Cell contact, prostaglandin E₂ and transforming growth factor beta 1 play non-redundant roles in human mesenchymal stem cell induction of CD4⁺CD25^Highforkhead box P3⁺ regulatory T cells

Cell isolation, purification and culture

Human MSC were isolated and expanded from aspirates of bone marrow by direct plating, as described previously [2,34]. Cultures were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal calf serum and 1% (v/v) penicillin/streptomycin (Invitrogen-Gibco, Paisley, Scotland, UK). Contamination in MSC populations is a key confounding variable, therefore rigorous quality control was adopted to ensure that MSC were not contaminated with haematopoietic or other cell types, and that cells retained differentiation capacity as described previously [2,34]. Human CD4⁺ T cells were isolated from peripheral blood mononuclear cells (PBMC) using a MagCellect CD4⁺ T cell negative selection, isolation kit according to the manufacturer's instructions (R&D Systems, Abingdon, Oxon, UK). CD4⁺ T cells were co-cultured with MSC at a ratio of 3:1 (CD4⁺ : MSC), a ratio determined to be the optimal dose for MSC modulation of T cells and DC. Similarly, for transwell experiments, MSC were cultured in the tissue culture insert incorporating a 0·4 µm membrane placed in the wells of six- or 24-well plates (Greiner Bio-one, Stonehouse, Glos, UK), while CD4⁺ T cells were cultured in the well beneath the insert. Study design and use of human mesenchymal stem cells was approved by the bioethics committees of the National University of Ireland Maynooth and the National University of Ireland Galway.

Characterization of human MSC surface marker expression and differentiation potential

Human MSC were detached from substrate with trypsin/ethylenediamine tetraacetic acid (EDTA), washed and resuspended in phosphate-buffered saline (PBS) with 1% (v/v) bovine serum albumin. Cells (1 × 10⁵) were incubated with fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-conjugated monoclonal antibodies for 15 min at 4°C, washed and then analysed by flow cytometry. Antibodies recognizing the following human antigens or appropriate isotype controls were used: human leucocyte antigen (HLA)-ABC, HLA-DR, CD106, CD117 (eBioscience, San Diego, CA, USA), CD11b, CD29, CD31, CD34, CD44, CD45, CD54, CD105 (Immunotools, Friesoythe, Germany) and CD90 (BD Biosciences, Oxford, UK).

Human MSC were examined for their ability to differentiate into adipocytes, osteoblasts and chondrocytes. Adipogenic differentiation was induced by 1 µM dexamethasone, 200 µM indomethacin, 10 µg insulin and 500 µM 3-isobutyl-methyl-xanthine, and osteogenic differentiation was induced by 100 nM dexamethasone, 10 mM β-glycerolphosphate and 50 µM ascorbic acid-2-phosphate. For chondrogenic differentiation a pellet culture system was used. Some 2 × 10⁵ cells were centrifuged in a 15 ml polypropylene tube. The pellet was cultured at 37°C in 5% CO₂ in 500 µl chondrogenic media containing high-glucose DMEM supplemented with 10 ng/ml human TGF-B3 (TS:beta) (R&D Systems), 100 nM dexamethasone, 50 ug/ml ascorbic acid-2-phosphate, 40 ug/ml proline, 1 mM sodium pyruvate and (1:99) ITS + supplement (BD Biosciences). Differentiation cultures were harvested after 21 days. Oil red O and Alizarin red S were used to identify adipocytes and osteoblasts respectively. Chondrogenic pellets were harvested on day 21 for glycosaminoglycan (GAG) quantitation. Briefly, chondrogenic pellets were washed with Dulbecco's PBS and digested with papain solution (1 mg/ml in 50 mM sodium phosphate, pH 6·5 containing 2 mM N-acetyl cysteine and 2 mM EDTA) for 16 h at 65°C. GAG was measured by reaction with 1,9-dimethylmethylene blue using shark chondroitin sulphate as standard. DNA quantitation was carried out using the PicoGreen® dsDNA Quantitation Kit (Invitrogen) with phage lambda DNA as standard as described previously [35]. Data was expressed as a ratio of GAG/DNA.

