SEARCH

SEARCH BY CITATION

Keywords:

  • Bone marrow stromal cells;
  • Mesenchymal stem cells;
  • Monocyte;
  • T cells;
  • Bone marrow stromal;
  • CCL18

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Multipotent stromal cells (MSC) have been shown to possess immunomodulatory capacities and are therefore explored as a novel cellular therapy. One of the mechanisms through which MSC modulate immune responses is by the promotion of regulatory T cell (Treg) formation. In this study, we focused on the cellular interactions and secreted factors that are essential in this process. Using an in vitro culture system, we showed that culture-expanded bone marrow-derived MSC promote the generation of CD4+CD25hiFoxP3+ T cells in human PBMC populations and that these populations are functionally suppressive. Similar results were obtained with MSC-conditioned medium, indicating that this process is dependent on soluble factors secreted by the MSC. Antibody neutralization studies showed that TGF-β1 mediates induction of Tregs. TGF-β1 is constitutively secreted by MSC, suggesting that the MSC-induced generation of Tregs by TGF-β1 was independent of the interaction between MSC and PBMC. Monocyte-depletion studies showed that monocytes are indispensable for MSC-induced Treg formation. MSC promote the survival of monocytes and induce differentiation toward macrophage type 2 cells that express CD206 and CD163 and secrete high levels of IL-10 and CCL-18, which is mediated by as yet unidentified MSC-derived soluble factors. CCL18 proved to be responsible for the observed Treg induction. These data indicate that MSC promote the generation of Tregs. Both the direct pathway through the constitutive production of TGF-β1 and the indirect novel pathway involving the differentiation of monocytes toward CCL18 producing type 2 macrophages are essential for the generation of Tregs induced by MSC. Stem Cells 2013;31:1980-1991


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Regulatory T cells (Tregs) are the primary mediators of peripheral tolerance, which maintain immune homeostasis and are moderators of inflammation. Several types of Tregs have been described, including naturally occurring Tregs (nTregs), that develop in the thymus, and induced Tregs (iTregs), that are generated in the periphery and arise from conventional T cells [1, 2]. The mechanisms by which iTregs are induced have not been fully elucidated. The transcription factor forkhead box P3 (FoxP3), which is a key regulatory gene for the development of Tregs [3], is expressed in both nTregs and iTregs.

Multipotent stromal cells (MSC) display immunomodulatory capacities, including the capacity to suppress the proliferation of T cells, B cells, and interleukin-2 (IL-2) stimulated natural killer cells [4-6] and to inhibit the differentiation and maturation of dendritic cells [7]. Another important mechanism by which MSC exert their immunomodulatory functions is the induction of Tregs. In in vivo studies regarding allergic airway inflammation and heart transplantation, MSC were shown to induce the generation of Tregs [8, 9]. Understanding the mechanisms behind the MSC-induced formation of Tregs is important for the application of MSC as a therapeutic agent.

The mechanisms by which MSC induce the generation of Tregs have been studied in vitro. These studies involved cocultures of MSC and peripheral blood mononuclear cells (PBMC) and showed the generation of functionally suppressive CD4+CD25hiFoxP3+ Tregs [10-14]. According to some studies soluble factors are involved, while other studies indicate the requirement of cell–cell contact. The molecular pathways that are involved in the MSC-induced Treg formation have not yet been fully identified. English et al. [12] claimed the involvement of transforming growth factor beta 1 (TGF-β1) and prostaglandin E2 (PGE2), while others found that MSC-induced Treg formation was not reduced by neutralizing antibodies against these factors [13].

In this study, we explored the cellular and molecular pathways underlying the Treg promoting effect of MSC. We found that the MSC-induced formation of Tregs is driven by TGF-β1, which is produced by MSC, and by CCL18, which is produced by monocytes upon interaction with MSC.

Materials and Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Generation of Human MSC and PBMC

After obtaining informed consent, adult bone marrow was harvested from healthy donors or orthopedic patients. Mononuclear cells were isolated using a Ficoll-Paque density gradient (1.077 g/cm3) and were plated at 1.3 × 105 per centimeter square in Dulbecco's modified Eagle's medium-low glucose (Invitrogen Corp., Paisley, U.K, http://www.invitrogen.com) supplemented with 10% fetal calf serum (FCS; Greiner Bio-one, Alphen a/d Rijn, The Netherlands, www.greinerbioone.com/ne/netherlands) and Penicillin/Streptomycin (P/S; Invitrogen Corp.). After 3–4 days, nonadherent cells were removed and medium was refreshed every 3–4 days until cells reached approximately 90% confluence. The MSC monolayer was detached using trypsin/EDTA (Invitrogen Corp.) and reseeded at 4,000 per cm2 for further expansion. The MSC were characterized by flowcytometric analysis and used in the experiments at passage 2–5.

Human PBMC were isolated from buffy coats from healthy donors obtained from Sanquin Blood Supply (Leiden, The Netherlands, www.sanquin.nl) using a Ficoll-Paque density gradient. CD4+ T cells were isolated from PBMC by magnetic activated cell sorting (MACS) using a cocktail of biotin-labeled antibodies against CD8, CD14, CD16, CD19, CD36, CD56, CD123, TCRγ/δ, and CD235a and anti-biotin microbeads followed by separation on a MACS LS column according to the manufacturer's recommendations (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com). Monocytes were isolated or depleted from PBMC populations by several rounds of plastic adherence or by MACS using CD14 microbeads (Miltenyi Biotec GmbH), followed by MACS LS column separation according to the manufacturer's recommendations.

