SEARCH

SEARCH BY CITATION

Keywords:

  • Bone marrow;
  • Mesenchymal stem cell;
  • In vitro migration;
  • Growth factors;
  • Chemokines;
  • Tumor necrosis factor α

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Adult bone marrow (BM)-derived stem cells, including hematopoietic stem cells (HSCs) and MSCs, represent an important source of cells for the repair of a number of damaged tissues. In contrast to HSCs, the soluble factors able to induce MSC migration have not been extensively studied. In the present work, we compared the in vitro migration capacity of human BM-derived MSCs, preincubated or not with the inflammatory cytokines interleukin 1β (IL1β) and tumor necrosis factor α (TNFα), in response to 16 growth factors (GFs) and chemokines. We show that BM MSCs migrate in response to many chemotactic factors. The GFs platelet-derived growth factor-AB (PDGF-AB) and insulin-like growth factor 1 (IGF-1) are the most potent, whereas the chemokines RANTES, macrophage-derived chemokine (MDC), and stromal-derived factor-1 (SDF-1) have limited effect. Remarkably, preincubation with TNFα leads to increased MSC migration toward chemokines, whereas migration toward most GFs is unchanged. Consistent with these results, BM MSCs express the tyrosine kinase receptors PDGF-receptor (R) α, PDGF-Rβ, and IGF-R, as well as the RANTES and MDC receptors CCR2, CCR3, and CCR4 and the SDF-1 receptor CXCR4. TNFα increases CCR2, CCR3, and CCR4 expression (as opposed to that of CXCR4), together with RANTES membrane binding. These data indicate that the migration capacity of BM MSCs is under the control of a large range of receptor tyrosine kinase GFs and CC and CXC chemokines. Most chemokines are more effective on TNFα-primed cells. Our results suggest that the mobilization of MSCs and their subsequent homing to injured tissues may depend on the systemic and local inflammatory state.

Disclosure of potential conflicts of interest is found at the end of this article.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Adult bone marrow (BM)-derived stem cells represent an important source of cells for repair of a number of damaged tissues. As shown by studies in vitro [1, 2] and in vivo [3, 4], these cells include two distinct multipotential cell populations: hematopoietic stem cells (HSCs), which differentiate into all hematopoietic lineages, and MSCs, which give rise to the different hematopoietic microenvironmental cells, including vascular smooth muscle-like stromal cells, adipocytes, osteoblasts, and, more controversially, endothelial cells. MSCs also generate cells that do not belong to the hematopoietic microenvironment that are mesodermal in origin, such as chondrocytes [3] and cardiac or skeletal muscle cells [5, 6], but may also be neuroectodermal (neurons, astrocytes, oligodendrocytes) [7, 8] or endodermal (hepatocytes) [9, 10].

BM-derived stem cells have been currently used in humans for more than 30 years by transplanting HSCs for hematopoietic rescue after high-dose therapy [11]. By contrast, there are fewer reports on the clinical use of MSCs in patients. Preclinical models with infusion of MSCs in animals have provided evidence for their local engraftment and/or repair capacity in settings such as cerebral injury [12], myocardial ischemia/infarction [6], muscular dystrophy [5], bone disease [13], and bone fractures [14]. In addition, coadministration of human marrow MSCs and HSCs to preimmune fetal sheep has been shown to improve the hematopoietic reconstitution [15], and this finding has been confirmed after transplantation in humans [16]. Another recent application of MSC cotransplantation with allogenic HSCs is the cure of severe graft-versus-host disease [17].

As reported in murine [18], simian [19, [20]21], and human [22] models, following systemic infusion into the bloodstream, BM MSCs are able to colonize and persist long-term in a wide range of tissues, including, in addition to BM, muscle, skin, gut, liver, and lung. Otherwise, MSCs can be induced to circulate into peripheral blood in animal models under particular conditions, such as hypoxia [23]. The chemotactic signals that guide MSCs to appropriate microenvironments or induce their circulation are yet to be clarified, which is not the case for HSCs, for which the predominant role of the chemokine SDF-1 and its receptor CXCR-4 is now well-established [24]. Some soluble factors have recently been reported to exert chemotactic effects on BM MSCs, including chemokines [25, [26]27] and growth factors (GFs) [28, [29]30], but their respective physiological relevance remains unclear. A better knowledge of the chemotactic factors effective on MSCs would be of clinical interest since modulation of their activity could affect not only the engraftment efficiency to damaged sites but also the mobilization into PB. In the present work, we studied the in vitro migration capacity of human BM-derived MSCs in response to 16 chemokines and GFs. We show that, in contrast to what is known for HSCs, a wide range of soluble factors exert significant chemotactic activity on MSCs and that some GFs are better chemoattractants than chemokines. Remarkably, inflammatory cytokines, such as tumor necrosis factor α (TNFα), are able to increase the sensitivity of MSCs to chemokines.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Cell Cultures

