Bone marrow (BM) mesenchymal stem cells (MSCs) are multipotent nonhematopoietic progenitor cells that can differentiate into BM stromal cells, osteoblasts, adipocytes, chondrocytes, tenocytes, skeletal myocytes, neurons, and cells of visceral mesoderm [1–6]. They grow rapidly in culture as adherent stromal colonies and typically do not express hematopoietic stem cell (HSC) markers on their surface but a quite specific pattern of molecules, such as SH2 (CD105), SH3 and SH4 (CD73), CD106 (vascular cellular adhesion molecule-1 [VCAM-1]), CD54 (intercellular adhesion molecule-1), CD44, CD90, CD29, and STRO-1 [2–6]. Initially, MSCs have been considered simply the stem cell pool that supports the BM stromal microenvironment, as well as the regeneration of other tissues of mesenchymal origin. Indeed, they interact with HSCs by influencing their homing and differentiation through either cell contact or the production of soluble factors and the chemokine CXCL12, which may attract infused HSCs to the BM through its interaction with CXCR4 . For this reason and because of their capability of long-term engraftment and in vivo differentiation [8, 9], MSCs can be cotransplanted with HSCs to improve their engraftment in BM. In addition, they can be used to repair or regenerate damaged or mutated bone, cartilage, myocardial, hepatic, or neuronal tissues and can provide a target for gene therapy strategies [3, 10–19].
In the last few years, it has become clear that MSCs also possess immunoregulatory properties. The BM microenvironment has been found to provide appropriate support for T-cell development in the absence of thymus, because in this condition most T cells adhering to BM stroma exhibit an immature phenotype [20, 21]. After HSC transplantation, BM stromal cells appear to migrate to the thymus, where they participate in the positive selection of thymocytes [22, 23]. In addition, BM stromal cells inhibit the function of mature T cells after their activation by nonspecific mitogens . In mice, BM-derived MSCs can dramatically downregulate the response of naive and memory antigen–specific T cells to their cognate peptide, and this effect is primarily cell contact–dependent . In addition, MSCs significantly prolong the survival of MHC-mismatched skin grafts after infusion in baboons and reduce the incidence of graft-versus-host disease (GvHD) after allogeneic HSC transplantation in humans [26, 27]. Finally, third-party haploidentical MSCs can be safely infused to treat severe acute GvHD refractory to conventional immunosuppressive therapy . However, the mechanisms involved in the immunoregulatory activity of MSCs on T lymphocytes are still partially obscure.
In this study, we demonstrate that human BM-derived MSCs are able to inhibit the proliferation not only of CD4+ and CD8+ T lymphocytes but also of NK cells, whereas they have no effect on the proliferation of B lymphocytes. The inhibitory effect of MSCs on the proliferation of T lymphocytes was neither related to the lack of their activation nor to the direct induction of apoptosis. The suppressive activity of MSCs was completely abrogated by the addition of an anti–interferon-γ receptor (IFN-γR) monoclonal antibody (mAb). Taken together, these findings indicate that IFN-γ produced by T lymphocytes or NK cells may promote the immunomodulatory activity of MSCs, which in turn suppress T- or NK-cell proliferation. Accordingly, in the presence of exogenous IFN-γ, B lymphocytes also became susceptible to the inhibitory activity of MSCs. The ability of IFN-γ to induce the suppressive effect of MSCs on cell proliferation appeared to be, at least in part, related to the enhancement of the indoleamine 2,3-dioxygenase (IDO) activity. These findings provide an explanation of why coinfusion of MSCs may be beneficial in the treatment of GvHD, a disorder that is primarily mediated by IFN-γ–producing type 1 T helper (Th1) cells .
Materials and Methods
Generation of MSCs
MSCs were generated from BM aspirates of healthy donors, recruited after informed consent. BM cells were obtained with density-gradient centrifugation (Lymphoprep, Nycomed Pharm, Oslo, Norway) and cultured in 25-cm2 flasks (BD Falcon, Becton Dickinson, Milan, Italy) at a concentration of 30 × 106 nucleated cells in 5 ml of Dulbecco's modified Eagle's medium, with high glucose concentration, GLUTAMAX I, 15% heat-inactivated adult bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin (all from Gibco, Milan, Italy, http://www.invitrogen.com). Cultures were incubated at 37°C in a 5% CO2 atmosphere. After 72 hours, nonadherent cells were removed. When 70%–80% adherent cells were confluent, they were trypsinized (0.05% trypsin at 37°C for 5 minutes, Gibco), harvested, and expanded in larger flasks. A homogenous cell population is normally obtained after 2 to 3 weeks of culture.