PBMC-MSC Co-Cultures

To assess the capacity of MSC to induce the generation of Tregs, transwell cocultures were performed as described earlier [14]. In short, PBMC were cocultured with allogeneic MSC for 1 week followed by 1 week of culture in the absence of MSC. As a control experiment, a coculture of PBMC with autologous MSC was performed. At day 0, 100,000 MSC were plated in 12-well plates in RPMI medium (Invitrogen Corp.) containing P/S, l-glutamine (Invitrogen Corp.) and 10% FCS, and in a transwell insert (pore size 0.4 μM, Corning, Inc., Lowell, MA, http://www.corning.com/lifesciences) 500,000 PBMC were added. At day 7, PBMC were harvested from the inserts and washed (day 7-PBMC). Cells were counted and resuspended in fresh medium and cultured for 7 days in the absence of MSC. At day 14, PBMC were harvested (day 14-PBMC), counted, and analyzed by flow cytometry.

In addition to the transwell cocultures, MSC-conditioned medium (CM) was tested for its ability to promote generation of CD4+CD25hiFoxP3+ T cells. Near-confluent cultures of MSC were cultured with RPMI supplemented with P/S, l-glutamin, and 10% FCS for 5–7 days. The cells were harvested and counted and the cell-free culture supernatant was concentrated using 10K Centriprep Centrifugal filters (Millipore Corp., Billerica, MA, http://www.millipore.com). PBMC were cultured for 7 days in a 12-well plate with a CM equivalent of 100,000 MSC supplemented until 1 mL with fresh RPMI containing P/S and l-glutamin. At day 7, the PBMC were harvested and resuspended in fresh medium and cultured for another 7 days. At day 14, PBMC were harvested, counted, and analyzed by using flow cytometry.

The role of molecular factors in the induction of Treg formation by MSC was assessed by blocking experiments. In the first week of coculture, antibodies against IL-6 (0.2–1.0 μg/mL), interferon-gamma (IFN-γ) (1.0 μg/mL), and CCL18 (0.5–2.0 μg/mL; all from R&D Systems Europe Ltd., Abingdon, U.K., http://www.rndsystems.com) and an inhibitor of the TFG-β1-receptor (TFG-β1-R) (10 μM, Sigma-Aldrich Chemie BV, Zwijndrecht, The Netherlands, http://www.sigmaaldrich.com) were added to the medium. Control experiments were performed by addition of equal concentrations of a nonspecific antibody of the same isotype to the cocultures. The role of heme oxygenase-1 (HO-1) was determined by adding the inhibiting agent tin protoporphyrin IX dichloride (SnPP, 10 μM, Tocris Bioscience, Bristol, UK, www.tocris.com).

Monocyte-MSC Cocultures

To directly investigate the effect of MSC on monocytes, 3-day cultures were performed of freshly isolated monocytes in unconcentrated CM diluted with RPMI medium in a ratio 2:1. To reveal an effect of MSC on monocytes, apart from their monocyte survival-enhancing effect, we compared monocytes that were stimulated with CM versus RPMI medium with and without the addition of 5 ng/mL macrophage colony-stimulating factor (M-CSF) (Peprotech EC Ltd. London, U.K., http://www.peprotech.com), neutralizing antibodies against the M-CSF-receptor (M-CSF-R; concentration [1:100]–[1:500]), the inhibitor of the TFG-β1-receptor (10 μM) or recombinant TFG-β1 (10 ng/mL; BD Biosciences, San Diego, CA, http://www.bdbiosciences.com). At day 3 of the culture the monocytes were harvested and analyzed for surface marker expression using flow cytometry, counted, and assessed for viability using eosin staining. RNA was extracted for gene expression analysis using RNeasy columns (Qiagen GmbH, Hilden, Germany, http://www1.qiagen.com). Cell-free supernatants were collected and stored at −20°C for analysis of cytokine concentrations.

Functionality Assay

For some experiments, PBMC at day 14 of the PBMC-MSC coculture were separated by FACS (FACSAria III cell sorter; BD Biosciences) in a CD25 fraction and a CD25hi fraction. The complete population of PBMC that were cultured in the presence and absence of MSC and the sorted CD25 fraction and a CD25hi fraction PBMC cultured in the presence of MSC were tested for their immunosuppressive capacity in an autologous suppression assay. To this end, autologous CD4+CD25 cells were isolated from PBMC (using the CD4+CD25+ regulatory T-cell isolation kit [Miltenyi Biotec GmbH]). CD4 T cells were isolated from PBMC as described above, followed by separation of CD4+CD25+ and CD4+CD25 T cells by CD25 microbeads according to manufacturer's recommendations (Miltenyi Biotec GmbH). CD4+CD25 T cells (100,000 per well) were stimulated with human T-activator CD3/CD28 dynabeads (Invitrogen Corp.) in a bead:cell ratio 1:5. To this culture, day 14-PBMC, day 14-CD25, or day 14-CD25hi cells were added in a 1:1 ratio with CD4+CD25 T-cells. After 5 days, the cells were pulsed with [3H]-thymidine (0.5 μCi per well) and incubated for 16 hours at 37°C. The cultures were harvested on a glass fiber filter and thymidine incorporation was measured with a liquid scintillation counter (Wallac, Turku, Finland, www.perkinelmer.com).