BM MSCs were obtained from informed and consenting patients undergoing orthopedic surgery (Orthopedic Surgery Department, Trousseau Hospital, Tours, France), following a procedure approved by the local ethical committee. Cells harvested from the iliac crest were centrifuged, and nucleated cells were counted and seeded at density of 5 × 104 cells per cm2 in α-minimum essential medium (Invitrogen, Cergy-Pontoise, France, http://www.invitrogen.com) supplemented with 10% (vol/vol) screened fetal calf serum (FCS), 20 μmol/l l-glutamine (Invitrogen), and 100 units/ml penicillin G (Invitrogen). Cells were incubated at 37°C in 95% humidified air and 5% CO2. On day 2, all nonadherent cells were removed by changing the medium; thereafter, medium was changed twice a week. When fully confluent, the layer of adherent cells was trypsinized (0.25% trypsin-EDTA, Invitrogen), and the cells were resuspended in culture medium and seeded at 1 × 103 cells per cm2 (passage 1 [P1]). BM MSCs were used at P2 in all experiments (migration assays, flow cytometry analyses, and real time (RQ)-polymerase chain reaction [PCR]).

The three-lineage mesenchymal differentiation of the BM MSCs used was systematically checked by culturing cells in adipogenic, chondrogenic, and osteogenic induction media as previously described [31], and the lack of remaining hematopoietic CD45+ cells was verified by flow cytometry analysis.

Cytokine Prestimulation of BM MSCs

Confluent P2 BM MSCs were preincubated for 1 day with fresh culture medium alone or supplemented with 1 ng/ml TNFα (R&D Systems, Lille, France, http://www.rndsystems.com) or interleukin 1β (IL1β) (R&D Systems) and then trypsinized for flow cytometry analyses and RQ-PCR. For migration assays, confluent P2 BM MSCs were preincubated with migration medium consisting of RPMI medium with 0.25% bovine serum albumin (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) (instead of culture medium) supplemented with 1 ng/ml TNFα or IL1β.

Migration Assays

Migration assays were performed in transwell dishes (Corning Costar, Cambridge, MA, http://www.corning.com/lifesciences) 6.5 mm in diameter with 8-μm pore filters. The upper side of the transwell filter was coated for 1 hour at 37°C with 0.1% (wt/vol) bovine gelatin (Sigma-Aldrich) in phosphate-buffered saline (PBS). Unstimulated or stimulated P2 BM MSCs (5 × 105 cells) were added to the upper chamber, and 600 μl of migration medium with chemotactic factors was added to the bottom chamber. Migration observed in the presence of 30% FCS and with medium alone served as positive and negative controls, respectively. After overnight incubation of the transwells at 37°C, 5% CO2, the upper side of the filters was carefully washed with cold PBS, and cells remaining on the upper face of the filters were removed with a cotton wool swab. Transwell filters were stained using May-Grünwald-Giemsa, cut out with a scalpel, and mounted onto glass slides, putting the lower face on the top. The total number of cells that had migrated were counted using light microscopy at ×200 magnification. Each experiment was performed in duplicate. Data were expressed as numbers of total migrated cells per insert or as percentages of cells related to that of the negative control. We evaluated the chemotactic activity of CC chemokines (100 ng/ml MCP-1, 150 ng/ml RANTES, 50 ng/ml MIP-1α, 100 ng/ml Eotaxin-1, 100 ng/ml Eotaxin-2, 100 ng/ml macrophage-derived chemokine [MDC]), CXC chemokines (50 ng/ml GRO-α, 150 ng/ml SDF-1), CX3C chemokine (300 ng/ml Fractalkine), and growth factors (10 ng/ml fibroblast growth factor-2 [FGF-2], 10 ng/ml platelet-derived growth factor-AB [PDGF-AB], 50 ng/ml hepatocyte growth factor [HGF], 10 ng/ml epidermal growth factor [EGF], 10 ng/ml vascular endothelial growth factor [VEGF], 30 ng/ml insulin-like growth factor-1 [IGF-1], 10 ng/ml Angiopoietin-1 [Ang-1]). All cytokines used were recombinant factors of human origin (rh), purchased from PeproTech (Rocky Hill, NJ, http://www.peprotech.com) or from R&D Systems. Concentrations of GFs and chemokines used were chosen that showed optimal promigratory activity according to the literature.