Anti-CD105 (endoglin), -CD73, -CD106 (VCAM-1), -CD44, -CD90, -CD29, -CD45, -CD14, -CD34, -CD80, -CD86, -IFN-γ-R (CD119), -CD8, -CD4, -CD25, -CD69, -CD152, -HLA (class I and II), -perforin, and -granzyme A mAbs, as well as Annexin V and BrdU detection kits, were purchased from BD Biosciences (San Diego, CA, http://www.bdbiosciences.com). Recombinant human IFN-γ, interleukin (IL)-4, IL-7, and IL-15, as well as anti–transforming growth factor (TGF)-β1 mAb, were purchased from R&D System (Minneapolis, http://www.rndsystems.com). Human rIL-2 was a kind gift from Eurocetus (Milan, Italy).
Flow Cytometric Analysis
Flow cytometric analysis was performed as detailed elsewhere . Briefly, 105 cells were incubated with the specific or the isotype control mAb at +4°C for 30 minutes; cells were then washed and analyzed on a BDLSRII cytofluorimeter using the Diva software (BD Biosciences). A total of 104 events for each sample were acquired.
Intracellular synthesis of IL-4 and IFN-γ at single-cell level was performed on polyclonally stimulated T lymphocytes, as described . To assess the expression of perforin and granzyme A, cells were fixed, washed, permeabilized, and then incubated for 15 minutes at room temperature with the specific or the isotype control mAb. Cells were then washed and analyzed on a BDLSRII cytofluorimeter.
MSC Differentiation Assay
Cell stemness was assessed by testing the ability of MSCs to differentiate into adipocytes, osteoblasts, and chondrocytes, as previously described [2, 3, 6, 9, 18, 25]. Oil red O, von Kossa, and toluidine blue dyes were used to identify adipocytes, osteoblasts, and chondrocytes, respectively. More than 90% of the cells differentiated depending on the time left in culture with the differentiating agent.
Isolation from Peripheral Blood of CD4+, CD8+, CD4+CRTH2+, CD4+CD25− T Cells, NK Cells, and B Cells
Negative selection from peripheral blood (PB) of CD4+, CD8+ T cells, and NK cells was performed by high-gradient magnetic-activated cell sorting, as described elsewhere . Briefly, PB mononuclear cell (PBMC) suspensions were treated with the CD4, CD8, and NK isolation kits II (all from Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com), the purity of the negatively selected populations consistently being >98%. In three separate experiments, CD4+ T cells were depleted of CD25+ cells by using anti-CD25 mAb-conjugated microbeads (Miltenyi Biotec). B cell isolation was performed by positive selection of CD19+ cells by using anti-CD19 mAb-conjugated microbeads (Miltenyi). The isolation of CRTH2+CD4+ T lymphocytes has been detailed elsewhere .
A total of 105 purified CD4+ or CD8+ human T cells were stimulated with 105 irradiated (9,000 rad) allogeneic T-cell–depleted PBMCs and 1 μg/ml of anti-CD3 mAb (clone HIT3a) (BD Biosciences) in the absence or presence of different numbers of MSCs (104, 103, 102, 10/well). On day 5, cultures were pulsed for 8 hours with 0.5 μCi (0.0185 MBq) of 3H-TdR (Amersham, Little Chalfont, UK, http://www.amersham.com) and harvested, and radionuclide uptake was measured by scintillation counting. In some experiments, neutralizing anti–IFN-γR (5 μg/ml), neutralizing anti-CD152 (10 μg/ml), neutralizing anti–TGF-β1 (10 μg/ml), or the same amount of isotype-matched mAbs were added to the cultures. In another series of experiments, human rIL-2 (10 IU/ml), rIL-7 (1 ng/ml), rIL-4 (2 ng/ml), or rIL-15 (7 ng/ml) was added to the cultures. In another set of experiments, two IDO inhibitors, 1-methyl tryptophan (250–1,000 μM; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) and Norharmane (125–500 μM; Sigma-Aldrich) were added to the cultures.
