Mesenchymal Stem Cells Inhibit Generation and Function of Both CD34⁺-Derived and Monocyte-Derived Dendritic Cells¹

Mesenchymal stem cells (MSCs)³ are multipotential nonhemopoietic progenitor cells capable of differentiating into multiple lineages of the mesenchyme (1). Although the frequency of MSCs is low in human tissues, MSCs can be isolated and expanded as undifferentiated cells in vitro from adult bone marrow, adipose tissue and fetal liver, blood, bone marrow, lung, and cord blood (2). Phenotypically, MSCs are characterized by the absence of hemopoietic and endothelial markers (CD45, CD34, CD31) and the expression of various other markers (such as CD90, CD73, CD105, CD106, CD44). Furthermore, MSCs secrete a number of cytokines, growth factors, and matrix molecules that may play important roles in the proliferation and maturation of hemopoietic stem cells. These properties of MSCs make them of substantial clinical interest.

Studies performed in human and animal models have demonstrated that infused MSCs are capable of long-term engraftment and in vivo differentiation into several mesenchymal tissues (3) and encouraging results have been reported for ex vivo-expanded MSCs in early clinical use (4). Furthermore, the clinical interest in MSCs is enhanced by the observation that MSCs display immunosuppressive activities by inhibiting T cell proliferation to polyclonal mitogens and alloantigens in vitro (5, 6, 7, 8). A previous report demonstrated that in vivo i.v. administration of MSCs prolonged allogeneic skin graft survival in baboons with an efficiency comparable to immunosuppressive agents currently being used clinically (9). Furthermore, it has been demonstrated that coinjection of murine MSCs promoted the outgrowth of allogeneic tumor cells in mouse recipients (10). A recent case report has suggested that systemic infusion of MSCs resulted in suppression of severe treatment-resistant acute graft-vs-host disease after allogeneic bone marrow transplantation (11). The mechanisms underlying the immunosuppressive effect of MSCs are surrounded by controversy and remain to be clarified. MSC-mediated suppression of T cells is thought to be mediated by soluble factors (5, 8, 10, 12) and by mechanisms that require cell contact (6, 8, 10, 12, 13). Moreover, MSCs can arrest T cells in the early G₁ phase of the cell cycle and irreversibly inhibit cell proliferation (14). The immunosuppressive effects of MSCs have been generally ascribed to the inhibitory effect on T cells, but recently it was also shown that MSCs have inhibitory effects on other immune cells, including dendritic cells (DCs) (15, 16).

DCs are the most potent APCs specialized in the uptake, transport, and presentation of Ags and have the unique capacity to stimulate naive and memory T cells (17). In addition, DCs can interact with B cells (18) and NK cells (19). In vitro, human DCs can be grown from monocytes in the presence of GM-CSF and IL-4 (20), and from CD34⁺ bone marrow stem cells in medium supplemented with GM-CSF and TNF-α (21). DCs play a key role in the initiation of primary immune responses and in tolerance, depending on the activation and maturation stage of DCs (22). Immature DCs behave as sentinels in peripheral tissues, with high ability in Ag uptake and processing, and low ability in T cell stimulation. Locally produced inflammatory cytokines or microbial components promote the maturation from a processing to a presenting stage, characterized by up-regulation of MHC and costimulatory molecules, production of IL-12, and migration to lymphoid tissue. DC maturation is a prerequisite to induce immunogenic T cell responses, whereas tolerance is observed when Ags are presented by immature or semimature DCs (23). Therefore, DCs serve as potential targets for suppression of alloimmune reactivity and promotion of tolerance induction.

Although the effects of MSCs on T cells have been extensively studied, the effects of MSCs on DCs are relatively unknown (24). In the present study, we have investigated the effects of MSCs on the differentiation of both CD34⁺-progenitors and monocytes into DCs with respect to phenotype and function. The present study demonstrates that MSCs inhibit the generation and maturation of DCs and, in consistence, impairs the capacity to stimulate T cell proliferation.

