SEARCH

SEARCH BY CITATION

Keywords:

  • Mesenchymal stem cells;
  • Immunosuppression;
  • Indoleamine 2,3-dioxygenase;
  • Nitric oxide;
  • Chemokine

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

Bone marrow-derived mesenchymal stem cells (MSCs) hold great promise for treating immune disorders because of their immunoregulatory capacity, but the mechanism remains controversial. As we show here, the mechanism of MSC-mediated immunosuppression varies among different species. Immunosuppression by human- or monkey-derived MSCs is mediated by indoleamine 2,3-dioxygenase (IDO), whereas mouse MSCs utilize nitric oxide, under the same culture conditions. When the expression of IDO and inducible nitric oxide synthase (iNOS) were examined in human and mouse MSCs after stimulation with their respective inflammatory cytokines, we found that human MSCs expressed extremely high levels of IDO, and very low levels of iNOS, whereas mouse MSCs expressed abundant iNOS and very little IDO. Immunosuppression by human MSCs was not intrinsic, but was induced by inflammatory cytokines and was chemokine-dependent, as it is in mouse. These findings provide critical information about the immunosuppression of MSCs and for better application of MSCs in treating immune disorders. STEM CELLS 2009;27:1954–1962


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

Bone marrow-derived mesenchymal stem cells (MSCs) are strongly immunosuppressive and have been demonstrated to be highly effective in suppressing transplant rejection, autoimmunity, and other immune disorders in animals and humans [1–6]. The mechanism of this immunosuppression remains controversial, however, as many different immunosuppressive mediators are suspected, including interleukin-10 (IL-10) [7], transforming growth factor-β (TGF-β) [8], prostaglandin E2 [9], nitric oxide (NO) [6, 10], indoleamine 2,3-dioxygenase (IDO) [11], hepatocyte growth factor (HGF) [8], and regulatory T cells [12]. These discrepancies may result from the different isolation methods, culture conditions, and assays used, but the most important factor may be interspecies variation, as many different species have been utilized in studies of MSC-mediated immunosuppression. Therefore, a better understanding of MSC-mediated immunosuppression in humans that distinguishes it from other mammalian species is critical for proper clinical application of this unique stem cell population.

We have recently reported that mouse bone marrow-derived MSCs are strongly immunosuppressive both in vitro and in vivo, and this effect depends on inflammatory cytokines and NO [6]. In these mouse models, we have shown that inflammatory cytokines induce significant production of NO, which then directly suppresses proliferation and cytokine production by lymphocytes. NO is a labile, bioactive, rapidly diffusing gaseous molecule [13]. Recent studies have shown that NO and NO-derived reactive nitrogen species can affect a broad array of enzymes, ion channels, and receptors [14, 15]. It is produced through the action of nitric oxide synthases (NOS), of which there are three species encoded by the human and mouse genomes: iNOS, found in macrophages and other cell types; nNOS, found in neurons; and eNOS, found in endothelial cells. The levels of nNOS and eNOS are relatively steady, whereas iNOS expression is inducible and plays a major role in immune regulation. Although the role of NO in the immune system is well established, it recently has been shown to also affect T cell receptor (TCR) signaling, cytokine receptor expression, and the phenotypes of T cells [16]. Importantly, since NO is highly unstable, it only acts locally; thus, it is conceivable that immune cells would need to be recruited into close proximity with MSCs to be affected by NO produced by MSCs. From our previous studies, we have concluded that chemotaxis is a critical component of NO-mediated immunosuppression by mouse MSCs [6].

The mouse system has been one of the most useful models to elucidate the molecular mechanisms of immunosuppression by MSCs. But findings derived from the murine system must be extendable to humans to be medically relevant. Therefore, we investigated the mechanism of immunosuppression by human MSCs. We found that, unlike that for mouse MSCs, human MSC-mediated inhibition of T-cell responses could not be reversed by NO synthase inhibitors. And, conversely, although inhibition of IDO had no effect on immunosuppression by mouse MSCs, it completely reversed the effect of human MSCs, indicating that mouse MSCs and human MSCs utilize different effector molecules in suppressing immune reactions.