Analysis of Cytokine Production

Cytokine concentrations were measured in the cell-free supernatant of day 14 from the PBMC culture in the presence and absence of CM (IL-10) and in the supernatant from the 3-day monocyte cultures with and without M-CSF and/or CM (IL-10, IL-12, and CCL18). Constitutive production by MSC of the cytokines M-CSF and TGF-β1 were measured in cell-free culture supernatant of unstimulated of MSC. Cytokine concentrations were measured using sandwich ELISA (IL-10, IL-12, and TGF-β1: BD Biosciences; M-CSF and CCL18: R&D systems Europe Ltd.) or a Bio-Plex Pro Human Cytokine 27-plex panel (Bio-Rad laboratories, Inc., Hercules, CA, http://www.bio-rad.com). Cytokine concentrations were determined using a standard curve. Statistical analysis was performed using a one-way ANOVA and the differences between groups were analyzed using a paired t test. A p < .05 was considered statistically significant.

Flow Cytometry

Antibodies used for flow cytometric analysis of PBMC and monocytes were anti-CD4 peridinin-chlorophyll protein complex-Cy5.5 (PerCP-Cy5.5), anti-CD69 fluorescein isothiocyanate (FITC), anti-CD25 phycoeryhrin-Cy7 (PE-Cy7), anti-CD14 phycoerythrin (PE), anti-CD127 biotin, anti-CD206 allophycocyanin APC (BD Biosciences), anti-CD25 PE (Miltenyi Biotec GmbH), anti-FoxP3 APC (eBioscience, Inc., San Diego, CA, www.ebioscience.com), anti-CD163 PerCP-Cy5.5, anti-CD80 PE-Cy7 (Biolegend, Inc., San Diego, CA, www.biolegend.com), and streptavidin-Pacific blue (Invitrogen Corp.). MSC were analyzed for expression of surface markers using FITC-conjugated and PE-conjugated antibodies against: CD90, CD73, CD45, CD34, HLA-DR, HLA-ABC, CD80 (BD Biosciences), and CD105 (Ancell Corp., Bayport, MN, www.ancell.com). For surface staining, primary antibodies were added and the cells were incubated for 30 minutes at 4°C in the dark. Cells were washed with phosphate buffered saline (PBS)/1% albumin, secondary antibodies were added and incubated for 30 minutes at 4°C in the dark. For intracellular FoxP3 staining, the FoxP3 staining buffer kit (eBioscience, Inc.) was used. All stainings were analyzed using a FACSCanto II (BD Biosciences). The analysis of the acquired data was done with FlowJo software version 7.6.1 (Tree Star, Inc., Ashland, OR, www.flowjo.com). Mean fluorescence intensity was analyzed and statistical analysis was performed using a one-way ANOVA, the differences between groups were analyzed using a paired t test and a p < .05 was considered statistically significant.

Gene Expression Analysis

To study the effect of MSC on monocytes, we performed a microarray analysis on monocytes cultured for 3 days with CM and/or M-CSF. From these monocyte populations, RNA was extracted using the RNeasy micro kit (Qiagen GmbH). For microarray analysis, the RNA was amplified and biotinylated using the Illumina Totalprep RNA amplification kit (Life Technologies Europe BV, Bleiswijk, The Netherlands, http://www.lifetech.com). The samples were hybridized on a HumanHT-12 v4 Expression BeadChip according to the manufacturer's instructions. Chips were scanned on an Illumina BeadArray 500GX reader and analyzed with BeadStudio software (all from Illumina Inc., San Diego, CA, www.illumina.com). Gene expression data were obtained from BeadStudio and processed in R using Bioconductor [15] and the R beadarray package [16], using the BASH algorithm for outlier detection. The remaining bead information was summarized and normalized using quantile normalization. Due to considerable donor variation between the different monocyte sources, differentially expressed genes were determined using the RankProd [17] method, considering a percentage false-positive rate of 0.05 significant. Gene set enrichment analysis (GSEA) was performed using the Broad Institute GSEA tool [18].

For quantitative reverse transcriptase polymerase chain reaction (RT-qPCR) analysis cDNA was synthesized with Superscript III RT (Invitrogen GmbH, Hilden, Germany). RT-qPCR analyses were performed on a StepOnePlus real-time PCR system (Applied Biosystems, Foster city, CA, http://www.appliedbiosystems.com) using SYBR Green reagent (Applied Biosystems). The primers sets used for RT-qPCR are listed in Table 1. All RT-qPCR data were normalized to β-actin expression and analyzed using the δ-Ct method.