Analysis of FoxP3 and TGF-β1 mRNA by reverse transcription–polymerase chain reaction (RT–PCR) and real-time PCR

Characterization of FoxP3 and TGF-β1 mRNA expression by CD4⁺ T cells was performed using semi-quantitative and quantitative (qRT–PCR). Purified CD4⁺ T cells (3 × 10⁵/ml) and MSC (1 × 10⁵/ml) were co-cultured in 24-well plates (Nunc, Roskilde, Denmark). In additional studies, anti-TGF-β (R&D Systems) or indomethacin (Sigma-Aldrich, Dublin, Ireland) were added to cultures at concentrations of 4 µg/ml and 40 µM respectively. After 24 h, non-adherent CD4⁺ T cells were removed by aspiration leaving the adherent MSC monolayer untouched using a simple method described previously [15]. Purification of total RNA from CD4⁺ T cell cultures was performed using the RNeasy® Plus Mini kit (Qiagen, Crawley, UK) according to manufacturer's instructions, reverse-transcribed to cDNA and analysed for the expression of human FoxP3 and TGF-β1 by a hot-start RT–PCR reaction (Promega, Southampton, UK). Glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) was detected as an internal control. Primer sequences were as follows: GAPDH forward 5′-GGTGAAGGTCGGAGTCAACG-3′ and reverse 5′-CAAAGTTGTCATGGATGACC-3′, with an annealing temperature of 55°C; TGF-β1, annealing at 54°C: forward 5′-CAGATCCTGTCCAAGCTG-3′; reverse 5′- TCGGAGCTCTGATGTGTT-3′; and FoxP3 annealing at 57°C: forward 5′-AGGTGGCAGGATGGTTTCT-3′; reverse 5′-AACAGCACATTCCCAGAGTTC-3′. Quantitative real-time PCR analysis was performed using a QuantiTect™ SYBR Green PCR Kit (Qiagen) with a DNA Opticon™ quantitative thermocycler (MJ Research, Waltham, MA, USA) [20].

Standard curves for each target gene were generated by amplifying 10-fold serial dilutions of known quantities of target gene PCR product standards. Quantification of target gene expression was obtained using sequence detector system software (MJ Research, Inc.). In this study the resultant target mRNA concentrations were expressed as fg/500 ng cDNA.

Analysis of CD25 and FoxP3 protein expression by flow cytometry

CD4⁺ T cells or PBMC were co-cultured with MSC in 2 ml volumes in six-well plates (Nunc) or in transwells for 72 h as described above. CD4⁺ T cells or PBMC were incubated at 0·5 × 10⁶ cells/ml and MSC at 1·5 × 10⁵/ml. Cells were labelled for surface CD4 and CD25 and intracellular FoxP3. Briefly, cells were incubated with CD4-PE or CD4-PE/Cy5 and CD25-PE or CD25-FITC or isotype-matched control antibodies (eBioscience) for 15 min at 4°C then washed and fixed by 2% (v/v) paraformaldehyde/PBS. Cells were permeabilized using cold PBS and 0·2% (v/v) Tween20/PBS then blocked with normal rat serum. After blocking, cells were incubated with fluorochrome-conjugated anti-human FoxP3 (eBioscience) or isotype control for at least 30 min at 4°C in the dark. At the end of the incubation period, cells were washed and resuspended in PBS containing 1% v/v formaldehyde. Analysis was performed within 4 h of preparation using BD fluorescence activated cell sorter (FACS)Calibur and CellQuest software (BD Biosciences).

CD4⁺CD25⁺ functional assay

After 72 h co-culture of purified CD4⁺ cells and MSC, CD4⁺ T cells were removed by aspiration and CD4⁺CD25⁺ cells isolated using CD25 microbeads according to the manufacturer's instructions (Miltenyi Biotech, Bisley, Surrey, UK). CD4⁺CD25⁺ T cells were evaluated for their ability to suppress allodriven proliferation in mixed lymphocyte reaction (MLR) by co-culturing PBMC from two MHC mismatched donors in 96-well plates, as described previously [20]. Proliferation was measured using [³H]-thymidine incorporation for the final 6 h of culture represented as the incorporated radioactivity in counts per minute (cpm). Results are expressed as the mean of triplicate values ± standard error (s.e.).