IDO can exert immunosuppressive effects in a number of settings, such as maternal tolerance of the fetus, immunoprivilege of the cornea, and immunosuppression inside tumors [17, 18]. IDO catalyzes the rate-limiting step in the degradation of tryptophan, an essential amino acid, along the kynurenine pathway. It is believed that the resulting reduction in local tryptophan concentration and the production of immunomodulatory tryptophan metabolites contribute to the immunosuppressive effects of IDO-expressing cells. Depletion of tryptophan could lead to an increase in the kinase activity of GCN2, which induces the expression of several genes that shut down cell proliferation [19, 20]. Some tryptophan metabolites suppress T-cell proliferation in vitro or induce T-cell apoptosis. A number of these metabolites are immunosuppressive when administered in vivo. They may act through the orphan G protein-coupled receptor, GPR35, which is highly expressed on immune cells [21]. Thus, immunosuppression by IDO may be derived from its ability to deplete tryptophan and produce immune regulatory kynurenine metabolites. It seems probable that both the GCN2 pathway and the metabolite pathways function synergistically to culminate the full biologic effects of IDO.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

Cells

Human MSCs were generated in house from human bone marrow following the approved protocol by the Institutional Review Board of Robert Wood Johnson Medical School. Monkey MSCs were derived in house from bone marrow obtained from rhesus monkeys following the protocol by the Institutional Animal Care and Use Committee at Kunming Zoology Institute. Briefly, human or monkey bone marrow aspirates were cultured in MesenPRO RS medium (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Nonadherent cells were removed after 24 hours, and adherent cells were maintained with medium replenishment every 3 days. All such MSCs were used before the seventh passage. Primary mouse MSCs were derived as previously reported [6]. All MSCs (before the seventh passage) were examined for free of hematopoietic stem (progenitor) cell and lineage cell surface makers. The “stemness” of all MSCs was determined by their capabilities to differentiate into adipocytes, osteoblasts, and chondrocytes and by their expression of specific cell surface markers (data not shown).

CD4+ lymphocytes were purified from fresh human peripheral blood mononuclear cells (PBMCs) (The Blood Center of New Jersey, East Orange, NJ) by magnetic-activated cell sorting (Miltenyi Biotec, Auburn, CA, http://www.miltenyibiotec.com). For T-cell blasts, T cells (1 × 106 cells per milliliter) were activated by plastic-bound OKT3 antibody for 72 hours and then cultured with IL-2 (200 U/ml) alone for 48 hours. All T-cell cultures were maintained in RPMI-1640 medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μM β-mercaptoethanol (complete medium).

Supernatant from activated human PBMC (“Sup”) was harvested for 72 hours after stimulation of PBMC (2 × 106/ml) by plastic-bound OKT3, 0.2-μm filtered, and frozen at −80°C.

Reagents

Recombinant mouse and human interferon-γ (IFNγ), tumor necrosis factor-α (TNFα), and IL-1α and monoclonal antibody against human IL-10 were from eBiosciences (La Jolla, CA). Monoclonal antibodies against human IFNγ and human TGF-β were purchased from PBL Biomedical Laboratories (New Brunswick, NJ) and Abcam Inc. (Cambridge, U.K., http://www.abcam.com), respectively. Anti-human CXCR3 was from R&D Systems Inc. (Minneapolis, http://www.rndsystems.com). Monoclonal antibodies against human IDO and mouse iNOS were from Millipore (Billerica, MA) and BD Bioscience (San Jose, CA). Indomethacin, 1-methyl-DL-tryptophan (1-MT), and NG-monomethyl-L-arginine (L-NMMA) were from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com).

Proliferation Assay

PBMCs (1 × 105) were cultured in 100 μl of complete medium in 96-well plates for the indicated times. To assay de novo cell proliferation, 0.5 μCi of 3H-thymidine (Tdr; GE Healthcare Bio-Sciences Corporation, Piscataway, NJ, http://www.gelifesciences.com) was added to each well 8 hour before termination of the cultures by freezing. Plates were then thawed, harvested, and incorporated 3H-Tdr was counted using a Wallac Microbeta scintillation counter (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com). Cell proliferation was also analyzed in cells stained with carboxyfluorescein diacetate succinimidyl ester (CFSE): freshly isolated PBMCs (106/ml in phosphate-buffered saline [PBS]) were labeled with 5 μM CFSE for 8 minutes at room temperature. The labeling was then terminated by adding 2% fetal calf serum in PBS. After washing, cells were cultured with MSCs and cell division, as evidenced by halving of the fluorescence intensity, was analyzed by flow cytometry.