Table 1. Primer sets used for RT-qPCR
GeneForwardReverse
IL-105'-CCGAGATGCCTTCAGCAGAG-3'5'-GGTCTTGGTTCTCAGCTTGG-3'
GAS65'-GCCTTTCAGGTCTTCGAGGAG-3'5'-GTCAGGCAGGTTTTGCACG-3'
CCL185'-ACAAAGAGCTCTGCTGCCTC-3'5'-CCCACTTCTTATTGGGGTCA-3'
MMP95'-TGGCAGAGGAATACCTGTACC-3'5'-GAGTAGTTTTGGATCCAATAGG-3'
β-Actin5'-AGGCATCCTCACCCTGAAGTA-3'5'- CACACGCAGCTCATTGTAGA-3'

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Soluble Factors Produced by MSC Promote the Formation of CD4+CD25HiFoxP3+ T Cells

The formation of Treg was studied in cocultures of PBMC and MSC. At day 7 of these cocultures, PBMC that were cultured in the presence and absence of MSC both contained a population of CD4+CD25hiFoxP3+ T cells (Fig. 1A). At day 14, however, PBMC cultured during the first week in the presence of MSC contained a significantly higher percentage of CD4+CD25hiFoxP3+ T-cells than PBMC cultured in the absence of MSC (Fig. 1A, 1B; 0.36% vs. 10.04%, n = 8, p < .001). The CD4+CD25hiFoxP3+ T cells did not express the T-cell activation marker CD69 and CD127 (Supporting Information Fig. 1), compliant with the reported Treg phenotype [19]. The formation of Tregs induced by MSC was observed when PBMC were cultured in the presence of both autologous and allogeneic MSC. Similar results were obtained with MSC CM (Fig. 1C), indicating that factors secreted by MSC were responsible for this effect.

image

Figure 1. Coculture of MSC with PBMC results in the induction of functional CD4+CD25hiFoxP3+ T cells. (A): Representative dot plots of expression of CD4, CD25, and FoxP3 on PBMC on day 7 and 14. (B): The percentage of CD4+CD25hi T cells at day 14 is significantly higher in the population that was cultured during the first week in coculture with MSC compared to PBMC that were cultured without MSC (data are means ± SEM from eight different experiments with nine different MSC and seven different PBMC. Statistical analysis was performed with a paired t test; ***, p < .001). (C): The presence of CM in the PBMC culture obtains similar results as the presence of MSC (data are means ± SEM from five different experiments with five different MSC and two different PBMC. Statistical analysis was performed with a paired t test; ***, p < .001). (D): The IL-10 protein concentrations measured at day 14 were significantly higher in the culture supernatants of PBMC cultures in the presence of MSC compared to cultures in the absence of MSC (data are means ± SEM from two different experiments with three different MSC and two different PBMC. Statistical analysis was performed using a paired t test (**, p < .01). (E, F): Autologous CD25 T cells were stimulated with anti-CD3/CD28 beads (control) with or without PBMC from the coculture experiment at day 14 (E) or with sorted CD4+CD25 and CD4+CD25+ T cells from the day 14-PBMC cultured with MSC (F). Day 14-PBMC were also tested separately for their proliferation upon CD3/CD28 stimulation (negative control). Proliferation was measured using 3H-thymidine incorporation after 5 days (data are mean from triplicates ± SEM). Statistical analysis was performed with a Student's t test (**, p < .01; ***, p < .001). Abbreviations: CM, MSC-derived conditioned media; CCPM, corrected counts per minute; IL, interleukin; MSC, multipotent stromal cell; PBMC, peripheral blood mononuclear cells.

Download figure to PowerPoint

Cytokine measurements in culture supernatants of MSC and PBMC cocultures indicated a threefold increase in IL-10 concentrations in comparison with culture supernatants of PBMC that were cultured in the absence of MSC (Fig. 1D; 27.74 pg/mL vs. 10.49 pg/mL, n = 4, p = .001). These concentrations fell within the standard curve. This indicates that CM induced the cells in the PBMC population to secrete IL-10.

CD4+CD25HiFoxP3+ T Cells Induced by MSC Exhibit Regulatory Function

To study the function of the MSC-induced CD4+CD25hiFoxP3+ T cells, the PBMC populations generated in the absence and presence of MSC were tested for their immunoregulatory capacity. The total PBMC populations obtained at day 14 after coculture with MSC exhibited a significantly enhanced capacity to suppress T-cell proliferation compared to PBMC populations that were generated in the absence of MSC (Fig. 1E). To test specifically whether the CD25hi T cells were responsible for this enhanced immunosuppression, PBMC from day 14 of culture in the presence of CM were separated in a CD4+CD25 and a CD4+CD25hi T-cell fraction and tested in a separate autologous suppression assay. At day 14, the CD4+CD25hi T cells showed a significantly higher suppression of T-cell proliferation compared to the CD4+CD25 T cells (Fig. 1F), indicating that the CD4+CD25hiFoxP3+ T-cell population that is formed during coculture with MSC or CM has regulatory function.