Flow Cytometry Analysis

Unstimulated and TNFα-stimulated BM MCSs were analyzed for membrane receptor expression using a three-step labeling procedure. Cells were incubated at 4°C for 45 minutes with the following primary mouse anti-human monoclonal antibodies (McAbs) to growth factor and chemokine receptors: anti-PDGF receptor (R) α (clone PRa292), anti-PDGF-Rβ (clone PR7212), anti-EGF-R (clone 102618), anti-FGF-R2 (clone 98725), anti-HGF-R (clone D0-24), anti-TIE-2 (clone 83715), anti-CCR1 (clone 53504), anti-CCR2 (clone 48607), anti-CCR3 (clone 61828), anti-CCR4 (clone 205410), anti-CCR5 (clone 45502.111), anti-CXCR1 (clone 5A12), anti-CXCR2 (clone 6C6), anti-CXCR4 (clone 44717), anti-CXCR5 (clone 51505), and anti-CX3CR1 (clone 2A9-1). All McAbs were purchased from R&D Systems, except for anti-HGF-R, from Upstate (Billerica, MA, http://www.upstate.com), anti-CXCR1 and anti-CXCR2, from Becton, Dickinson and Company (Le-Pont-de-Claix, France, http://www.bd.com), and anti-CX3CR1, from MBL (Nagoya, Japan, http://www.mblintl.com). Cells were subsequently incubated at 4°C with biotinylated goat anti-mouse Ig antibody (Ab) (Becton Dickinson) for 1 hour and with R-phycoerythrin-conjugated streptavidin (Dako, Trappes, France, http://www.dako.com) for 45 minutes. As a negative control, cells were incubated with irrelevant mouse IgG1, IgG2a, or IgG2b McAbs (Becton Dickinson). A minimum of 20,000 events was recorded for each analysis, using a FACSCalibur (Becton Dickinson) flow cytometer. To exclude dead cells, cells were stained before analysis with 7-aminoactinomycin D (7-AAD) in PBS. Appropriate gating was performed using the forward scatter versus side scatter dot plot and gating out 7-AAD-positive dead cells. Expression of membrane receptors was studied on a FL2-H logarithmic scale. For each antibody, the histogram was compared with that of the irrelevant negative control.

Binding Assays with Labeled Cytokines

The binding of RANTES and stromal-derived factor (SDF)-1α to unstimulated and TNFα-stimulated BM MSCs was measured with the Fluorokine kit (R&D Systems). Cells were incubated with a biotinylated rhRANTES or rhSDF-1α at 4°C for 60 minutes. Stained cells were pretreated with purified mouse Ig (10 μg per 106 cells) for 15 minutes at room temperature to block Fc-mediated interactions. Cells were then incubated (without washing) with avidin-fluorescein isothiocyanate reagent for 30 minutes. All samples were labeled with 7-AAD before analysis with a FACSCalibur (Becton Dickinson) flow cytometer. In parallel experiments, for specificity testing, polyclonal goat IgG anti-human RANTES or SDF-1α Ab was mixed with biotinylated rhRANTES or rhSDF-1 and incubated for 15 minutes at room temperature. As negative binding control, an identical sample of cells was stained with a biotinylated soybean trypsin inhibitor, a membrane-binding molecule irrelevant for animal cells. The KG1a cell line, obtained from a patient with acute myeloblastic leukemia, was used as positive control.

Western Blotting

Unstimulated and TNFα-stimulated BM MCS confluent layers were detached in PBS-EDTA (1 mM) and then lysed with lysing buffer (Sigma-Aldrich). Protein concentration was measured using the Bradford protein assay reagent. Proteins were denatured by boiling for 3 minutes in the presence of β-mercaptoethanol. Proteins were then separated in 12% (wt/vol) SDS-polyacrylamide gel. Fifteen micrograms of proteins was loaded per lane. After separation, proteins were electrophoretically transferred onto polyvinylidene difluoride membrane. Nonspecific binding was blocked by incubation with 5% (wt/vol) dried milk in TTBS buffer (10 mM Tris/HCl [pH 8.3], 0.05% Tween-20, and 150 mM NaCl) for 60 minutes at room temperature. Membranes were then incubated with the following primary antibodies: rabbit anti-human polyclonal Abs directed against CCR1 and CCR2 and mouse anti-human McAb directed against CCR4. All Abs were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com) or from R&D Systems. Secondary antibodies were mouse or rabbit anti-mouse IgG conjugated to horseradish peroxidase (Bio-Rad, Hercules, CA, http://www.bio-rad.com). Revelation by chemiluminescence was performed with the ECLplus Western blotting detection kit (Amersham Biosciences, Saclay-Orsay, France, http://www.amersham.com), and acquisition was performed with Chemi-Smart 2000 using Chemocapt software (Vilber Lourmat, Marne-la-Vallée, France, http://www.vilber.com). A lysate of total white blood cells was used as positive control.