A total of 105 human NK cells were stimulated with 105 irradiated (9,000 rad) allogeneic T cell–depleted PBMCs and IL-2 (100 U/ml) in the absence or presence of different numbers of MSCs (104, 103,102, 10/well). On day 5, after 8 hours of 0.5 μCi 3H-TdR pulsing, cultures were harvested and radionuclide uptake was measured by scintillation counting. In some experiments, the proliferation assay was performed using different numbers of NK cells (50 × 103, 25 × 103, 12.5 × 103, 6.25 × 103/well). In other experiments, an anti-CD119 (5 μg/ml) or the same amount of an isotype-matched mAb was added to the cultures.
A total of 105 human B cells were stimulated with 105 irradiated (9,000 rad) allogeneic T cell–depleted PBMCs and 2 μg/ml of the CpG-containing DSP30F oligodeoxynucleotide (ODN) (MWG Biotech, Ebersberg, Germany, http://www.mwg-biotech.com) in the absence or presence of different concentrations of MSCs (104, 103, 102, 10/well). On day 5, after 8-hour pulse with 0.5 μCi 3H-TdR, cultures were harvested and radio-nuclide uptake was measured by scintillation counting. In some experiment, 2 ng/ml of human rIFN-γ was added at the beginning of cultures.
CarboxyFluorescein Diacetate, Succinimidyl Ester (CFSE) Labeling and BrdU Uptake
Labeling of lymphocytes with CFSE was performed as described previously . Briefly, cells were extensively washed and resuspended at a final concentration of 107 cells/ml in phosphate-buffered saline. CFSE was added at a final concentration of 5 μM and incubated for 4 minutes at room temperature. The reaction was stopped by cell washing with 10% heat-inactivated fetal calf serum– containing RPMI 1640. Cells then received the allogeneic stimulus in the presence or absence of MSCs for 5 days. BrdU was added to the medium during the last 6 hours of culture at a final concentration of 10 μM. Cells were then harvested and treated according to manufacturer's instructions (BrdU Flow Kits; BD Biosciences). Cells were then analyzed on a BDLSR cytofluorimeter (BD Biosciences) using both the FACSDiva and the ModFit LT3.0 softwares. Cell division was characterized by sequential halving of CFSE fluorescence, generating equally spaced peaks on a logarithmic scale. Seven individual peaks and corresponding regions were identified (G0 –G6). The symbols related to each peak (G0 –G6) were indicated, the undivided T cells (G0) residing in the rightmost peak and T cells that have divided six times (G6) residing in the leftmost peak.
Transwell experiments were performed in 24-well transwell plates (0.22-μm pore size, Costar, Corning, NY). A total of 5 × 105 CFSE-labeled CD4+ T lymphocytes were stimulated with 5 × 105 irradiated T cell–depleted allogeneic PBMCs and 1 μg/ml of anti-CD3 mAb in the absence or presence of 5 × 104 MSCs, placed in the same or in another chamber. On day 4, after 6 hours of pulsing with BrdU, cells were harvested and evaluated by flow cytometry for CFSE and BrdU contents. In additional experiments, an anti-CD119 or an isotype-matched mAb (5 μg/ml) was added to the cultures.
Evaluation of Apoptosis and NK Cytotoxicity
The evaluation of apoptosis in CD4+ and CD8+ T cell populations was performed using the annexin V kit (BD Biosciences) following the manufacturer's instructions. Cells were then analyzed on a BDLSRII cytofluorimeter (BD Biosciences) using the FACSDiva software.
To evaluate the effects of MSCs on the cytotoxic activity of NK cells, these cells were purified and cultured in the absence or presence of MSCs (ratio, 10:1) for 5 days. On day 5, NK cells were collected and tested for their ability to kill K562 target cells by using flow cytometry. Briefly, K562 target cells were cultured for 5 hours in the absence or presence of different numbers of untreated or MSC-precultured NK cells (ratio, 1:1, 1:2, and 1:4). Cells were then collected, stained with anti-CD56 mAb and annexin V, as detailed elsewhere , and finally analyzed on a BDLSRII cytofluorimeter using the FACSDiva software. The cytolitic potential of NK cells was evaluated by gating on CD56-negative cells (K562 target cells) and looking at the percentage of annexin V–positive cells, as described elsewhere .
