO2 for 24 hours. Cells that invaded the Matrigel were evaluated on the undersides of filters after being fixed with 11% glutaraldehyde and stained with 1% crystal violet (both from Sigma-Aldrich). Three random fields were selected for microscopic count at ×20 magnification (Leitz Diavert, Ottawa). Each experiment was performed using at least five chambers for each condition, and counts were performed in duplicate. The migration index was calculated as the ratio of the number of cells invading the Matrigel toward the HGF or SDF-1 gradient to the number of cells migrating toward media alone. To assess the individual or combined effects of SDF-1 and HGF in cell migration, some experiments were carried out using increasing doses of HGF (0–40 ng/ml) in the presence or absence of a constant suboptimal dose (50 ng/ml) of SDF-1. Furthermore, we used 1 μM K-252a (Calbiochem-Novabiochem, San Diego, http://www.emdbiosciences.com) and 100 μg/ml AMD3100 (Sigma), which are specific blocking agents for c-met and CXCR4, respectively, to determine whether these molecules inhibit the chemotactic responses of MSCs toward HGF and SDF-1. To assess the role of MT1-MMP in the chemoinvasion process, MSCs were incubated with 50 μg/ml (−) epigallocatechin gallate (EGCG, Sigma). EGCG has been shown to directly inhibit MT1-MMP and prevent activation of MMP-2 [30]. The experiments were carried out as before, with at least three chambers being used for each condition.

Statistical Analysis

Arithmetic means and standard deviations were calculated for our data, and statistical significance was defined as p < .05 using Student's t test.


Characteristics of MSCs Derived from BM and CB

MSC cultures were initiated using MNCs isolated from normal BM samples and maintained for up to 18 passages. It took between 14 and 21 days to obtain a homogeneous adherent monolayer and to establish primary culture of MSCs derived from BM. In contrast, the MSC cultures established from CB showed more variable degrees of confluency, and only 19 of 70 CB samples yielded cultures that survived past the 10th passage. We found no correlation between gestational age and the successful establishment of MSC cultures from CB; however, a cell density of less than 1 × 107 MNCs per ml failed to generate MSC-like cells, indicating that the concentration of MNCs may be an important parameter for MSC expansion. Establishment of primary CB MSC cultures took approximately 30 days. However, in contrast to BM MSCs, CB MSCs showed more than a twofold increase in proliferation at each subcultivation from the start of passaging up to passage 14, while BM-derived MSCs displayed markedly reduced proliferation, especially after the 10th passage (Fig. 1A). BM- and CB-derived MSCs were CD45-negative and STRO-1-positive (65% in BM MSCs and 25% in CB MSCs) (Fig. 1B). These CB-and BM-derived MSCs were capable of differentiating into bone and fat after a 2-week culture in appropriate media (Fig. 1C).

Moreover, using gel-based RT-PCR, we detected expression of early markers specific for cardiac (Nkx2.5, GATA-4), skeletal muscle (myogenin), and endothelial (VE-CAD) cells in both CB- and BM-derived MSC cultures, which persisted during 15–17 passages (Fig. 2A). Interestingly, CB-derived MSCs consistently expressed VEGFR-2, whereas the BM-derived MSCs did not. Furthermore, using real-time PCR, we found expression of early skeletal muscle (Myf5 and MyoD), neural (GFAP, nestin), and liver (α-fetoprotein) markers both in BM- and CB-derived MSCs, and that relative to the control (passage 1), all markers except one were upregulated for up to passage 15 (Fig. 2B).