TaqMan Real-Time Reverse Transcriptase-Polymerase Chain Reaction

P2 confluent layers of unstimulated and stimulated BM MSCs were lysed directly in the culture dishes, and total RNA was isolated using the RNeasy kit (Qiagen, Chatsworth, CA, http://www1.qiagen.com). TaqMan low density arrays (TLDAs) (Micro Fluidic Cards; Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com) were used in a two-step reverse transcription (RT)-PCR process. First-strand cDNA was synthesized from 3 μg of total RNA using the High-Capacity cDNA Archive Kit (Applied Biosystems). PCRs were then carried out in Micro Fluidic Cards using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). We studied 44 genes related to cell migration, including 22 chemokine, 12 cytokine, and 10 protease and inhibitor genes. The probes were labeled with the fluorescent reporter dye 6-carboxyfluorescein (FAM; Applera Corp., Norwalk, CT, http://www.applera.com) on the 5′-end and with nonfluorescent quencher on the 3′-end. cDNA (100 ng) combined with 1× TaqMan Universal Master Mix (Applied Biosystems) was loaded into each well. Micro Fluidic Cards were thermal cycled at 50°C for 2 minutes and 94.5°C for 10 minutes, followed by 40 cycles at 97°C for 30 seconds and 59.7°C for 1 minute. Data were collected with instrument spectral compensations by the Applied Biosystems SDS 2.2.1 software and analyzed using the threshold cycle (CT) relative quantification method. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) reference gene was used for normalizing the data. The 2−Δ·CT that corresponds to the ratio of each gene expression to GAPDH expression was calculated. The data presented are mean ± SEM from three independent experiments.

Statistics

Statistical differences between two groups were evaluated by the nonparametric Mann-Whitney U test for independent values and by the Wilcoxon t test for paired values (using StatView software; SAS Institute Inc., Cary, NC, http://www.statview.com). Differences were considered significant when the p value was <.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

Before performing specific studies, we verified that culture-amplified cells were negative for CD34, CD45, and CD14 surface markers and positive for CD73, CD90, and CD106 and gave rise to adipocytes, osteoblasts, and chondrocytes when placed in adequate differentiating conditions to show their multipotentiality (data not shown).

Growth Factors Are More Potent Chemotactic Agents Than Chemokines for Unstimulated BM MSCs

To investigate the capacity of GFs and chemokines to induce migration of BM MSCs, we tested seven GFs and nine chemokines (CC, CXC, and CX3C chemokines) using an in vitro migration model. As shown in Figure 1A, GFs were better chemoattractants than chemokines. The highest and significant (p < .01) chemotactic activity was observed with PDGF-AB and IGF-1; a lesser but significant (p < .01) activity was observed with EGF and HGF. Values obtained with PDGF-AB were even superior to values with medium supplemented with 30% FCS (positive control). Among chemokines, only RANTES, MDC, and SDF-1 had significant (p < .05) chemotactic activity. However, chemokine values did not reach those obtained with PDGF-AB or IGF-1.