Real-Time Quantitative Reverse Transcription–Polymerase Chain Reaction (TaqMan)
Foxp3 quantitation was performed using Assay on Demand (Applied BioSystems, Foster City, CA, http://www.appliedbiosystems.com) as described elsewhere .
Quantitation of IFN-γ Concentration in Supernatants
For the quantitation of IFN-γ in cell supernatants, a commercially available ELISA kit was used (R&D System), according to the manufacturer's instructions.
Statistical comparison of the proliferation assay arms was carried out according to the Student's t-test. Differences were considered statistically significant with p < .05.
MSCs Inhibit the Proliferative Response of T and NK but Not B Cells
Human MSCs were obtained from BM aspirates of healthy volunteers, appropriately cultured, and purified to homogeneity. Figure 1 shows the immunophenotype and the differentiative potential of MSCs used in our experimental system. Constitutive expression of CD105, CD73, CD29, CD44, CD90, CD106, HLA class I, and CD119, but not of CD80, CD86, CD45, and HLA-DR, was observed (Fig. 1A). The same cells exhibited multilineage differentiation potential, as assessed by culturing in adipogenic, osteogenic, or chondrogenic medium (Fig. 1B).
The activity of increasing numbers of MSCs on the proliferative response to allogeneic stimulation of total T lymphocytes purified from PB of healthy donors was then evaluated. As shown in Figure 2A, the addition of MSCs in cultures of activated T lymphocytes significantly inhibited their proliferation only at 1:10 ratio (104 MSCs vs. 105 T lymphocytes), whereas at lower ratios it was ineffective. To evaluate the kinetics of T cell proliferation in the absence or presence of MSCs, we performed parallel experiments in which the proliferation tracer CFDA-SE and the uptake of the thymidine analogue BrdU at different time points of cell cultures were contemporaneously evaluated. As shown in Figures 2B and 2C, during the first 3 days of culture, T-cell proliferation was not affected by the presence of MSCs. The suppressive effects of MSCs became significant on day 4 and even more on day 5 (Figs. 2B, 2C). As shown in Figure 2D, after 5 days of culture in the presence of MSCs, there was a significant accumulation of T cells in generation (G) 1, G2, and G3, with a significant decrease in G4, G5, and G6.
To establish whether MSCs can exert their inhibitory effect on both CD4+ and CD8+ T lymphocytes, the proliferation in response to allogeneic stimulation of CD4+ or CD8+ T cells, purified from PB of the same donors and cultured in absence or presence of increasing numbers of MSCs, was assessed. Again, the proliferation of both CD4+ and CD8+ T lymphocytes was inhibited only when 104 MSCs were added to 105 CD4+ or CD8+ T cells. Of note, CD8+ T cells seemed to be more susceptible than CD4+ lymphocytes to the MSC suppression (p < .005 vs. p < .05) (supplemental online Fig. 1). Furthermore, the inhibitory activity of MSCs, as evaluated on day 5, resulted in a significant accumulation of both CD4+ and CD8+ T cell subsets in the first generations (Figs. 3A, 3B).
We then asked whether MSCs were also able to suppress the proliferation and the functional activities of NK and B cells. To this end, purified CD56+ cells (purity > 98%) were stimulated with allogeneic cells in the presence of rIL-2. As shown in Figure 3C, MSCs significantly suppressed the NK cell proliferation. In addition, measurement of IFN-γ levels in culture supernatants revealed a significant decrease of this cytokine when NK cells were cultured in the presence of MSCs (Fig. 3D). More importantly, the ability of NK cells to lyse K562 target cells seemed to be significantly impaired, at least at the 4:1 (NK:K562) cell ratio, when NK cells were precultured for 5 days in the presence of MSCs (Fig. 3E). Finally, to test the effect of MSCs on B cell proliferation, CD19+ cells purified from PB (purity > 98%) were stimulated with the DSP30F CpG-containing ODN plus allogeneic T-cell–depleted PBMCs. As shown in Figure 3F, MSCs did not exhibit any suppressive effect on B cell proliferation.