Statistical methods

Statistical significance was assessed using Prism3 software (GraphPad Software, San Diego, CA, USA). Paired data were analysed by paired t-test and three or more data sets were compared using one-way anova, with Tukey's multiple comparison with measure significance. Data are presented as the mean ± s.e. P-values of P < 0·05 (*), P < 0·01 (**) or P < 0·001 (***) were considered statistically significant.

Characterization of human bone marrow-derived MSC

Human MSC derived from bone marrow displayed a fibroblastoid morphology and expressed surface markers typical of adult MSC [36,37], including HLA-ABC, CD44, CD90 and CD106 (Fig. 1a). Human MSC did not express HLA-DR, the haematopoietic cell markers CD34, CD117 or the co-stimulatory molecules CD40, CD80 or CD86 (Fig. 1a). Human MSC did not express CD29, CD54, CD105 or CD154 (Fig. 1a). Human MSC had the ability to differentiate along the chondrogenic, osteoblastic or adipocytic pathways after culture in the appropriate conditions (Fig. 1b and c).

Allogeneic MSC induce human CD4⁺CD25^HighFoxP3⁺ T cells

Previous studies have suggested that MSC induce T cells with a regulatory phenotype when co-cultured with PBMC. However, these data are difficult to interpret, as multiple cell types (CD8⁺ T cells, etc.) are present in such systems. Therefore, this study sought to investigate whether allogeneic MSC induced T cells with suppressive/regulatory activity using defined purified populations of human CD4⁺ T cells. MSC were co-cultured with MHC mismatched CD4⁺ T cells and qRT–PCR used to quantify expression of the T_reg transcription factor, FoxP3. CD4⁺ T cells were cultured in the presence or absence of MSC for 24 h. FoxP3 expression could be observed in CD4⁺ T cells, cultured in the presence or absence of MSC (Fig. 2a); however, there was a significant increase (**P < 0·01) of FoxP3 in purified CD4⁺ T cells cultured previously in the presence of allogeneic MSC compared with CD4⁺ T cells cultured alone (Fig. 2b).

mRNA concentration does not always correlate with protein levels; therefore, FoxP3 and CD25 protein expression were determined by flow cytometry. Purified CD4⁺ T cells were cultured in the presence or absence of MSC for 72 h. In the absence of MSC, small numbers of CD4⁺ T cells expressed FoxP3 (Fig. 2c) or CD25 (Fig. 2c). In contrast, there was a consistent increase in FoxP3 and CD25^High expression after co-culture with MSC (Fig. 2c). Thus MSC induce human CD4⁺CD25^HighFoxP3⁺ expression in allogeneic T cells.

Allogeneic MSC induction of FoxP3 and CD25^High expression by human T cells involves a cell contact-dependent mechanism

In other systems, MSC-mediated immunomodulation involves both cell contact-dependent and -independent mechanisms mediated through the release of soluble factors. In order to probe the role of cell contact, transwell experiments were performed that prevented direct contact between purified CD4⁺ T cells and allogeneic MSC. FoxP3 mRNA expression was examined after 24 h using qRT–PCR. Separation of MSC from CD4⁺ T cells prevented the induction of FoxP3 mRNA (Fig. 3a). To investigate the correlation at the protein level, CD4⁺ cells were recovered from the lower chamber after 72 h of culture, and both FoxP3 and CD25 expression were analysed. When cell–cell contact was prevented, allogeneic MSC did not induce FoxP3⁺ CD25⁺ expression in purified CD4⁺ cells (Fig. 3b). Consistently fewer CD4⁺ T cells expressed FoxP3 and CD25^High when compared with cultures where cell contact was permitted (Fig. 3a and b). In separate experiments, the importance of other immune cells in MSC induction of FoxP3 expression was examined. Unseparated PBMC were cultured in the presence or absence of MSC in regular cultures or in a transwell system that prevented cell contact. PBMC were recovered after 72 h and FoxP3 and CD25 expression analysed by flow cytometry. Although co-culture of PBMC with MSC resulted in increased expression of FoxP3 and CD25, prevention of cell contact did not interfere with MSC induction of FoxP3 and CD25, indicating that in the presence of other immune cells, cell contact is not required (Fig. 3c). Taken together, these data show that cell–cell contact contributes to MSC-driven T_reg induction in purified CD4⁺ T cells, but that in a multi-cell system such as PBMC additional factors can substitute for cell contact-derived signals.