CD34⁺ hemopoietic progenitor cells were isolated from umbilical cord blood samples to generate immature DCs as previously described (21). In brief, CD34⁺ cells were isolated from mononuclear fractions through positive selection using anti-CD34-coated microbeads and Midi-Macs separation columns (Miltenyi Biotec) to a purity of 90–96% as determined by flow cytometry. Cells were cultured in RPMI 1640 containing 5% FCS, penicillin/streptomycin (P/S), 2 mM l-glutamine, 50 μM 2-ME, 100 ng/ml GM-CSF, 25 ng/ml stem cell factor (SCF; Amgen), 2.5 ng/ml TNF-α (ITK Diagnostics), and 5% AB⁺ pooled human serum. After 6 days, cells were harvested and further cultured in the absence of AB⁺ serum but in the presence of GM-CSF and TNF-α for 6–8 additional days.

In some experiments, cells were collected at day 6 and labeled with FITC-conjugated anti-CD1a (clone HI149; BD Pharmingen) and PE-conjugated anti-CD14 (clone MP9; BD Biosciences). Cells were separated according to CD1a and CD14 expression into CD1a⁺CD14⁻ and CD1a⁻CD14⁺ fractions using a FACS Starplus (BD Biosciences). Sorted cells (purity of 95–98%) were further cultured in the presence of GM-CSF and TNF-α for 6–8 additional days.

Human PBMC were isolated from buffy coats obtained from healthy donors using Ficoll-Hypaque (Sigma-Aldrich). Monocytes were purified from PBMC by using the MACS monocyte Isolation kit (Miltenyi Biotec), and DCs were generated by culturing monocytes in RPMI 1640 containing 10% FCS, P/S, GM-CSF (5 ng/ml; Novartis), and IL-4 (10 ng/ml; Schering-Plough) for at least 6 days.

Maturation of monocyte-derived cells was induced with 100 ng/ml LPS (Salmonella typhosa; Sigma-Aldrich) or CD40L. CD40L activation was performed with CD40L-transfected L cells (25) at a DC:L cell ratio of 4:1. Nontransfected L cells served as control cells. Cells and supernatants were analyzed after 48 h.

The Medical Ethics Review Board of the Leiden University Medical Center approved the protocol for collecting fetal tissues for research purposes (01/116-E). Fetal tissues were obtained after informed consent from women undergoing elective termination of pregnancy between 15 and 22 wk of gestation. Fetal lung and fetal bone marrow MSCs were obtained as previously described (26). In brief, single-cell suspensions of fetal bone marrow and lung were cultured in M199 (Invitrogen Life Technologies) supplemented with 10% FCS, P/S, endothelial cell growth factor (20 μg/ml; Roche Diagnostics) and heparin (8 U/ml) in tissue culture flasks coated with 1% gelatin. When grown to confluency, adherent cells were detached with trypsin/EDTA. Data presented were performed with irradiated (60 Gy) fetal MSCs although similar results were obtained with adult MSCs.

Adult bone marrow was obtained after informed consent from harvests of healthy donors for allogeneic transplantation. Mononuclear cells were isolated using Ficoll-Hypaque density gradient centrifugation and plated at 2 × 10⁵/cm² in DMEM-low glucose (Invitrogen Life Technologies) supplemented with 10% FCS and P/S. Medium was replaced, and the nonadherent cells were removed after 2 days of initial culture and every 3 or 4 days thereafter. When grown to confluency, adherent cells were detached with trypsin/EDTA and reseeded for expansion. A homogenous population was obtained after 4 wk of culture. The MSCs expanded in culture showed positive surface staining for CD105, CD90, and CD44, and were negative for CD14, CD31, CD34, and CD45 surface expression. The expanded MSCs retained the capacity to differentiate into adipogenic, osteogenic, and chondrogenic lineages (data not shown).