Inhibitory Effect of MSCs Is Not Contact-Dependent and Is Not Due to Their Interference on Cell Activation, Induction of Apoptosis, or Involvement of T Regulatory Cells
The mechanisms possibly responsible for the inhibitory effects of MCSs on human T-cell proliferation were then investigated. To evaluate whether the suppressive activity of MSCs was mediated by cell contact or via the release of soluble factors, CFDA-SE–labeled CD4+ cells were stimulated with allogeneic cells in the absence or presence of MSCs (10:1 cell ratio), which were added in the same well of a Transwell plate, separated or not by a porous septum. The inhibitory effect of MSCs on CD4+ T-cell proliferation was not affected by the physical separation of the two populations, as shown by both CFDA-SE and BrdU stainings, suggesting that the inhibitory effect of MSCs was not mediated by cell contact but was dependent on the release of soluble factors (Figs. 4A, 4B).
In previous reports [32, 34], we have shown that the suppressive activity of human CD4+CD25+ and CD8+CD25+ thymocytes on the proliferation of autologous CD4+CD25− thymocytes was contact dependent and could be related to the combined activity of CTLA-4 and TGF-β1 expressed on the surface of activated Treg cells. Moreover, such a regulatory effect was rescued by the addition in culture of cytokines able to stimulate the proliferation of T lymphocytes, such as IL-2, IL-4, IL-7, or IL-15 [30, 32, 34]. However, a mixture of the same anti–CTLA-4 and anti–TGF-β1 mAbs that had been found able to revert the suppressive effect of Treg cells showed no effect on the suppressive activity of MSCs (Fig. 5A). Moreover, none of the above-mentioned cytokines, at least at the concentration tested, was able to rescue T cells from the suppressive activity exerted by MSCs (Fig. 5B).
We also tried to establish whether the inhibitory effect of MSCs was due to their interference with cell activation processes. To this end, the expression of activation markers, such as CD69, CD25, and CTLA-4, by both CD4+ and CD8+ T-cell subsets stimulated with allogeneic cells in the absence or presence of MSCs was assessed. No significant difference in CD69, CD25, or CTLA-4 expression by CD4+ and CD8+ T cells was detected at any time (6, 24, 48 hours) (Fig. 5C). To exclude the possibility that the inhibition of T-cell proliferation induced by MSCs was due to cell death, the presence of apoptotic cells on day 5 was evaluated by flow cytometry in cocultures of MSCs/ CD4+ or MSCs/CD8+ cells, cultured at a 1:10 cell ratio. No significant differences in the proportion of annexin V–positive cells was observed in both CD4+ and CD8+ T cell populations (Fig. 5D).
Finally, to evaluate a possible involvement of CD4+CD25+ Treg cells in the inhibitory activity of MSCs, we tested the ability of MSCs to suppress the proliferation induced by allogeneic stimulation of CD4+CD25− T cells. As shown in Figure 5E, MSCs exerted the same suppressive effects on the proliferation of either total CD4+ or CD4+CD25− T cells. In addition, the evaluation on day 5 of mRNA for Foxp3, a marker highly expressed by CD4+CD25+ Treg cells (13,083 ± 3,727 fgs of Foxp3 cDNA/50,000 cells, n = 3), showed low levels of expression on total CD4+ T cells cultured in the absence or presence of MSCs. More importantly, no significant differences in Foxp3 mRNA levels between CD4+ T cells cultured under the two conditions were detected (Fig. 5F).