BM- and CB-Derived MSCs Express Functional CXCR4 and c-met Receptors

We found that transcripts for CXCR4 and c-met (receptors involved in cell migration toward SDF-1 and HGF gradients, respectively), are strongly expressed by BM- and CB-derived MSCs regardless of number of passages (Fig. 3A). FACS analysis showed that these MSCs stained positively for the CXCR4 and c-met antigens, albeit weakly (Fig. 3B). In fact, we found that surface expression of CXCR4 and c-met diminished with passage. CXCR4 expression in CB-derived MSCs was 21.6% at passage 5 and 11.5% at passage 14; in BM-derived MSCs, CXCR4 expression went down from 21.5% at passage 5 to 14.7% at passage 14. A decrease in c-met expression with passage was also observed in CB-derived MSCs, from 18.5% at passage 5 to 5% at passage 14, and in BM-derived MSCs from 30.4% at passage 5 to 3.6% at passage 14. To check whether these receptors are functional, we employed a migration assay in which the membrane filters were covered with Matrigel to provide an adherent surface for the MSCs (prior to their penetration and migration toward chemotactic gradients). At passage 10, CB-derived MSCs exhibited a higher increase in chemotaxis toward an HGF gradient (4.1-fold) than BM-derived MSCs (2.4-fold) (Fig. 4A). These strong chemotactic responses toward both HGF and SDF-1 (relative to medium alone) were not diminished for up to 15 passages. However, the absolute number of cells migrating toward SDF-1 or HGF decreased from passages 5 to 12 (Fig. 4B), which is consistent with the observed reduction of surface expression of CXCR4 and c-met with passage. We also observed that CB-derived MSCs from passages 5 and 12 responded to HGF in a dose-dependent manner and showed increased chemotactic response in the presence of a suboptimal dose of SDF-1 (50 ng/ml) (Fig. 4B). This chemotaxis was significantly inhibited by the specific blocking agents for c-met (K-252a) and CXCR4 (AMD3100) used singly or in combination (Fig. 4B).

Early- and Late-Passage MSCs Express MMP-2 and MT1-MMP

This observation that MSCs were able to migrate across a Matrigel barrier suggested that they produce matrix-degrading enzymes. We have previously shown that MMPs facilitate migration of HSPCs [25, 29, 31]. Here we show that both BM-derived and CB-derived MSCs express MMP-2 and MT1-MMP transcripts throughout passages 1 to 15 (Fig. 5A). MMP-9 mRNA and MMP-9 protein also appeared during the early passages of BM-derived MSCs, which could be explained by contamination of cultures by nonadherent MNCs (Fig. 5B). Importantly, both the latent and active forms of MMP-2 were found in media conditioned by BM- and CB-derived MSCs for up to late passages (Fig. 5B). Furthermore, MT1-MMP protein was detected by Western immunoblotting of lysate proteins from early-, middle-, and late-passage MSCs derived from BM and CB (Fig. 5C). To further evaluate the role of MMPs and especially MT1-MMP in Matrigel chemoinvasion toward gradients of SDF-1 and HGF, CB-derived MSCs were incubated with the potent MT1-MMP inhibitor EGCG. This inhibitor reduced CB MSC trans-Matrigel chemoinvasion by 56%–67%, and the extent of inhibition was similar for both early (passage 4) and late (passage 12) CB MSC passages (Fig. 5D).


Mesenchymal stem cells are nonhematopoietic stem cells that are able to differentiate into various mesoderm-type cell lineages such as osteoblasts, chondrocytes, adipocytes, myocytes, and endothelial cells, and they hold significant promise for cellular therapies. Recently, however, a population of MSCs highly purified from BM, characterized as STRO-1BRIGHT/VCAM-1+, was found to differentiate into osteo-, adipo-, and chondrogenic cell lineages only and not into other cell types [32].

Although BM has been the main source of MSCs for both experimental and clinical studies, recent work has shown that MSCs could also be isolated from umbilical cord vein [33, 34] and CB [3537]. In our studies, we were able to establish MSC cultures from CB which survived past the 10th passage 30% of the time, but we were unable to culture MSCs from mobilized PB (data not shown) as others have done [8]. The BM- and CB-derived MSCs we established were negative for the hematopoietic CD45 marker and positive for the stromal cell marker STRO-1 and were able to differentiate into bone and fat cells.