Generation of Tregs Is Mediated by TGF-β1

In order to identify the factors involved in the MSC-induced Treg formation, antibody neutralization and inhibitor experiments were performed. Addition of neutralizing antibodies to the TGF-β1-receptor to the PBMC-MSC cocultures resulted in a significantly reduced generation of Tregs compared to control cultures without antibodies (4.6%, vs. 20.2% p = .02, n = 5, Fig. 2A). MSC constitutively secreted TGF-β1 as was measured with ELISA in cell free culture supernatant of unstimulated MSC (Supporting Information Fig. 2). Addition of neutralizing antibodies against IL-6 did not result in a significant reduction (p = .38) of the MSC-induced generation of Tregs (Fig. 2B). As expected, neutralizing antibodies against IFN-γ, which is not constitutively expressed by MSC, did not influence the effect of MSC on the generation of Tregs (Fig. 2C). Finally, addition of the HO-1 inhibitor SnPP significantly reduced the population of Tregs at day 14 (13.6%, vs. 41.5% p = .04, n = 3) (Fig. 2D).

image

Figure 2. The induction of Treg formation by MSC is mediated by several factors. (A): Addition of blocking antibodies against the TGF-β1-receptor to the coculture during the first week significantly reduced the formation of Tregs. (B): Addition of neutralizing antibodies against IL-6 slightly diminished the MSC-induced formation of Tregs but not significant (p = .38). (C): Addition of neutralizing antibodies against IFNγ did not affect the MSC-mediated induction of Tregs. (D): Inhibition of HO-1 by SnPP resulted in a significant lower percentage of Tregs. Means ± SEM are shown of three to five different experiments with different MSC ad PBMC donors in each experiment, statistical analysis was performed using a Student's t test (*, p < .05). Abbreviations: TGF-β, transforming growth factor beta; IFN, interferon; SnPP, tin protoporphyrin IX dichloride.

Download figure to PowerPoint

Monocytes Are Required for Induction of CD4+CD25HiFoxP3+ T Cells

Since CD4+CD25hiFoxP3+ Tregs are derived from the CD4+ T-cell population [3], we investigated whether a Treg population could be generated by coculture of purified CD4+ T cells with MSC (Fig. 3). MSC-induced Tregs could not be generated from the purified CD4+ population, as opposed to PBMC, indicating that the PBMC population contains accessory cells mediating the formation of Tregs (Fig. 3A). Following depletion of monocytes from the PBMC population, no Treg population could be formed (Fig. 3A, 3B). To exclude the possibility that the depletion procedure interfered with the formation of Tregs, we added monocytes back to the monocyte-depleted PBMC populations. This restored the formation of the Treg population (Fig. 3A, 3B). These data show that monocytes are essential in mediating MSC-induced formation of Tregs.

image

Figure 3. Monocytes are indispensable for the induction of CD4+CD25hiFoxP3+ T cells. (A): Total PBMC (Ai), purified CD4+ T cells (Aii), and monocyte-depleted (Aiii) and repleted PBMC (Aiv) were cocultured with MSC and tested for their ability to generate CD4+CD25hiFoxP3+ T cells (n = 5). Results are shown as representative dot plots. (B): Cumulative data of five experiments. The population CD4+CD25hi T cells is significantly lower when either MSC or monocytes are depleted from the culture (data are means ± SEM of five different experiments, statistical analysis was performed using a Student's t test [*, p < .05]). Abbreviations: MSC, multipotent stromal cells; PBMC, peripheral blood mononuclear cells.

Download figure to PowerPoint

MSC Promote the Survival of Monocytes

Since monocytes were essential for the MSC-induced Treg formation, we focused on the effect of MSC and CM on monocytes in PBMC-MSC cocultures and by performing cultures of freshly isolated monocytes stimulated with CM. In the presence of MSC, the PBMC cultures contained a higher percentage of CD14+CD16+ cells after 4 and 7 days compared to cultures in the absence of MSC (Fig. 4A). Staining with Annexin V and CD1a showed that the monocytes that lose CD14 undergo apoptosis and do not differentiate toward dendritic cells (data not shown), indicating that MSC promote survival of monocytes in the culture. This survival-promoting effect was confirmed in 3-day cultures of freshly isolated monocytes in the presence of CM. Similar results were observed in cultures of freshly isolated monocytes stimulated with M-CSF. M-CSF, a factor that is produced by MSC (Supporting Information Fig. 2) and is known to promote monocyte survival, did not further enhance the effect of CM on monocyte survival (Fig. 4B).

image

Figure 4. MSC promote survival of monocytes. (A): In the presence of MSC, a higher percentage of CD14+CD16+ monocytes is present in the PBMC population after 7 days of culture than in PBMC cultured without MSC. (B): Stimulation with either M-CSF or CM increases survival of monocytes in a 3-day culture compared to control cultures (data are means ± SEM from three different MSC; ***, p = .001). (C): Heatmap representing upregulated genes in monocytes cultured in the presence of CM compared to control cultures of monocytes cultured in the absence CM. The Venn-diagram represents the overlapping differentially regulated genes between the comparison of CM versus control cultures and the CM+M-CSF versus M-CSF cultures. Abbreviations: CM, conditioned medium; M-CSF, macrophage colony-stimulating factor; MSC, multipotent stromal cell; PBMC, peripheral blood mononuclear cells.

Download figure to PowerPoint

To explore the transcriptional changes in monocytes in response to MSC exposure, we performed a genome-wide gene expression profiling of monocytes from three different donors. Comparison of the two datasets (“CM stimulated monocytes vs. unstimulated monocytes” and “CM + M-CSF stimulated monocytes vs. M-CSF stimulated monocytes”) resulted in 43 overlapping unique genes that were differentially expressed after stimulation with CM (Fig. 4C). GSEA showed downregulation of apoptosis-related genes [20] in monocytes in response to CM, in line with the survival promoting effect of MSC on monocytes (Supporting Information Fig. 3).