Real-Time Polymerase Chain Reaction

Real-time polymerase chain reaction (PCR) was conducted as per a previous protocol using a RNeasy Mini Kit [22]. Briefly, the first-strand cDNA synthesis was performed using a Sensiscript RT Kit with random hexamer primers (all kits from Qiagen, Valencia, CA, http://www1.qiagen.com). mRNA of the genes of interest were quantitated by real-time PCR (MX-4000 from Stratagene, La Jolla, CA, http://www.stratagene.com) using SYBR Green Master Mix (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Total amount of mRNA was normalized to endogenous GAPDH mRNA. Primer sequences are shown in Table 1.

Table 1. Primers used for real-time polymerase chain reaction
inline image

Chemotaxis Assay

Chemotaxis was tested with the NeuroProbe ChemoTx Chemotaxis System (NeuroProbe, Gaithersburg, MD), as described previously [6, 23]. The lower chambers of the 96-well plate were filled with supernatant from MSCs stimulated with IFNγ, IFNγ + TNFα (20 ng/ml each), or OKT3-activated PBMC supernatant (Sup) (1:2 dilution). An upper chamber with a polyvinylpyrrolidine-free polycarbonate membrane with 5-μm pores was then inserted in the well, and T-cell blasts (1.25 × 105) were added to the top chambers. After a 3-hour incubation, cells that had migrated through pores and into bottom wells were quantitated using MTT assay. A chemotaxis index was calculated as the ratio of the number of T-cell blasts migrated in response to MSCs compared to the number migrating in response to medium alone.

The immunosuppression resulting from T-cell migration toward inflammatory cytokine-activated MSCs was examined in a similar setup. MSCs (2 × 104) were added to the lower chamber with or without stimulation with IFNγ and TNFα (20 ng/ml each) for 24 hours. Activated T-cell blasts were then added to the upper chamber, as above, and IL-2 was added to both chambers. After 1 hour, both chambers were pulsed with 3H-thymidine and cell proliferation was assessed 3 hours later.

Western Blot Analysis

Protein samples were diluted in Laemmli buffer (62.5 mM Tris-HCl, pH 6.9, 2% SDS, 1% β-mercaptoethanol, 10% glycerol, and 0.04% bromphenol blue) and separated on a 10% SDS-polyacrylamide gel. Proteins were then electroblotted onto a nitrocellulose membrane (Roche Molecular Biochemicals, Laval, QC) and revealed using monoclonal antibodies against iNOS, IDO, or β-actin and chemiluminescent detection (Amersham ECL, GE Healthcare, Piscataway, NJ, http://www.gelifesciences.com) according to the manufacturer's instructions.

Statistical Analysis

Significance was assessed by unpaired two-tailed Student's t test or analysis of variance (ANOVA). ∗∗, p < .01; ∗∗∗, p < .001.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

Nitric Oxide Plays a Critical Role in Immunosuppression Mediated by MSCs Derived from Mouse, but Not from Human

We recently reported that immunosuppression by mouse MSCs is achieved through the concerted action of nitric oxide and T-cell specific chemokines, both of which are induced by inflammatory cytokines in MSCs. These chemokines drove T-cell migration into the proximity of MSCs, where T-cell responsiveness was suppressed by high concentration of nitric oxide in the activated MSCs niche [1, 6]. To establish the clinical relevance of these findings, the mechanism of immunosuppression mediated by mouse MSCs was compared to that of human MSCs. Therefore, we derived MSCs from human bone marrow. When tested for their immunosuppressive capability, human MSCs strongly inhibited TCR-triggered proliferation of freshly isolated human PBMCs (Fig. 1A). Similar to that of mouse MSCs [6], this immunosuppressive effect largely depended on IFNγ since anti-IFNγ reversed the inhibition of T-cell activation by human MSCs (Fig. 1A). Unlike that in mouse MSCs, however, the immunosuppressive effect of human MSCs was unaffected by inhibition of NO production by L-NMMA, an effective iNOS inhibitor (Fig. 1A). Other well-known factors were also ruled out by using blocking antibodies or inhibitors against TGFβ, IL-10, and COX2 (Fig. 1A). In contrast, an inhibitor of IDO, 1-methyl-tryptophan (1-MT), effectively blocked immunosuppression by human MSCs (Fig. 1A). Therefore, whereas immunosuppression mediated by human MSCs (like mouse) requires IFNγ, the mediator is IDO instead of NO. This species variation in the mechanism of MSC-mediated immunosuppression may explain, at least in part, previous conflicting reports regarding the role of NO and IDO in MSC-mediated immunosuppression.