Early studies of transplantation of human MSCs into animals indicated that these cells engrafted and differentiated into tissue-specific cells including cartilage, fat, and cardiac muscle, as well as BM and thymic stromas [38] and epithelium of liver, lung, and gut [39]. Here we present evidence that both BM- and CB-derived MSCs are not only capable of differentiating into bone and fat but also express mRNA specific for early skeletal muscle, smooth muscle, neural, liver, and endothelial cells and that this expression even increases with subsequent passages. This phenomenon could be explained by the fact that, as suggested by other investigators, MSCs are capable of shifting differentiation courses that may include endodermal and ectodermal fates [40, 41], supporting the notion that MSCs do in fact have true pluripotent potential. On the other hand, we have recently suggested that cultures of MSCs may contain from the beginning a small admixture of tissue-committed stem cells (TCSCs) that are co-isolated with MSCs and could survive/expand during subsequent passages [42]. Also, the possibility of promiscuous gene expression of these markers for TCSCs cannot be excluded. One difference between BM- and CB-derived MSCs that we observed was expression of VEGFR-2, which we found in MSCs from CB but not from BM. Recently, MSCs derived from umbilical cord vein were shown to express endothelial markers [34]. In our study, CB (not vein) MSCs showed both VE-CAD and VEGFR-2 transcripts even though the culture conditions in which we carried out MSC expansion were not conducive to endothelial cell growth. This suggests that cultures of CB-derived MSCs are more enriched in endothelial progenitor cells than cultures of BM MSCs. On the other hand, this variation in expression of VEGFR-2 may be a consequence of the different growth characteristics of BM- and CB-derived MSCs in culture. Accordingly, we observed that MSCs from CB proliferated better than those from BM, and our results are consistent with the notion that confluence of MSCs results in the upregulation of VEGFR-2 expression.

Because tissue damage appears to be an important signaling cue in the migratory responses of MSCs, we next investigated in BM and CB MSCs the expression of CXCR4 and c-met, which are the receptors for SDF-1 and HGF chemokines, respectively, and whose levels are upregulated at injured sites [14, 15, 17, 18]. It has been shown in a rat model that SDF-1-CXCR4 interactions mediate the homing of BM-derived MSCs to impaired sites in the brain [43]. We found that the CXCR4 transcript is strongly expressed by both BM- and CB-derived MSCs regardless of the number of passages. However, the surface expression of the CXCR4 antigen was quite low, confirming the findings of others [44, 45] and suggesting that this protein can be expressed intracellularly rather than on the surface. The majority of CXCR4 (83%–98%) is localized in endosomal compartments and cycles continuously to and from the cell surface via endocytosis involving clathrin-coated pits [46], and we can assume that CXCR4 sequestered intracellularly in MSCs is mobilized to the cell surface, for example, during cytokine stimulation [45]. Nevertheless, we demonstrated the CXCR4 receptor to be functional, as evaluated by the chemotactic responses of MSCs to an SDF-1 gradient, although the chemotactic response diminished with passage consistent with the decrease in CXCR4 surface expression. The CXCR4 antagonist AMD3100 significantly inhibited chemotaxis of MSCs toward SDF-1, further confirming that the SDF-1-CXCR4 axis regulates the migratory responses of MSCs. Recently we reported that, after stimulation of HSPCs with various inflammatory molecules, CXCR4 incorporates into membrane lipid rafts and HSPCs primed in such a manner were better able to sense an SDF-1 gradient and home to BM [47]. Whether MSCs can be primed by inflammatory molecules requires further investigation.

HGF and its high-affinity receptor c-met were previously shown to be upregulated following myocardial ischemia and reperfusion in a rat model [48]. We have also observed that HGF expression is increased in injured heart relative to a control, and we suggested that nonhematopoietic cell populations are mobilized into PB and chemoattracted to infarcted myocardium in an HGF-c-met-dependent manner [18]. Here we present evidence that BM and CB MSCs express the c-met receptor, albeit weakly, which is functional as evaluated through chemotaxis assay. Similar to SDF-1, chemotactic response of MSCs toward HGF diminished with passage, a finding that is consistent with the concurrent decrease in c-met surface expression. Furthermore, the specific c-met blocking agent K-252a significantly inhibited the chemotactic responses of MSCs toward HGF. Hence we suggest that the HGF-c-met axis is also involved in the directed migration of MSCs, and this HGF-induced chemoattraction may direct MSCs into the HGF-rich environment of injured sites (e.g., infarcted myocardium). Moreover, this chemotactic response could be potentiated by SDF-1, also known to be upregulated in these sites, and here we demonstrate the additive chemotactic effects of HGF and SDF-1 on CB MSCs obtained from early and late passages.