MSC Skew Monocytes Toward an Anti-Inflammatory Profile

Phenotype Analysis

To further study the effect of MSC on monocytes, phenotypic changes of monocytes during cocultures with MSC were studied. At day 7 of the cocultures, the expression of the macrophage type 2 markers CD206 and CD163 and of CD80 was increased on monocytes compared to the cultures without MSC (Fig. 5A). Further investigation of the effect of MSC on freshly isolated monocytes showed that M-CSF as well as CM increased the expression of the surface markers CD163, CD206, and CD80 on monocytes, indicating a skewing of monocytes toward a macrophage type 2 phenotype (Fig. 5B). The expression of CD163 was further increased in the presence of both M-CSF and CM, which shows an additive effect of CM. Expression of CD86 was not changed by M-CSF alone but was significantly decreased after culture with CM. To investigate the involvement of M-CSF in the effect of CM we added neutralizing antibodies against M-CSF-R to monocytes cultured in the presence of CM. We found that the increased expression of the surface markers CD80, CD163, and CD206 on monocytes cultured in the presence of CM (Fig. 6A) as well as the increase in survival (Fig. 6B) was reversed in the presence of neutralizing antibodies against M-CSF-R, showing that M-CSF is responsible for these effects. Altogether, these surface marker expression data indicate that some of the effects of CM can be explained by M-CSF, but that additional soluble factors, present in the CM, are involved in skewing of monocytes towards type 2 macrophages.

image

Figure 5. MSC-CM induces skewing of monocytes toward differentiation into a type-II macrophage. (A): The MFI of CD206, CD163, and CD80 within the CD14+ population are shown (data are means ± SEM from two experiments with three different MSC donors; **, p < .01). Expression of surface markers on monocytes (B) and IL-10 concentrations in culture supernatant (C) after a 3-day stimulation with M-CSF and/or CM (data are means ± SEM from five different monocyte donors and CM from two different MSC; *, compared to control; **, p < .01; ***, p < .001; #, compared to M-CSF; #, p < .05; ##, p < .01; ###, p < .001). (D): IL-10 gene expression is increased in monocytes that were cultured with CM compared to M-CSF and control cultures (data are means ± SD from three different monocyte donors and two different MSC donors; *, p < .05; **, p < .01). (E): CCL18, GAS6, and MMP9 gene expression in monocytes that were cultured for 3 days in CM and/or M-CSF (data are means ± SD from three different monocyte donors and two MSC donors; *, compared to control; *, p < .05; **, p < .01; #, compared to M-CSF; #, p < .05; ##, p < .01; ###, p < .001;). Abbreviations: CM, conditioned medium; IL, interleukin; M-CSF, macrophage colony-stimulating factor; MSC, multipotent stromal cell; MFI, mean fluorescence intensity.

Download figure to PowerPoint

image

Figure 6. M-CSF is partly responsible for the MSC-induced skewing of monocytes toward type-II macrophages. (A): The MFI of CD206, CD163, and CD80 within the CD14+ population are shown (data are means ± SEM from two different MSC donors and two different monocyte donors; **, p < .01). (B): Addition of an antibody against the M-CSF-R reversed the CM-induced increased survival of monocytes in a 3-day culture. (C): IL-10 concentrations in culture supernatant after 3 days of culture of monocytes in the presence of CM and CM + anti-M-CSF-R antibodies (data are means ± SEM from two different multipotent stromal cell donors and two different monocyte donors; *, p < .05). Abbreviations: CM, conditioned medium; IL, interleukin; M-CSF, macrophage colony-stimulating factor; M-CSF-R, M-CSF-receptor; MFI, mean fluorescence intensity.

Download figure to PowerPoint

Cytokine Secretion

Monocytes were also functionally tested for cytokine secretion after stimulation with CM and/or M-CSF. Stimulation of monocytes with CM or M-CSF resulted in significantly (p < .001) increased concentrations of IL-10 in the culture supernatant at day 3. IL-10 concentrations were further increased following stimulation of monocytes with both CM and M-CSF (Fig. 5C). Expression levels of IL-10 mRNA were similar in unstimulated monocytes and monocytes that were stimulated with M-CSF (Fig. 5D), suggesting that the observed increase in IL-10 concentrations in the culture supernatant was a result of increased monocyte survival. Stimulation of monocytes with CM, however, significantly increased the expression of IL-10 mRNA by monocytes compared to the cultures of monocytes in the absence of CM. The combination of CM and M-CSF did not further increase the mRNA expression of IL-10 in monocytes compared to CM alone, suggesting that the increased IL-10 secretion that was observed after monocyte stimulation with the combination of CM and M-CSF was regulated at a post-transcriptional level. Moreover, the CM-induced increase in IL-10 secretion was not significantly reversed by addition of neutralizing antibodies against M-CSF-R (Fig. 6C), indicating that the M-CSF is not responsible for the observed IL-10 secretion by monocytes. Taken together, these data indicate that CM induced IL-10 production in monocytes and that CM is superior to M-CSF alone in this effect. IL-12 protein was not detectable in supernatant of monocyte cultures stimulated with CM. Further analysis of cytokine concentrations in the culture supernatant of monocytes showed reduced concentrations of the proinflammatory cytokines IL-1b, MIP-1a, MIP-1b, and RANTES in the culture supernatant of monocytes that were cultured in the presence of CM compared to M-CSF (Supporting Information Fig. 4).