thumbnail image

Figure Figure 1.. Migration capacity of Unst and TNFα-primed bone marrow (BM) MSCs toward growth factors (GFs) and chemokines. In vitro migration of BM MSC took place through gelatin-coated transwells. After overnight incubation, migrated cells that remained on the lower face of the filters were stained and counted as indicated in Materials and Methods. The following concentrations of soluble factors were used: 10 ng/ml VEGF, 10 ng/ml EGF, 50 ng/ml HGF, 30 ng/ml IGF-1, 10 ng/ml PDGF-AB, 10 ng/ml FGF-2, 10 ng/ml ANGPT-1, 100 ng/ml MCP-1, 50 ng/ml MIP-1α, 150 ng/ml RANTES, 100 ng/ml Eotaxin-1, 100 ng/ml Eotaxin-2, 100 ng/ml MDC, 300 ng/ml Fractalkine, 50 ng/ml GROα, 150 ng/ml SDF-1. (A): Migration of Unst BM MSCs. Different recombinant factors were compared with migration medium alone (neg control). The same medium supplemented with 30% FCS served as positive control. Columns represent mean percentage of migrated cells related to negative control, and bars represent SEM from seven independent experiments. ∗, p < .05; ∗∗, p < .01. (B): Migration of Unst and prestimulated BM MSCs. BM MSCs were prestimulated for 24 hours with 1 ng/ml TNFα or IL1β and allowed to migrate toward medium alone supplemented or not with 30% FCS. Columns represent mean total Nb (±SEM) of migrated cells per filter from three independent experiments. #, Migration values of prestimulated MSCs were higher than values of Unst MSCs in the three experiments. (C): Migration of TNFα-stimulated BM MSCs toward purified chemotactic factors. BM MSCs were prestimulated for 24 hours with 1 ng/ml TNFα or not and were allowed to migrate toward medium alone or medium supplemented with recombinant factors. Columns represent mean total Nb (±SEM) of migrated cells per filter from three independent experiments. #, Migration values of TNFα-stimulated MSCs were higher than values of Unst MSCs in the three experiments. (D): May-Grünwald-Giemsa-stained membrane. Representative stained filters of Unst MSCs toward medium alone and of Unst and TNFα-stimulated MSCs toward RANTES. Abbreviations: ANGPT-1, angiopoietin-1; EGF, epidermal growth factor; FCS, fetal calf serum; FGF, fibroblast growth factor; HGF, hepatocyte growth factor; IGF, insulin-like growth factor; IL, interleukin; MDC, macrophage-derived chemokine; Nb, number; Neg Ctrl, negative control; PDGF, platelet-derived growth factor; SDF, stromal-derived factor; TNFα, tumor necrosis factor α; Unst, unstimulated; VEGF, vascular endothelial growth factor.

Download figure to PowerPoint

TNFα-Primed BM MSCs Show Enhanced Migration Capacity

The “spontaneous” migration capacity of BM MSCs in the presence of medium alone (without FCS, GFs, or chemokines) was low (226 ± 100 cells per insert); it increased in presence of 30% FCS (1336 ± 121 cells per insert) (Fig. 1B). We hypothesized that inflammatory mediators could influence MSC migration. We therefore pretreated BM MSCs with TNFα (1 ng/ml) or IL1β (1 ng/ml) for 24 hours before the migration assay. Only TNFα was found to clearly modify MSC migration by increasing both spontaneous (by 71%) and FCS-induced (by 170%) MSC migration. These results prompted us to evaluate the effect of TNFα-stimulation on MSC migration induced by purified cytokines including GFs and chemokines.

TNFα-primed BM MSCs showed increased migration capacity toward the chemokines RANTES, SDF-1, and MDC (Fig. 1C) by 415%, 267%, and 181%, respectively. Interestingly, the absolute number of migrated MSCs with RANTES after TNFα-stimulation was comparable to that of unstimulated MSCs with IGF-1. On the contrary, in the presence of GFs, the migration capacity of TNFα-primed BM MSCs was not significantly different from that of unstimulated cells, except for Ang-1.

BM MSCs Express a Number of Growth Factor and Chemokine Receptors that Are Modulated by TNFα Stimulation

According to the chemotactic response of BM MSCs to GFs and chemokines, we examined by flow cytometry (Fig. 2A) and Western blotting (Fig. 2B) the membrane expression of relevant receptors. We found high-level expression of PDGF-Rα, PDGF-Rβ, FGF-R2, EGF-R, TIE-2, and Ang-1 receptor and low/intermediate levels of HGF-R (c-Met) and IGF1-R. TNF-R1 and TNF-R2 were found expressed at a low level (data not shown). Unstimulated BM MSCs expressed the following chemokine receptors: among CC receptors, CCR3 at a high level and CCR2, CCR4, and CCR5 at a low/intermediate level; among CXC receptors, only CXCR4 and CXCR5 at a high level, whereas CXCR1 and CXCR2 were not detected. TNFα-stimulation of BM MSCs enhanced membrane expression only for FGF-R2 and CCR3, leading to an increase in signal/noise ratio by 214% and 231%, respectively.