Another mechanism contributing to cell migration to various sites involves MMPs. We provide evidence that BM and CB MSCs from early up to late passages secrete latent and active MMP-2 and produce MT1-MMP. MT1-MMP has been shown in our and others' preliminary studies to be involved in mobilization and homing of HSPCs [49, 50], and is also known to influence the migration of monocytes and endothelial cells [51, 52]. The expression of MT1-MMP by MSCs is important, as it has recently been shown to be upregulated under hypoxic conditions and to promote the migration of, and capillary-tube formation by, BM MSCs [53]. Moreover, it has been proposed that MT1-MMP confers upon cancer cells the ability to penetrate connective tissue barriers, proliferate within the confines of the three-dimensional extracellular matrix, and initiate the metastatic process. Apparently, the cancer cells adopt a fibroblast-like phenotype that allows for an amoeboid form of movement [54]. It is believed that only MT1-MMP is directed specifically to sites targeted for degradation by a combination of motifs embedded not only in the transmembrane domain and intracellular cytosolic tail but also in the collagen-binding hemopexin domain of the enzyme. This is not to suggest that secreted proteinases such as MMP-2 and MMP-9 are not contributing to the extravasation process, but MT1-MMP is able to promote cell migration also through its ability to process and activate cell surface molecules such as CD44 [55], integrin αvβ3 [56], and tissue transglutaminase [57]. In our study, chemoinvasion by CB-derived MSCs toward SDF-1 and HGF gradients across a reconstituted basement membrane barrier was strongly inhibited by EGCG, further underlining the role of MT1-MMP in MSC migration, although the involvement of other MMPs, especially MMP-2, is also likely.

In conclusion, our work provides evidence that MSCs could home to injured tissues following signaling cues regulated by gradients of SDF-1 and HGF in an MMP-dependent manner. Expression/retention of various markers for cardiac and skeletal muscle, neural, liver, and endothelial cells by MSC cultures derived from both CB and BM holds promise for deployment of these cells in tissue/organ regeneration.

Table Table 1.. Primers (sense and antisense) used for gene characterization
original image
Figure Figure 1..

Characteristics of MSCs derived from BM and CB. (A): Growth kinetics of mesenchymal stem cells (MSCs) derived from BM and CB as evaluated from cell counts from different passages (P2 to P14). Numbers represent fold increase in cell counts relative to previous passage. Cell enumeration of trypsinized cells was carried out using a Neubauer hemocytometer. *, statistically significant difference (p < .05) between CB MSCs and BM MSCs. (B): Representative flow cytometric analysis of CD45 and STRO-1 expression (blue curves) on BM-derived MSCs (passage 3) and CB-derived MSCs (passage 5); red curves, isotype controls. (C): CB-derived MSCs (passage 7) differentiate into adipocytes (left panel) and bone (right panel). Photographs were taken at ×20 magnification, and similar images were observed with BM-derived MSCs. Abbreviations: BM, bone marrow; CB, cord blood; P, passage.

Figure Figure 2..

Expression of early tissue-specific markers in BM- and CB-derived MSCs. (A): Gel-based reverse transcription-polymerase chain reaction (PCR) analysis of tissue-specific markers for cardiac (Nkx2.5, GATA-4, and MEF2C), skeletal muscle (myogenin), and endothelial (VE-CAD and VEGFR-2) cells in MSCs derived from BM and CB. GAPDH was used as the internal mRNA control. (B): Real-time PCR analysis of expression of mRNA for early muscle, neural, and liver markers in BM- and CB-derived MSCs. Data represent results of two samples each of BM and CB used to establish MSCs and cells obtained from passages 0 (control, P1), 5 (P5), 10 (P10), and 15 (P15) were analyzed. Abbreviations: BM, bone marrow; CB, cord blood; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MSC, mesenchymal stem cell; P, passage.

Figure Figure 3..