Inhibitory Effect of MSCs on Cell Proliferation Requires the Presence of IFN-γ, Which, at Least in Part, Acts by Enhancing Their IDO Activity
In the course of this study, we were intrigued by the observation that MSCs could inhibit the proliferation of either T or NK cells, whereas they did not exhibit any effect on B cell proliferation. Therefore, the possibility that some factors produced by both T and NK cells but not by B cells may allow, or even trigger, the suppressive activity exerted by MSCs was hypothesized. We focused our attention on IFN-γ because this cytokine represents one of the factors produced by several CD4+ and CD8+ T lymphocytes, as well as by NK cells, but not by B lymphocytes. In addition, IFN-γ can potentially modify the MSC immunological phenotype, as it is able to upregulate the expression of HLA class I molecules and to induce de novo expression of HLA class II molecules. To provide evidence on the possible role of IFN-γ, the suppressive activity of MSCs toward CD4+ and CD8+ T cells in the absence or presence of a neutralizing anti– IFN-γR mAb was assessed. The addition in culture of the anti– IFN-γR mAb completely reverted the suppressive effect exerted by MSCs on CD4+ and at least partially even on CD8+ T cells, as assessed by both thymidine uptake (Fig. 6A) and CFSE-BrdU assays (data not shown). In addition, the capacity of the anti– IFN-γR mAb to revert the suppressive activity exerted by MSCs on NK cell proliferation was achieved only when the number of NK cells in culture was reduced, in agreement with the notion that NK cells are able to produce high IFN-γ concentrations (Fig. 6B). More importantly, the addition of the anti– IFN-γR mAb in the condition in which MSCs and CD4+ T cells were separated by the porous septum in the Transwell system completely restored the proliferation of the CD4+ T cells (Fig. 6C).
To further support the role of IFN-γ on the suppressive activity of MSCs, two additional experiments were performed. First, the suppressive activity of MSCs on the proliferation of B lymphocytes (which are unable to produce this cytokine) in the absence or presence of exogenously added IFN-γ was assessed. As shown in Figure 6D, the addition in culture of IFN-γ allowed MSCs to suppress even the proliferation of B lymphocytes. Second, purified PB CD4+ T lymphocytes were separated into CRTH2+ and CRTH2− fractions, a cell marker that has been previously associated with Th2 cells [31, 35], and then assessed for their responsiveness to allogeneic stimulation in the absence or presence of MSCs. As expected, purified CRTH2+CD4+ T cells produced IL-4, but not IFN-γ, whereas CRTH2−CD4+ T cells produced mainly IFN-γ (supplemental online Fig. 2). As shown in Figure 6E, only the proliferation of total CD4+, as well as of CD4+CRTH2−, T cells was inhibited in a dose-dependent fashion by MSCs, whereas the proliferation of CD4+CRTH2+ cells was not. Of note, even if the proliferation of CD4+CRTH2+ cells was lower than the proliferation of total CD4+ or CD4+CRTH2− T cells, the ability of all T cell populations to respond to allogeneic stimulation was similar, the stimulation index (cpm stimulated − cpm unstimulated/cpm unstimulated cells) of CD4+CRTH2+, total CD4+, and CD4+CRTH2− T cells being equal to 90.5 ± 19.6, 106.0 ± 26.3, and 129.7 ± 29.8, respectively.
Finally, the mechanisms by which IFN-γ can trigger, or contribute to, the immunosuppressive activity of MSCs were investigated. It has recently been demonstrated that human MSCs express IDO, which catalyzes the conversion from tryptophan to kynurenine, and whose expression is upregulated by IFN-γ . To this end, the suppressive activity of MSCs on the proliferative response of T cells was assessed in the presence of two different competitive inhibitors of the IDO pathway, 1-methyl tryptophan and norharmane [37, 38]. As shown in Figure 7, both these molecules exerted a partial, but consistent, inhibition of the suppressive activity of MSCs on T-cell proliferation.
BM transplantation is largely used for the treatment of several pathological conditions, despite the lack of suitable BM donors, representing a limit to further expand this kind of therapeutic approach. The donor-recipient histoincompatibility is associated with a high risk of both graft rejection and GvHD, situations in which strategies to diminish immune responses after transplantation are mandatory. Recently, a great interest in MSCs contained in BM has been developed, inasmuch as several reports suggest that these cells not only have multipotential differentiation ability  but seem to be capable also of modulating immune responses, both in vitro and in vivo [24–28, 39–41]. More importantly, MSCs are able to prolong skin graft survival in nonhuman primates, and their coinfusion with HSCs has been found to reduce the incidence of GvHD in patients affected by hematologic malignancies [26, 27]. Thus, the infusion of MSCs in conjunction with the donor organ or BM might provide a useful tool for favoring the engraftment and to reduce the incidence and/or the intensity of GvHD. However, the development of this novel therapeutic strategy requires that the mechanisms involved in the immunosuppressive activity of MSCs are better clarified.