thumbnail image

Figure Figure 2.. Protein expression of growth factor and chemokine receptors on Unst and tumor necrosis factor α (TNFα)-primed BM MSC. (A): Flow cytometry. Passage 2 (P2) BM MSCs were detached using trypsin and incubated with specific or isotypic negative control antibody (Ab). Live 7-aminoactinomycin D cells were analyzed for fluorescence. The results presented are from one representative experiment of three. Black and red curves represent the distribution of fluorescence intensity with specific Ab on Unst and TNFα-Stm MSCs, respectively, and gray areas represent that with negative control Ab. Values indicated in each panel are the ratio of mean fluorescence intensity of Unst BM MSCs with specific Abs and negative control Ab. (B): Western blotting. Proteins from Unst and TNFα-Stm P2 BM MSCs were separated by SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membrane. The blot was incubated with Abs against CCR1, CCR2, or CCR4. WBC lysates were used as positive control. (C): RANTES and SDF-1α binding assays. Cells were preincubated with biotinylated RANTES and SDF-1 or with irrelevant vegetal cell molecule as negative binding control. Membrane binding of cytokines was revealed by incubation with avidin-fluorescein isothiocyanate (as described in Materials and Methods). KG1a cells were used as positive control cells. a, Histogram plots show the level of RANTES and SDF-1 bound to KG1a cells (black curve) as compared with negative control cytokine (gray area). The specificity of the binding assays was measured after preincubation with blocking Abs against RANTES and SDF-1 (blue curve). b, Histogram plots show the level of RANTES and SDF-1 bound to Unst (black curve) and TNFα-Stm MSCs (red curve) as compared with negative control cytokine (gray area). Abbreviations: BM, bone marrow; EGF-R, epidermal growth factor receptor; FGF-R, fibroblast growth factor receptor; HGF-R, hepatocyte growth factor receptor; IGF1-R, insulin-like growth factor 1 receptor; PDGF-R, platelet-derived growth factor receptor; S/N, mean signal/noise ratios from the three experiments; Stm, stimulated; Unst, unstimulated; WBC, white blood cell.

Download figure to PowerPoint

Because of the low level or the lack of expression of CCR1, CCR2, and CCR4 found by flow cytometry, we also evaluated these receptors by Western blotting. CCR2 and CCR4 were clearly detected, and the band intensity was increased in TNFα-primed cells.

Considering the redundancy of chemokines for their receptors, we decided to complete these studies on chemokine receptors by functional assays that evaluated the binding of labeled RANTES and SDF-1 on MSCs. As shown in Figure 2C, both RANTES and SDF-1 bound to unstimulated BM MSCs. After TNFα stimulation, an increase in binding was observed for RANTES but not for SDF-1. These findings are consistent with the clear TNFα-enhanced MSC migration toward RANTES and the increased expression of CCR3 (one of the RANTES receptors) observed in TNFα-primed MSCs.

mRNAs for Growth Factors, Chemokines, and Proteases Are Differentially Expressed by Unstimulated and TNFα-Stimulated BM MSCs

Although TNFα-enhanced BM MSC migration could be explained by the increase in cytokine receptor expression, we investigated whether the modulation of soluble factors could also contribute to this phenomenon. We studied the expression pattern of a total of 44 genes known to be implicated in cell migration: 22 CC, CXC, CX3C, and XC chemokines; 12 cytokines (11 GFs and TNFα); and 10 proteases and inhibitors. We used TLDAs, which allow the level of mRNA to be quantified compared with a housekeeping gene (GAPDH). Several genes encoding chemokines (Fig. 3A) were found to be expressed constitutively at a high level (MCP-1, GROα, IL8, and SDF-1) or at a low/intermediate level (RANTES, MCP-3, Eotaxin-3, Fractalkine, GROβ, ENA-78, IP-10, and I-TAC). TNFα stimulation (a) induced the expression of I-309, MIP-1α, Eotaxin-1, TARC, and MIG; (b) increased the expression of MCP-1, RANTES, Fractalkine, GROα, GROβ, ENA-78, IL8, IP-10, and i-TAC; and (c) did not modulate the expression of MCP-3, Eotaxin-3, and SDF-1. Some chemokine genes were not expressed at all (MDC, Eotaxin-2, TECK, BCA-1, and Lymphotactin).

thumbnail image

Figure Figure 3.. mRNA expression of chemokines and growth factors on unstimulated and TNFa-stimulated bone marrow MSCs. Columns represent mean values (and bars represent the SEM) of three separated experiments assessed by TaqMan low density arrays, in black for unstimulated MSCs and in white for TNFa-stimulated MSCs. Twenty-two chemokines, 12 growth factors, and 10 proteases and inhibitors were evaluated. Expression level of each gene is represented by 2−ΔCT, where ΔCT = CT (gene X) − CT (GAPDH). #, Values of TNFa-stimulated MSCs were higher than values of unstimulated MSCs in the three experiments. Abbreviations: EGF, epidermal growth factor; FGF, fibroblast growth factor; HGF, hepatocyte growth factor; IGF, insulin-like growth factor; IL, interleukin; MMP, matrix metalloproteinase; PDGFa, platelet-derived growth factor α; PDGFb, platelet-derived growth factor β; TNFa, tumor necrosis factor α; VEGF, vascular endothelial growth factor.