Expression of CXCR4 and c-met receptors in BM- and CB-derived MSCs. (A): Reverse transcription-polymerase chain reaction analysis of CXCR4 and c-met receptors in BM-derived and CB-derived MSCs. Negative reactions (N) were carried out without adding the reverse-transcribed cDNA; GAPDH was used as the internal mRNA control. Passage numbers are indicated on top of the gel; data are representative of three experiments each using BM- and CB-derived MSCs. (B): Expression of CXCR4 and c-met on BM-and CB-derived MSCs (both at passage 5) as evaluated by flow cytometry (bold font), iso-type controls (regular font). Abbreviations: BM, bone marrow; CB, cord blood; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MSC, mesenchymal stem cell.

Figure Figure 4..

Chemotaxis of BM- and CB-derived MSCs across Matrigel-covered membrane filters and the effect of CXCR4 and c-met antagonists. (A): The mesenchymal stem cells (MSCs) were collected during their early (P3), middle (P10), and late (P15) passages and allowed to migrate toward gradients of HGF (40 ng/ml) or SDF-1 (100 ng/ml). Cells that invaded Matrigel were counted microscopically using three random fields on the undersides of filters. Data are presented as fold increase in migration relative to control (toward media alone). (B): Additive effect of HGF and SDF-1 on chemotaxis of CB-derived MSCs, passages 5 and 12. Incremental doses of HGF (0–40 ng/ml) were added to a constant suboptimal dose of SDF-1 (50 ng/ml). Chemotaxis toward 100 ng/ml SDF-1 was inhibited when cells were pretreated with A. Likewise, chemotaxis toward 40 ng/ml HGF was inhibited in the presence of K and a combination of both antagonists (A + K). *, p < .01 (indicates statistically significant difference in chemotactic response toward gradient when compared to media alone); **, p < .001 (indicates statistically significant inhibition of chemotactic response by AMD3100 and K-252a). Abbreviations: A, the CXCR4 antagonist AMD3100; BM, bone marrow; CB, cord blood; HGF, hepatocyte growth factor; K, the c-met blocking agent K-252a; P, passage; SDF, stromal-derived factor.

Figure Figure 5..

Expression of MMPs in BM- and CB-derived MSCs and the effect of MTI-MMP inhibitor on chemoinvasion. (A): Reverse transcription-polymerase chain reaction analysis of MMP-9, MT1-MMP, and MMP-2 expression in BM and CB MSCs obtained from various passages. Negative reactions (N) were carried out without adding the reverse-transcribed cDNA; GAPDH was used as the internal mRNA control, and passage numbers are indicated on top of the gels. (B): Protein secretion of MMP-9 and MMP-2 as analyzed by zymography. Medium conditioned by fibrosarcoma HT-1080 cells was used as a standard to indicate the position of the latent and active forms of MMP-9 and MMP-2. Passage numbers are indicated on top of the gels. (C): Western immunoblotting showing MT1-MMP present in lysate proteins from BM-derived MSCs (passages 3, 6, and 15) and CB-derived MSCs (passages 3, 10, and 15). Lysate proteins from HT-1080 cells were used as positive control (Pos) and from T47D cells for negative control (Neg). (D): Inhibition of chemoinvasion of CB-derived MSCs across Matrigel-covered membranes by EGCG. The MSCs were collected from early (P4) and late (P12) passages and allowed to migrate toward gradients of HGF or SDF-1 (in the lower chambers) in the absence (white bars) or presence (black bars) of EGCG. Data are representative of two experiments using three chambers per condition. CB MSCs from passage 12 had higher chemoinvasive potential than those from passage 4 (*, p < .05) and chemoinvasion toward HGF was higher than toward SDF-1 (**, p > .05). Abbreviations: BM, bone marrow; CB, cord blood; EGCG, epigallocatechin gallate; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; MMP, matrix metalloproteinase; MSC, mesenchymal stem cell; MT1, membrane type 1; Neg, negative control; P, passage; Pos, positive control; SDF, stromal-derived factor; Std, standard.


The authors indicate no potential conflicts of interest.


B.-R. S. was on leave of absence from the College of Medicine, Chungbuk National University, South Korea. We thank Jencet Montaño and Neeta Shirvaikar for technical assistance.