In the present study, we evaluated the activity of human MSCs on the proliferation of different lymphocyte populations and also tried to provide further information on the mechanisms possibly involved in their suppressive effect. To address these points, we tested the activity exerted by MSCs obtained from human BM on the proliferation of T lymphocytes, both CD4+ and CD8+, NK cells, and B lymphocytes. MSCs were able to inhibit in a dose-dependent fashion the proliferative response of CD4+ and CD8+ T lymphocytes, as well as of NK cells, whereas they did not exert any inhibitory effect on B-cell proliferation. The evaluation of the proliferation tracer CFDA-SE and the uptake of the thymidine analogue BrdU on allogeneic-stimulated T cells at different time points of cell cultures revealed that the proliferation of T cells began to be suppressed by MSCs only after 3 days. These data indicate that to trigger the suppressive activity of MSCs, a crosstalk between MSCs and target cells is needed during the first phase of culture, an event that cannot occur with B lymphocytes. Moreover, we clearly showed that NK cells precultured for 5 days with MSCs were partially inhibited in their ability to lyse the K562 target cells. Why these data are apparently at variance with those reported by Rasmusson et al.  is unclear. One possible explanation may be that in this study the ability of MSCs to suppress NK cell–mediated lysis of K562 cells was directly assessed during the 4 hours of MSC/NK cell coculture of the cytolytic assay, which may be a time not sufficient to reveal the suppressive effect.
Because the suppressive activity of MSCs seems to be limited to T lymphocytes and NK cells, as that exerted by natural CD4+CD25+ Treg cells [32, 34, 42], we asked whether the same mechanisms were involved in this phenomenon. Some reports have suggested that natural Treg cells act through cell-to-cell contact and that TGF-β1 and/or CTLA-4 may be responsible for the suppressive activity of CD4+CD25+ Treg cells [32, 34, 43]. However, by using a double-chamber culture system in which MSCs and target cells are separated by a semipermeable membrane that allows the diffusion of molecules in the absence of cell contact, we clearly demonstrated, in accordance with previous reports [24, 40, 44, 45], that the suppressive activity of MSCs did not depend on cell contact. In addition, as already reported [24, 44, 46], neutralization of both TGF-β1 and CTLA-4 did not reverse the suppressive capacity of MSCs. It has been shown recently that MSCs can exert their suppressive activity by inducing an increase in the numbers of CD4+CD25+ Treg cells within the target population . To test this possibility, the expression of Foxp3 mRNA, CD25, and CTLA-4 molecules by T lymphocytes cultured in the presence of MSCs, as well as the ability of these latter to suppress the proliferation of CD4+CD25− T cells induced by allogeneic stimulation, was evaluated. After allogeneic stimulation, low levels of Foxp3 mRNA were found independently of whether CD4+ T cells were cultured in the absence or presence of MSCs. Furthermore, high CD25 expression was found in T cells, but in the presence of MSCs there was no significant difference in the expression of either CD25 or CTLA-4 compared with T cells stimulated in the absence of MSCs. More importantly, no significant differences were observed in the suppressive activity of MSCs on the proliferation induced by allogeneic stimulation of either total CD4+ T cells or CD25-depleted CD4+ T cells. In addition, the proliferation of T cells could not be rescued by T-cell growth factors, such as IL-2, IL-4, IL-7, and IL-15, at least at the concentration used, which had been found to be able to overcome the suppressive activity of Treg cells . Thus, at least in this experimental system, MSCs seem to exert their suppressive activity through mechanism(s) different from those described for natural Treg cells [32, 34, 43]. We also asked whether the inhibition of T cell proliferation induced by MSCs could result from their interference on T-cell activation or the induction of apoptosis. However, neither inhibition of activation markers such as CD25 and CD69 nor induction of apoptotic processes could be observed. The results that CD25 and CD69 expression on allostimulated T cells is not influenced by the presence in culture of MSCs are in accordance with those recently reported in other studies [45, 48].