Download figure to PowerPoint

We detected mRNAs at intermediate/high levels for all GF genes assayed (Fig. 3B), except those for PDGF-B and TNFα, which were expressed only after TNFα stimulation.

Proteases are well known to play a critical role in cell migration, because of their ability to degrade extracellular matrix components or locally produced chemokines. We therefore investigated the mRNA expression of some proteases, such as matrix metalloproteinases (MMPs), in BM MSC migration, particularly after TNFα stimulation (Fig. 3C). MMP-2, its activator MMP-14 (MT1-MMP), and the inhibitors TIMP-1 and TIMP-2 were found constitutively expressed at a high level in BM MSCs. A similar expression pattern was seen for CD26/dipeptidyl peptidase IV (DPPIV), an ectoenzyme able to inactivate a number of chemokines (including SDF-1) by limited cleavage of the molecule [32]. In contrast, the mRNA expression found at intermediate/low levels for MMP-1, -3, -7, and -13 was increased to a high level after TNFα stimulation. MMP-9 was the only MMP not expressed at steady state or induced by TNFα stimulation.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

This study shows that BM MSCs are able to migrate in vitro in response to a large set of chemotactic factors, including both GFs and chemokines. The GFs PDGF-AB and IGF-1 appear to be the most potent factors, whereas some chemokines (RANTES, MDC, and SDF-1) have limited effect. However, preincubation with the inflammatory cytokine TNFα leads to increased MSC migration toward chemokines, especially RANTES. The migration capacity of MSCs clearly stands in contrast to that of HSCs whose migration is induced mainly by a single chemokine, SDF-1 [33], whereas hematopoietic GFs display little or no chemotactic activity [34].

We tested seven GFs known to be produced by mesenchymal cells and known for their promigratory, proliferating, or differentiating activity on mesenchymal cells. In our study, the two most chemoattractive GFs were PDGF-AB and IGF-1, which confirms reports by Fiedler et al. [28, 35]. The three other GFs with chemoattractive capacity were HGF, EGF, and Ang-1. HGF and EGF are already reported as chemoattractants for MSCs [29, 30, 36]. Ang-1 is a crucial chemotactic factor for endothelial cells through its receptor Tie-2 [37]. Recently, HSCs expressing Tie-2 have been shown to respond to Ang-1 by inducing cobblestone area formation and maintaining long-term repopulating activity [38]. In the present study, we show, for the first time to the best of our knowledge, that human (Hu) BM-derived MSCs express Tie-1 and are chemoattracted by Ang-1 under proinflammatory conditions (i.e., in the presence of TNFα). This result is in contrast with the negative results we observed for VEGF, another potent angiogenic factor, underlining a major difference between MSCs and endothelial cells. As expected from the promigratory response, MSCs expressed, at a moderate to high level, receptors for PDGF-AB, IGF-1, HGF, and EGF. MSC migration capacity to these major chemotactic GFs was unchanged after TNFα exposure, as was the membrane receptor expression.

The chemotactic activity of chemokines on Hu BM MSCs appeared clearly less efficient than that of GFs, with only three of the nine chemokines tested displaying a significant activity (RANTES/CCL5, SDF-1/CXCL12, and MDC/CCL22). The most active was MDC, with induced migration levels equivalent to those obtained with the less active GFs (EGF and HGF). RANTES and SDF-1 showed a lesser, although significant, effect. Consistent with these results, we detected on MSCs the receptors for RANTES (CCR3, CCR4, and CCR5, but not CCR1), MDC (CCR4), and SDF-1 (CXCR4). Although found at low levels by flow cytometry, CCR2 and CCR4 expression was more evident using Western blotting, probably because of the concomitant detection of the intracellular pool. In contrast, CCR1 was not detected by either technique. The most striking finding was the dramatic increase, after TNFα stimulation, in MSC sensitivity to RANTES, MDC, and SDF-1. In TNFα-primed cells, we observed not only increased migration in response to RANTES and MDC but also increased expression of the receptors CCR3 and CCR4, using flow cytometry or Western blotting. The functional relevance of these observations was confirmed by experiments showing that the membrane binding of RANTES was increased in TNFα-stimulated cells. In contrast, TNFα-stimulated migration of MSCs to SDF-1 was not associated with either increased membrane expression of its receptor CXCR4 or increased SDF-1 surface binding, a discrepancy already reported for dendritic cells submitted to a similar regimen [39]. TNFα might modify the MSC sensitivity to SDF-1 by modulating CXCR4 signal transduction without affecting receptor expression, as described after inhibition of the protein kinase C pathway [40]. Some reports have recently pointed out the potential role of the SDF-1/CXCR4 axis on MSC migration [25, [26]27, 41]. Our study not only confirms these reports but also indicates that the promigratory activity of this chemokine is increased in TNFα-primed cells, suggesting an increased role of SDF-1 under inflammatory conditions in which TNFα can be overexpressed. The activity of RANTES on BM MSCs has been reported [25]; our results clearly indicate that it is, together with SDF-1, the most efficient chemokine after TNFα-stimulation. Finally, we show, for the first time to the best of our knowledge, that MDC, a chemokine active on mature hematopoietic cells [42], is also a potent chemoattractant for Hu BM MSCs, even displaying the strongest activity among the chemokines tested on steady-state cells.