The phenotype of the cells was analyzed by flow cytometry by means of a FACSCalibur cytometer using CellQuest software (BD Biosciences). Cells were stained with FITC-conjugated or PE-conjugated mAbs against the following surface markers: CD1a, CD14, CD40 (Serotec), CD80 (Immunotech), CD83 (Sanbio), CD86 (BD Pharmingen), and HLA-DR (BD Biosciences).

CD34-derived cells cultured for 12–14 days were collected, and, after irradiation (30 Gy), used as stimulator cells for allogeneic naive CD4⁺ T cells (5 × 10⁴/well) purified from umbilical cord blood mononuclear cells by using the CD4 isolation kit (Miltenyi Biotec).

After 7 days of culture, monocyte-derived cells were collected, activated with CD40L for 2 days as described above, and after irradiation (30 Gy), used as stimulator cells for allogeneic T cells (1 × 10⁵/well) isolated from PBMC by negative selection using the pan T cell isolation kit (Miltenyi Biotec).

Stimulator cells were added in graded doses to the T cells in 96-well flat-bottom plates in RPMI 1640 containing 10% FCS. After 3 (monocyte-derived DCs) or 5 (CD34-derived DCs) days of incubation, cells were pulsed with [³H]thymidine (1 μCi/well) for the last 18 h, harvested, and counted using a Topcount NXT (Canberra Packard). Results are expressed in cpm and presented as means (± SD) obtained from triplicate cultures.

Day 10 CD1a- and CD14-derived DCs were activated with CD40L with CD40L-transfected L cells. Monocyte-derived DCs were activated at day 7 of culture. After 48 h, cytokines were detected in cell-free culture supernatants by ELISA. The measurement of IL-6 was performed as described previously (27). The analysis of IL-10 (Sanquin Reagents), IL-12 (U-Cytech), TNF-α (Sanquin Reagents), and IFN-γ (Sanquin Reagents) was performed according to the instructions of the manufacturer.

Statistical analysis was performed by the Student t test or two-way ANOVA using GraphPad PRISM (GraphPad Software). Differences were considered statistically significant when p values were <0.05.

The effect of MSCs on the differentiation of DCs derived from CD34⁺ progenitor cells was studied by culturing CD34⁺ cells in the presence of GM-CSF, TNF-α, and SCF with or without MSCs. In response to GM-CSF and TNF-α, CD34⁺ progenitor cells differentiate along two unrelated DC pathways into Langerhans cells, which differentiate directly into CD1a⁺ DCs, and into dermal/interstitial DCs, which differentiate via an intermediate CD14⁺CD1a⁻ phenotype into CD14⁻CD1a⁺ DCs (21). The presence of MSCs during differentiation did not affect cell yield of CD34⁺-derived DCs. The addition of MSCs during the proliferation phase (days 0–6) stimulated the differentiation of CD34⁺ cells to CD14⁺CD1a precursor cells and was associated with an increase of CD14 expression intensity (Fig. 1,A). The continuous presence of MSCs (days 0–12; Fig. 1 A) or the addition of MSCs during the maturation phase (days 6–12) (data not shown) almost completely prevented the differentiation of CD14⁺CD1a⁻ into CD14⁻CD1a⁺ DCs. In addition to the phenotypical modifications, the morphology of DCs generated in the presence of MSCs was also affected, showing more round cells with numerous vacuoles (data not shown). The suppressive effect was observed with different donors of fetal and adult MSCs at different passages (P2-P9; data not shown). However, fetal MSCs from one donor lost the capacity to inhibit the differentiation after at a very early passage (P4).

The inhibitory effect of MSCs on differentiation of DCs from CD34⁺ progenitor cells was dose dependent, with an inhibitory effect ranging from a MSC:CD34⁺ cells ratio of 1:1 to 1:300 (Fig. 1 B).

During the maturation phase, DC precursors differentiate into DCs in terms of morphology and phenotype (Fig. 2), as described previously (21). However, in the presence of MSCs, the expression of typical DC Ags, including CD80, CD86, CD83, HLA-DR, and CD40, was strongly reduced (Fig. 2).