thumbnail image

Figure 1. Immunosuppression by human MSCs is mediated by indoleamine 2,3-dioxygenase (IDO), not nitric oxide. (A): Human MSCs (passage five) were co-cultured with fresh human PBMCs at a ratio of 1:10 (MSC/PBMC) and stimulated by anti-human CD3 (OKT3, 1 μg/ml). Antibodies against human IFNγ, TGFβ, and IL-10 (20 μg/ml each), nitric oxide synthase inhibitor L-NMMA (1 mM), prostaglandin E2 inhibitor indomethacin (10 μM), or IDO inhibitor 1-MT (0.5 mM) were added as indicated. Cell proliferation was assessed by 3H-thymidine incorporation after 72 hours. Values represent mean ± SD of five wells from a representative of four experiments using different human MSCs. (B): Human MSCs (passage five) were treated with recombinant human IFNγ, alone or in combination with TNFα (20 ng/ml each) for 24 hours, and then co-cultured with human T-cell blasts (1:10 ratio) with or without 1-MT (0.5 mM). Cell proliferation was assessed after an additional 8 hours. Values represent mean ± SD of five wells from a representative of three experiments with different human MSCs. ∗∗, p < .01. Abbreviations: IFNγ, interferon-γ; IL, interleukin; L-NMMA, NG-monomethyl-L-arginine; MSCs, mesenchymal stem cells; 1-MT, 1-methyl-DL-tryptophan; PBMCs, peripheral blood mononuclear cells; TGFβ, transforming growth factor-β; TNFα, tumor necrosis factor-α.

Download figure to PowerPoint

Immunosuppression by Human MSCs Was Not Intrinsic, but Induced by Inflammatory Cytokines

With the mouse system, we have established that the immunosuppressive effect of MSCs is not intrinsic, but is inducible by IFNγ in combination with either TNFα or IL-1 [6]. Therefore, to test whether immunosuppression by human MSCs also requires activation by inflammatory cytokines, we co-cultured human MSCs with activated T-cell blasts derived from freshly isolated human PBMCs. These T-cell blasts do not produce cytokines unless re-activated. We found that, like mouse MSCs, human MSCs do not inhibit the T-cell blast proliferation unless the MSCs are pre-activated by inflammatory cytokines (Fig. 1B). Unlike that for mouse MSCs, however, IFNγ alone was sufficient to induce significant immunosuppression by human MSCs, although the addition of TNFα could further enhance this function (Fig. 1B). This enhancement of IFNγ-induced immunosuppression by TNFα might be due to the fact that chemokine induction in MSCs requires both TNFα and IFNγ (see below). Furthermore, 1-MT reversed the inhibition of T-cell proliferation by human MSCs stimulated by either IFNγ alone or IFNγ + TNFα (Fig. 1B). Therefore, similar to mouse MSCs, human MSCs must also first be stimulated by inflammatory cytokines to gain immunosuppressive function.

IDO Plays a Key Role in Immunosuppression by MSCs Derived from Humans and Non-Human Primates

The difference among various species in inflammatory cytokine-induced expression of iNOS was previously noted in macrophages [24]. NO was found to be induced by inflammatory cytokines in macrophages of mouse, rat, and bovine origin, but not caprine, lapin, porcine, and human macrophages [24, 25]. Importantly, the roles of iNOS [6, 10] and IDO [11, 26, 27] in MSC-mediated immunosuppression are still being debated. We believe that the disparate findings regarding the roles of iNOS and IDO in MSC-mediated immunosuppression might in fact be due to the different species studied. Therefore, we evaluated the roles of IDO and NO in the T-cell-inhibitory function of MSCs from mouse, human, and rhesus monkey, side by side. We found that immunosuppression by mouse MSCs was completely reversed by inhibition of NO by L-NMMA, whereas immunosuppression by human and monkey MSCs was reversed by the IDO inhibitor, 1-MT, indicating that MSCs from non-human primates and humans utilize IDO as the major effector of immunosuppression, whereas mouse MSCs utilize NO (Fig. 2). Therefore, it is IDO, rather than NO, that plays a critical role in immunosuppression mediated by human and non-human primate MSCs.

thumbnail image

Figure 2. Immunosuppression by human, non-human primate and mouse MSCs was mediated by indoleamine 2,3-dioxygenase and nitric oxide, respectively. PBMCs isolated from fresh human, monkey, or mouse blood were stained with CFSE and co-cultured with their corresponding MSCs at a 1:10 ratio (MSC/PBMC) in wells pre-coated with anti-CD3 and anti-CD28. L-NMMA (1 mM) or 1-MT (0.5 mM) was added at the beginning of the co-culture. Cell proliferation was assessed by flow cytometry after 48 hours. Each halving of fluorescence intensity (log scale) represents one complete cell division. Results represent four experiments with different preparations of human, monkey, and mouse MSCs. Abbreviations: CFSE, carboxyfluorescein diacetate succinimidyl ester; L-NMMA, NG-monomethyl-L-arginine; MSCs, mesenchymal stem cells; 1-MT, 1-methyl-DL-tryptophan; PBMCs, peripheral blood mononuclear cells.