Using large-scale quantitative RT-PCR, we evaluated, in TNFα-primed cells and in unstimulated cells, the expression at the mRNA level of a large panel of factors, including not only chemokines (from the CC, CXC, XC, and CX3C families) and cytokines but also proteases and inhibitors. Of the 22 chemokines tested, only 5 were not expressed (including MDC), 14 were induced or increased after TNFα treatment (including MCP-1 and RANTES), and no change was observed for 3 (including SDF-1). Of the 11 GFs tested, 10 were constitutively expressed (with TNFα-increased levels only for IGF-1), and only 1 was induced (PDGF-B). Considering these results, we may have underevaluated the observed promigratory activity of certain chemokines and GFs because of the autocrine MSC production leading to the reduction of the imposed concentration gradient. Proteases, such as MMPs and CD26/DPPIV, are able to favor cell locomotion and tissue reconstitution by breaking down the extracellular matrix but also by degrading locally a number of chemokines [43]. In particular, MMP-2 and MT1-MMP (MMP-14) have been reported to be involved in MSC migration induced by PDGF [44], HGF, or SDF-1 [41], whereas MT1-MMP cellular expression has been shown very recently to be upregulated by inflammatory cytokines, including TNFα [45]. In the present study, we showed that MMP-1, -2, -3, -13, MT1-MMP, CD26, and the inhibitors TIMP-1 and -2 are constitutively expressed by BM MSCs, in contrast to MMP-7 and -9. TNFα treatment strongly enhances the expression of MMP-1, -3, and -13 and induces that of MMP-9, as found by other authors [45]. Thus, MSC migration might be favored by increased local concentrations of collagenases and gelatinases, particularly in inflammatory situation.

Our results on chemotactic factors, either GFs or chemokines, are summarized in Table 1. The chemotactic factors identified here are particularly relevant for MSC homing and tissue repair, since tissue damage has been associated with local increase in mediators, such as those reported for SDF-1 or HGF [26, 46, 47]. In vivo use of chemotactic factors for MSCs may be of clinical interest, and they now need to be tested in animal models. On the one hand, local administration might improve MSC homing to specific tissues, as recently reported with IGF-1, HGF, or PDGF after myocardial damage [48], and, on the other hand, systemic administration might trigger endogenous MSC mobilization for tissue repair, avoiding ex vivo amplification. In addition, the local or systemic production of inflammatory mediators might influence not only MSC migration, as shown in this study and another one [45], but also MSC proliferation and differentiation [49] and MSC engraftment [50].

Table Table 1.. Summary of the sensitivity pattern of human bone marrow MSCs to chemotactic factors compared to relevant membrane receptor expression
Thumbnail image of

In conclusion, our data indicate that the migration capacity of BM MSCs is under the control of a large range of receptor tyrosine kinase receptor GFs and CC and CXC chemokines. These chemokines appear mostly effective on TNFα-primed cells, suggesting that the mobilization of MSCs and their subsequent homing to injured tissues may depend on the systemic and local inflammatory state. This parameter has to be taken into account in cell therapy protocols using BM MSCs.

Disclosures of Potential Conflicts of Interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

The authors indicate no potential conflicts of interest.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References

We are grateful to Elfi Ducrocq and Marie-Christine Bernard for expert technical assistance. This work was supported by grants from the European Integrated projects FIRST (contract no. 503436) and GENOSTEM (contract no. 503161), from the French Cancéropôle Grand-Ouest network, and from the Committees of Cher and Vendée of the French League against Cancer.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures of Potential Conflicts of Interest
  8. Acknowledgements
  9. References