Considering the inhibitory effect of MSCs on the expression of costimulatory molecules on DCs, it was of interest to examine the functional activity of DCs with regard to their ability to induce proliferation of allogeneic naive T cells. DCs generated in the presence of MSCs were significantly impaired in their ability to induce allogeneic T cell proliferation compared with cells generated in the absence of MSCs (Fig. 3).

To examine the effect of MSCs on the two different pathways of DC development, CD14⁺ precursors and CD1a⁺ precursors were FACSorted at day 6 of the cultures. As shown in Fig. 4,A, CD14⁺ precursors cultured for an additional 3 days in GM-CSF and TNF-α spontaneously differentiated into CD1a⁺ DCs. Consistent with the results obtained from the bulk population, culturing the CD14⁺ sorted cells in the presence of MSCs completely prevented the differentiation into CD1a⁺ DCs (Fig. 4,A). MSCs did not affect the development of Langerhans cells. In line with the inhibitory effect of MSCs on the differentiation of CD34⁺-derived CD14⁺ precursors into CD1a⁺ DCs, the effect of MSCs on the differentiation of monocytes into DCs was similar. After 7 days of culture, nontreated cells differentiated into nonadherent DCs characterized by protruding veils, abundant cytoplasm, and by the acquisition of CD1a and the loss of CD14 Ags (Fig. 4,A). In the presence of MSCs, cultured monocytes displayed a macrophage morphology, round and lacking veiled appearance (data not shown). Cells cultured in the presence of MSCs retained high CD14⁺ without acquisition of CD1a (Fig. 4,A). The presence of MSCs did not affect the cell yield and viability. Kinetic experiments with monocyte-derived DCs revealed that addition at days 2 and 4 partially blocked DC generation. Only when MSCs were added at day 0 was DC development completely prevented (Fig. 4,B). Importantly, when MSCs were removed after 2 days, CD14 expression was lost but the cells did not acquire expression of CD1a, suggesting that the suppressive effect of MSCs was not completely reversible (Fig. 4 B).

Next, the phenotype of CD14⁺- and CD1a⁺-derived DCs was analyzed. Both CD14⁺ precursors and CD1a⁺ precursors acquired expression of characteristic DC Ags (Fig. 5). Culturing the sorted CD14⁺ precursors in the presence of MSCs resulted in a decreased expression of CD80 and CD86 and a moderate suppressed expression level of CD83 and HLA-DR (Fig. 5). The phenotype of CD1a⁺ precursor cells did not differ between control and MSC cultures (Fig. 5). Similarly, the response of monocyte-derived DCs generated in the presence of MSCs to maturation signals was investigated. Activation of immature DCs by CD40L resulted in DC maturation as shown by an induction of CD83 expression and increased expression of costimulatory molecules CD80, CD86, and HLA-DR (Fig. 5). However, DCs generated in the presence of MSCs did not show expression of CD83 or up-regulation of costimulatory molecules in response to CD40L (Fig. 5). In addition, LPS and TNF-α were unable to activate DCs generated in the presence of MSCs, whereas maturation was induced in control DCs (data not shown).

Because cytokines produced by DCs are especially important in determining subsequent T cell responses, supernatants of nonstimulated or CD40L-activated DCs were assessed for the production of IL-12, IL-10, and TNF-α. Activation of DCs with CD40L, derived from CD34⁺ progenitors or from monocytes, strongly increased the production of IL-12 and IL-10 (Fig. 6). Monocyte-derived DCs generated in the presence of MSCs showed a decrease in CD40L-induced IL-12, IL-10 (Fig. 6), and TNF-α production (1368 ± 239 pg/ml vs 240 ± 83 pg/ml in the absence and presence of MSCs, respectively, p < 0.05). In contrast, the production of IL-12 and IL-10 by CD40L-activated CD14⁺-derived DCs was not affected by the presence of MSCs (Fig. 6), although MSCs inhibited the spontaneous secretion of IL-12 by unstimulated CD14⁺-derived DCs. These results suggest that activation with CD40L antagonizes the MSC-mediated inhibition on the function of CD14⁺-derived DCs. Moreover, the presence of MSCs did not impair the production of IL-12 by CD1a⁺-derived DCs. Consistent with previous publications, CD1a⁺-derived DCs did not produce IL-10 (28).