The TGF-β1 and PGE₂ play a non-redundant role in MSC induction of human CD4⁺ CD25^High FoxP3⁺ T cells

The observed role for cell contact in T_reg induction in CD4⁺ T cells does not exclude the possibility that soluble factors play a non-redundant role in T_reg induction, and indeed the data in Fig. 3c support an important role for soluble factors in MSC induction of T_reg. In other systems, PGE₂ contributes to the induction of FoxP3 [33], and we have demonstrated previously that PGE₂ secreted by human MSC was required for modulation of allodriven proliferation in MLR [20]. TGF-β1 is also known to be involved in T_reg induction [38], and moreover is up-regulated in MSC ‘licensed’ by IFN-γ stimulation [20]. Therefore, the effects of blocking both factors in allogeneic MSC : purified CD4⁺ T cell co-cultures were examined. Cyclooxygenase activity and therefore PGE₂ induction was ablated using the antagonist indomethacin, whereas TGF-β1 was neutralized by a well-characterized antibody. The presence of indomethacin or anti-TGF-β1 alone resulted in a significant decrease in FoxP3 mRNA expression by CD4⁺ T cells cultured with allogeneic MSC (Fig. 4a). However, a combination of both reduced FoxP3 mRNA to background levels (Fig. 4a). Neutralization of TGF-β resulted in decreased protein expression of both intracellular FoxP3 and surface CD25^High by purified CD4⁺ cells following co-culture (Fig. 4b and c) and antagonism of PGE₂ production produced a similar effect (Fig. 4b and c). This indicates that although cell contact is involved in MSC induction of human CD4⁺ CD25^High FoxP3⁺ T cells, both TGF-β1 and PGE₂ play non-redundant roles in this process. In effect, both cell contact and soluble factors are required for MSC induction of CD4⁺ T_reg cells. The transwell experiments show that cell contact is required and soluble factors are not sufficient to achieve this. The neutralization study which allowed cell contact demonstrated that TGF-β1 and PGE₂ are the effector molecules involved in the induction of T_reg in the absence of other immune cells.

Human CD4⁺CD25⁺ T cells induced by allogeneic MSC act as conventional T_reg

Blockade of cytokine production from CD4⁺ T cell/MSC co-cultures is a straightforward technique to determine a role for cytokines in T_reg induction; however, this method does not provide information of the functional activity of the induced T cell. Therefore CD4⁺ T cells, cultured in previously the presence or absence of MSC for 24 h, were re-isolated from the adherent monolayer of MSC. TGF-β1 mRNA was significantly higher in CD4⁺ T cells that had been co-cultured with MSC (Fig. 5a and b), supporting hypotheses that MSC-derived signals condition the local tissue environment to promote induction of T cells secreting this immunosuppressive cytokine [16,39]. Nevertheless, the induction of CD4⁺ cells secreting TGF-β1 is insufficient to describe these cells as T_regs. Similarly, the induced expression of CD25 does not necessarily define these as T_regs, because CD25 is also expressed by recently activated T cells. To ascertain whether MSC generate functional T_regs, CD4⁺CD25⁺ T cells induced by allogeneic MSC were purified and examined for their suppressive activity in MLR. After 72 h MSC co-culture, CD4⁺CD25⁺ cells were retrieved using CD25⁺ microbeads and were added subsequently to MLR containing PBMC from mismatched donors. CD4⁺CD25⁺ cells induced by allogeneic MSC secreted TGF-β1 (Fig. 5a and b) and reduced allodriven proliferation significantly (Fig. 5c). Thus CD4⁺CD25⁺ T cells induced by allogenic MSC displayed conventional suppressive activity consistent with definitions of T_regs.