Transcriptional Profiling of Monocytes

Gene expression profiling of monocytes showed that by CM stimulation of monocytes resulted in upregulation of several genes, including CD163, GAS6, CCL18, PLTP, that are associated with the type 2 macrophage phenotype (Fig. 4C). Additionally, GSEA showed that the STAT3 pathway was activated in monocytes that were cultured in the presence of CM [21], which is another characteristic of type 2 macrophages (Supporting Information Fig. 3). Microarray results were confirmed and validated by RT-qPCR in samples from other monocyte donors and stimulated with CM from another MSC donor than those used for the microarray analysis. The expression of genes encoding CCL18 and GAS6 was indeed upregulated in monocytes that were stimulated with CM, but not in monocytes that were stimulated with M-CSF (Fig. 5E). As expected, MMP9 gene expression was downregulated in response to stimulation with CM (Fig. 5E). The changes that were observed in phenotype, cytokine concentrations, and transcriptional profile of monocytes stimulated with CM indicate that CM not only promoted survival of monocytes but also skew monocytes towards type 2 macrophages.

MSC-Induced and Monocyte-Dependent Treg Formation Is Mediated by CCL18

Since CCL18 has been implicated in the generation of Tregs [22, 23], we hypothesized that the MSC-induced CCL18 production by monocytes was involved in the MSC-induced Treg induction that was observed in PBMC-MSC cocultures. Indeed, the culture supernatant of monocytes stimulated with CM contained measurable concentrations of CCL18. These CCL18 concentrations were significantly higher than those in culture supernatant of unstimulated monocytes, which were hardly detectable by ELISA (detection limit of the standard curve was 0.78 pg/mL) (Fig. 7A). Since addition of an inhibitor of the TGF-β1 receptor resulted in an inhibition of the MSC-induced Treg formation, we also tested whether TGF-β1 was involved in the MSC-induced secretion of CCL18 by monocytes. Addition of a TGF-β1-receptor inhibitor did not reverse the increase in CCL18 secretion in monocytes induced by CM. Moreover, addition of recombinant TGF-β1 did not result in CCL18 secretion by monocytes (Fig. 7A), indicating that the induction of CCL18 secretion by MSC is not mediated by TGF-β1.

image

Figure 7. MSC-induced and monocyte-dependent Treg formation is mediated by CCL18. (A): CCL18 protein concentrations measured with ELISA at day 3 of the monocyte culture are increased when monocytes were stimulated with CM. Addition of an inhibitor of the TGF-β1-receptor did not reverse the effect of CM and recombinant TGF-β1 did not induce CCL18 secretion by monocytes. (Data are means of five monocyte donors and two MSC donors; *, compared to control; *, p < .05; **, p < .01; #, compared to M-CSF; #, p < .05; ##, p < .01). (B): Representative dotplots of expression of CD4 and CD25 on day 14-PBMC from control cultures and from PBMC-CM cocultures in the presence of neutralizing antibodies against CCL18. (C): The percentage of CD4+CD25hi T cells at day 14 is significantly lower in the population that was cultured during the first week in the presence of aCCL18 compared to PBMC that were cultured in control PBMC-CM cultures (data are means ± SEM from two different PBMC. Statistical analysis was performed with a Student's t test; **, p < .01). (D): Hypothetical model of Treg formation induced by MSC, both the direct pathway via MSC-derived TGF-β1 and the indirect pathway via the modulation of monocytes are crucially involved in the MSC-induced Treg formation. Abbreviations: CM, conditioned medium; IL, interleukin; MSC, multipotent stromal cell; M-CSF, macrophage colony-stimulating factor; TGFβ, transforming growth factor beta.

Download figure to PowerPoint

Addition of neutralizing antibodies against CCL18 to the PBMC-CM cultures during the first week resulted in decreased MSC-induced Treg formation (Fig. 7B, 7C). These data indicate that MSC-induced CCL18 promotion by monocytes is responsible for the generation of Tregs.

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

In this study, we explored the factors and cellular interactions involved in the MSC-induced formation of CD4+CD25hiFoxP3+ Tregs derived from PBMC. We showed that MSC promote the generation of CD4+CD25hiFoxP3+ T cells with regulatory capacities, which is mediated directly by TFG-β1. In addition, this is dependent on a novel pathway for the MSC-induced generation of Tregs, for which the presence of monocytes is crucial. We showed that MSC promote monocyte survival and induce skewing of monocytes toward a cell type with an anti-inflammatory, macrophage type 2 phenotype. These cells produce IL-10 and mediate Treg formation through the production of CCL18. We hypothesize that both the TFG-β1 mediated pathway and the modulation of monocytes are essential for the MSC-induced generation of Tregs. Despite the variability in the frequency of Tregs generated by coculture of PBMC with MSC that was observed between experiments, the observation that addition of MSC or CM resulted in a higher frequency of Tregs was a robust finding.