Because DCs derived from CD34-derived CD14⁺ precursors and peripheral blood monocytes generated in the presence of MSCs display reduced expression of MHC class II and costimulatory molecules, the T cell stimulatory capacity of DCs cultured in the presence of MSCs was examined. CD1a⁺-derived DCs generated in the presence of MSCs induced T cell proliferation equally as well as control CD1a⁺-derived DCs (Fig. 7). Furthermore, the secretion of IFN-γ by T cells stimulated with both CD1a⁺-derived DC populations was similar (Fig. 7). In contrast, the capacity of DCs derived from CD14⁺ precursors and monocytes to stimulate proliferation of T cells was strongly reduced upon culturing in the presence of MSCs (Fig. 7). In addition, the secretion of IFN-γ by T cells stimulated with these cells was decreased (Fig. 7).

To investigate whether the inhibitory effect of MSCs on the differentiation of DCs was mediated by soluble factors, transwell culture systems were used with monocyte-derived DCs. MSCs physically separated from monocytes suppressed the generation of DCs to the same extent as that observed when MSCs were in direct contact with the monocytes (Fig. 8,A), thus suggesting the involvement of soluble factors. Supernatants from MSCs cultured in the presence of monocytes suppressed the generation of DCs, but the inhibitory effect was more variable and less profound (Fig. 8,A) than that observed in the transwell experiments. In contrast, supernatants from MSCs did not affect the differentiation of monocytes to DCs (data not shown). Earlier studies have demonstrated that IL-6 and M-CSF can inhibit the differentiation of DCs (29, 30). To examine the involvement of IL-6 and M-CSF that are known to be produced by MSCs, neutralizing Abs against IL-6 and M-CSF were examined for their capacity to restore DC differentiation. As shown in Fig. 8 B, saturating concentrations of anti-IL-6 and anti-M-CSF Abs partially reversed the phenotypic modifications induced by MSCs; although all the cells lost CD14 expression, CD1a expression was not completely restored. These data suggest that IL-6 and M-CSF are involved and that other factors also play a role in the inhibition of DC differentiation by MSCs.

Several studies have indicated that MSCs are able to modulate immune responses, through mechanisms that are largely unknown. Although the effects of MSCs on T lymphocytes have been extensively studied, their effects on the initiators of immune responses, the DCs, are relatively unknown. In the present study, we demonstrate that MSCs prevent differentiation of both CD34⁺ progenitors and monocytes into DCs and can inhibit their final maturation and function.

CD34⁺-derived CD14⁺ precursors and peripheral blood monocytes represent a large pool of circulating precursors that can differentiate into macrophages, the scavengers of the immune system, or DCs, the most potent APCs that can activate T cells and initiate immune responses. The differentiation is dependent on environmental factors and it has been demonstrated that transmigration through endothelial layers can induce differentiation toward DCs (17, 22, 31), while exposure to IL-6 and IL-10 shifts differentiation from DCs to macrophages (29, 30, 32, 33). In line with previous observations with fibroblasts (30) and epithelial cell lines (29), our study is the first to show that MSCs block the generation of interstitial/dermal DCs by inhibiting the differentiation of CD34⁺-derived CD14⁺ precursors into CD1a⁺ DCs, while the generation of Langerhans cells is not affected. Moreover, MSCs similarly inhibit the differentiation of monocytes into DCs, as shown previously (16, 24). CD34⁺-derived CD14⁺ precursors and monocytes cultured in the presence of MSCs were found to have characteristic features of macrophages in terms of morphology and phenotype with persistent CD14 expression and decreased CD1a expression.