To provide further information on the mechanisms involved in the suppressive activity of MSCs, we took advantage of two orders of observations. First, the suppressive activity of MSCs did not require cell-to-cell contact, suggesting the possibility that it was mediated through some soluble factors. Second, the suppressive effect of MSCs seemed to be stronger on CD8+ rather than on CD4+ T lymphocytes, and even more on NK cells, whereas it did not affect the proliferation of B lymphocytes. The latter finding allowed us to hypothesize that some mediators actively produced upon stimulation by both NK and T, but not by B, lymphocytes may trigger, or at least contribute to, the suppressive activity of MSCs. An obvious candidate for this activity was IFN-γ, which is produced at high concentrations by NK and CD8+ T cells and at lower concentrations by CD4+ T cells but is not produced by B cells. In agreement with this possibility, the addition in culture of a neutralizing anti– IFN-γR mAb consistently inhibited the suppressive activity of MSCs on both CD4+ and CD8+ T cells and even on NK cells, provided that lower numbers of these latter cells were tested in culture. This finding is consistent with the observation that NK cells are able to produce extraordinarily high amounts of IFN-γ, which can, at least partially, bypass the mAb-mediated blocking of the IFN-γR. The crucial role of IFN-γ in MSC-driven suppression was further supported by another series of experiments. First, MSCs were able to suppress the proliferation of PB CD4+CRTH2− T cells, which contain a mixture of cells, including those able to produce IFN-γ, whereas they had no inhibitory effect on the proliferation of purified CD4+CRTH2+ T cells, which represent a pure population of Th2 cells able to produce IL-4 but not IFN-γ . Finally, and most importantly, the addition in culture of exogenous IFN-γ also made the proliferative response of B lymphocytes susceptible to the inhibitory activity of MSCs.
The mechanisms by which IFN-γ can trigger, or contribute to, the immunosuppressive activity of MSCs were finally investigated. It has recently been demonstrated that human MSCs express IDO, which catalyzes the conversion from tryptophan to kynurenine and whose expression is upregulated by IFN-γ . Moreover IFN-γ–induced IDO expression by professional antigen-presenting cells has been identified as a major immunosuppressive effector pathway for the inhibition of T-cell proliferative responses [49–51]. In this study, by using competitive inhibitors of IDO activity, we could observe a partial but substantial inhibition of the suppressive activity exerted by MSCs on T-cell proliferation. Thus, it is highly likely that even if it does not represent the unique mechanism that activates suppression, IFN-γ contributes to promote IDO expression in MSCs, which in turn suppresses the proliferative response of effector cells through the tryptophan pathway. Consistent with these observations, the involvement of PGE2 in the suppressive activity exerted by human MSCs has recently been demonstrated [45, 47]. Since neither the competitive inhibitors of IDO activity nor PGE2 inhibitors are able to completely abrogate MSC-mediated immunosuppressive effects, it is reasonable to hypothesize that both pathways are involved in this phenomenon. Accordingly, inhibition of IFN-γ, which has been shown to induce both IDO and PGE2 [36, 47], results in a complete abrogation of the MSC suppressive activity.
The demonstration that human MSCs can interact with HLA-unrelated immune cells modulating their proliferative response and that IFN-γ is crucial to evoke this capacity may have important implications in transplantation biology. Acute GvHD is a Th1-mediated condition  in which NK and CD8+ T cells also play an important role. However, the results of this study provide evidence that the most important cytokine produced by these cells, i.e., IFN-γ, plays an important role in activating the immunomodulatory effects of MSCs, thus favoring, in turn, the suppression of GvHD itself. Based on these findings, it is likely that the coadministration of MSCs with the transplant could be useful in controlling both graft rejection and GvHD. However, additional studies are needed to clarify the cell contact signals by which MSCs exert their immunosuppressive activity on the proliferation of several cell types, including Th1 cells. Moreover, the finding that MSCs can even suppress the proliferation of B cells in the presence of exogenously added IFN-γ supports the possibility that lymphoproliferative disorders or autoimmune diseases, which are characterized by strong B cell activation, also could benefit from therapeutic approaches based on the use of allogeneic or autologous MSCs.
M.K. and L.C. contributed equally to this work. This work was supported by Italian Ministry of University and Scientific Research, Italian Association for Cancer Research (AIRC), Italian National Research Council (CNR), Fondazione Cariverona, and the Ministry of Health of Tuscany Region.
The authors indicate no potential conflicts of interest.