Transwell experiments and the use of CM showed that the MSC-induced Treg formation is dependent on soluble factors secreted by MSC. In contrast to other reports [24, 25], MSC produce these factors constitutively and do not require activation. It has been reported that TGF-β1 is involved in the generation and expansion of Tregs and it is constitutively expressed and secreted by MSC [26-28]. Our finding that HO-1 is involved in the MSC-induced Treg formation confirms previous studies [29]. However, since HO-1 is not a secreted factor, it plays no role in the effects reported in this study. Here we confirm that in this in vitro culture system TGF-β1 is an essential factor involved in MSC-mediated promotion of Treg induction, indicating that TGF-β1 produced by MSC might have a direct promoting effect on the formation of Tregs from T-cells.

We did not investigate whether the generated Tregs were induced from CD4+FoxP3 T cells or whether the natural Tregs preferentially survive or proliferate. Crop et al. [14], however, showed that in this in vitro system the CD4+CD25hi fraction of the PBMC is highly proliferative during the second week of culture, indicating proliferation of already existing Tregs. An important finding of this study is the crucial role of monocytes in the MSC-induced Treg formation. Other studies showed that monocytes are important mediators for the immunosuppressive effect of MSC on T-cell proliferation [30]. Here we have identified an additional pathway and we show that the MSC-induced formation of Tregs is also mediated by monocytes, revealing an indirect route for MSC to promote Treg formation. The essential involvement of monocyte modulation for MSC-induced Treg formation in addition to the presence of TGF-β1 also explains why we did not observe Treg formation in cocultures of purified CD4+T-cells and MSC.

MSC promoted the survival of monocytes, which was confirmed by gene expression profiling of the monocytes, that showed downregulation of apoptosis-related genes in monocytes in response to CM. M-CSF promotes monocyte survival and is secreted by MSC and we showed that this factor is involved in the promotion of monocyte survival by CM.

In addition to their effect on monocyte survival, we found that MSC skew the differentiation of monocytes toward type 2 macrophages, similar to what has been suggested for fully differentiated macrophages [31]. In accordance, soluble factors derived from MSC induce upregulation of the expression of the surface markers CD206 and CD163 and downregulation of expression of the costimulatory molecule CD86. This was confirmed by gene expression profiling that showed increased gene expression of CD206, CD163, GAS6, PLTP, and CCL18. These genes are associated with the immune modulating and anti-inflammatory properties of type 2 macrophages [32-35].

The hypothesis that MSC skew monocytes toward type 2 macrophages was further supported by the observed changes in cytokine concentrations. It is well-established that type 2 macrophages produce anti-inflammatory cytokines, including IL-10 and CCL18 [22, 32, 36]. Concentrations of these cytokines were increased in supernatants of monocytes cultured in the presence of CM, and antibody neutralization experiments showed that monocyte-derived CCL18 mediated the MSC-induced generation of Tregs. We previously showed that MSC do not produce IL-10 and that soluble factors produced by MSC induce IL-10 production by monocytes [37]. The data from this study suggested M-CSF as a candidate factor for this effect. However, addition of neutralizing anti-M-CSF-R antibodies did not prevent IL-10 secretion by monocytes, indicating that other factors produced by MSC are responsible for Il-10 induction.

Antibody neutralization experiments indicated that the increase in CCL18 concentrations that was observed in monocyte cultures in the presence of CM was not dependent on TGF-β1. Since Martinez et al. [38] showed that alternative stimulation of monocytes results in upregulation of CCL18, the role of IL-4 and IL-13, cytokines that induce alternative activation of monocytes [39], would be of interest for further research on the CCL18-inducing effect of MSC.

Interestingly, type 2 macrophages have been implicated in the formation of Treg in coculture [40], which could explain the requirement for monocytes in the MSC-induced Treg formation. Conversely, Tregs are able to induce differentiation of monocytes toward type 2 macrophages [41], thus creating a loop for further amplification of Treg production.

Conclusion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

In conclusion, we hypothesize that the MSC-induced formation of Tregs relies on multiple direct and indirect mechanisms (Fig. 7D). MSC-derived TGF-β1 might act directly on the T cells to promote their differentiation toward Tregs. In addition, the modulation of monocytes is a crucial mechanism through which MSC indirectly promote Treg formation. MSC promote survival of monocytes and skew monocyte differentiation toward modulatory type 2 macrophages that secrete IL-10 and that induce Treg formation through the production of CCL18. Further studies are required to identify MSC-derived factors that are responsible for the skewing of monocytes. Our results contribute to the understanding of the cellular networks that are involved in MSC-mediated Treg formation and show a crucial role for monocytes/macrophages in MSC-induced immune modulation.

Acknowledgments

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

This work received financial support from the Netherlands Organization for Scientific Research (NWO) ZonMW Translational Adult Stem Cell (TAS) program nr 11.600.1016.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1432-sup-0001-suppfig1.tiff4917KSupporting Information Figure 1
stem1432-sup-0002-suppfig2.tiff2227KSupporting Information Figure 2
stem1432-sup-0003-suppfig3.tiff7440KSupporting Information Figure 3
stem1432-sup-0004-suppfig4.tiff3286KSupporting Information Figure 4

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.