Previous studies demonstrated that MSCs have a moderate effect on already differentiated immature and mature DCs (16, 24). Importantly, our results demonstrate that DCs generated in the presence of MSCs showed impaired responses to signals inducing maturation, as demonstrated by the absence of CD83 costimulatory molecules and HLA-DR up-regulation. In agreement with these phenotypical changes, DCs generated in the presence of MSCs were strongly hampered in their capacity to induce activation of T cells. However, MSCs only suppress the production of cytokines by CD40L-activated monocyte-derived DCs and not by CD14⁺-derived DCs, indicating that functional differences exist between the two DC populations.

The mechanism underlying the inhibitory effect of MSC on DC differentiation and function remains to be elucidated. Transwell experiments and the addition of conditioned medium from MSCs cocultured with monocytes indicate the involvement of soluble factors. The production of these soluble factors most likely requires intercellular contact between MSCs and monocytes, because no inhibitory effect was observed with supernatant from MSCs alone. MSCs produce several cytokines, including IL-6, M-CSF, and IL-10, which may be important in DC differentiation. Although neutralizing Abs against IL-6 and M-CSF reduced the expression of CD14, the acquisition of CD1a was not completely restored, suggesting that other soluble factors produced by MSCs may cooperate with IL-6 and M-CSF to inhibit DC differentiation. Of interest is the observation that MSCs specifically inhibit the differentiation of CD34⁺-derived CD14⁺ precursors without affecting the generation of Langerhans cells from CD1a⁺ precursors. The former represent bipotent cells that can be induced to differentiate into macrophage-like cells in response to M-CSF, in contrast to the latter that do not respond to M-CSF, presumably because of absence of M-CSF receptor expression (34).

DCs and their precursors show remarkable plasticity and the generation of APCs with distinct differentiation and functional stages is tightly controlled by the microenvironment. It has been proposed that the capacity to induce tolerance or immunity is dependent on the maturation and activation state of DCs. Mature DCs, which are able to secrete IL-12, and express high levels of costimulatory molecules, can efficiently induce potent immune responses, whereas immature DCs, with low expression of costimulatory molecules, are poor stimulators of Th1-type responses and can lead to T cell anergy or deletion (35) or the development of regulatory T cells (36) that maintain tolerance. Thus, the microenvironment plays an important role in the regulation of immune responses. MSCs are one of the main components of the marrow microenvironment and the data presented in this study suggest that MSCs may contribute to a cellular microenvironment that influences the differentiation and function of DCs, which in turn may shape T cell responses. However, the influence of MSCs on the differentiation and function of DCs might be limited because of the low frequency of MSCs in the bone marrow (37).

Previous studies have demonstrated that MSCs can inhibit T cell responses. In vivo, infusion of MSCs leads to prolonged skin allograft survival and may decrease graft-vs-host disease. Whether the reported MSC-induced alterations of DC differentiation and function play any role in the immunosuppressive effects remains to be established. Of note, recently another immunoregulatory mechanism for MSCs is proposed wherein MSCs inhibit T cells by induction of regulatory APCs (38). Furthermore, MSCs can suppress the inflammatory response of various cells of the immune system (15), although it has been demonstrated recently that the immunosuppressive properties of MSCs can be reversed by TNF-α (39). Taken together, the complex interplay between MSCs and subsets of immunocompetent cells may provide an explanation for the in vivo MSC-mediated induction of tolerance.

In conclusion, our data show that, in addition to a direct effect on T lymphocytes, MSCs suppress the development and maturation of both interstitial/dermal DCs and monocyte-derived DCs without affecting the development and function of Langerhans cells and thereby reveal a new mechanism by which MSCs may alter the outcome of the immune response.

We thank Drs. A. M. Woltman and C. van Kooten for their helpful discussion. Furthermore, the midwives of the obstetrician practice de Kern, Leiden, The Netherlands, and the Cord Blood Bank are acknowledged for the collection of cord blood samples.

The authors have no financial conflict of interest.