Download figure to PowerPoint

Distinct Expression Patterns of iNOS and IDO Induced by Inflammatory Cytokines in Mouse and Human MSCs

As we have reported, iNOS is highly induced in mouse MSCs upon stimulation with inflammatory cytokines [6]. The most definitive proof of the central role for iNOS in immunosuppression by mouse MSCs comes from our experiments showing that iNOS inhibitors block immunosuppression and that MSCs derived from iNOS-deficient mice completely lack immunosuppressive capability [6]. As described above, immunosuppression by human MSCs can be reversed by 1-MT, but not by iNOS inhibitors (Fig. 2), whereas the converse is true with mouse MSCs [6]. Clearly, there are different roles for iNOS and IDO in mouse and human MSCs. We therefore examined the magnitude of expression of these molecules in human and mouse MSCs stimulated with various cytokine combinations side by side. As shown by RT-PCR analysis (Fig. 3A), after their respective optimal stimulation, human MSCs expressed very high levels of IDO (more than twofold that of GAPDH) and mouse MSCs expressed extremely high iNOS levels (fivefold that of GAPDH). In contrast, expression levels of either iNOS in human MSCs or IDO in mouse MSCs were minimal (only 10−4∼10−3 that of GAPDH). Furthermore, iNOS induction in mouse MSCs required TNFα or IL-1 concomitant with IFNγ, whereas IFNγ alone was sufficient for IDO upregulation in human MSCs since the addition of TNFα or IL-1 had no significant effect (Fig. 3A).

thumbnail image

Figure 3. Distinct expression pattern of iNOS and IDO in mouse and human MSCs after stimulation with inflammatory cytokines. (A): Human and mouse MSCs were supplemented with the indicated combinations of their respective species-specific recombinant cytokines (20 ng/ml each). Twelve hours later, iNOS and IDO mRNA were assayed by real-time polymerase chain reaction and GAPDH was used as an internal control. The expression levels are shown normalized to the level of GAPDH mRNA (defined as one arbitrary unit). (B): Protein levels of human IDO and mouse iNOS were determined by Western blot analysis. Results are representative of three experiments. Abbreviations: IDO, indoleamine 2,3-dioxygenase; IFNγ, interferon-γ; IL, interleukin; iNOS, inducible nitric oxide synthase; MSCs, mesenchymal stem cells; TNFα, tumor necrosis factor-α.

Download figure to PowerPoint

To further verify the intensity of expression of iNOS in mouse MSCs and IDO in human MSCs, we performed Western blot on whole cell extracts from MSC with or without cytokine stimulation. We found indeed that human IFNγ induced significant expression of IDO in human cells, whereas mouse IFNγ and TNFα induced iNOS protein in mouse MSCs. In contrast, IDO in mouse MSCs and iNOS in human MSCs were undetectable at the protein level, confirming our results from quantitative PCR (Fig. 3B and data not shown). Therefore, the effector molecule for MSC-mediated immunosuppression is NO in mouse, but IDO in human.

Since we have shown that MSCs from rhesus monkey bone marrow can inhibit T-cell receptor-induced proliferation of monkey PBMCs, and that this effect is reversed by 1-MT, we also examined IDO expression by these MSCs. We found that, similar to human MSCs, monkey MSCs treated with IFNγ showed a dramatic induction in IDO expression (data not shown). Thus, the role of IDO in MSC-mediated immunosuppression is likely to be evolutionally conserved in non-human primates and humans.

T-Cell Cytokines Induce Abundant Expression of CXCR3 Family Chemokines in Human MSCs

IDO-mediated immunosuppression is achieved through tryptophan degradation and formation of secondary metabolites in the local environment [28]. Interestingly, in our studies, we found that T-cell inhibition by human MSCs also required T cells in proximity to the human MSCs since the effect was significantly reduced when the cells were separated by a permeable membrane in Transwell culture plates (Fig. 4A). Using mouse MSCs, we previously showed that chemokines play a critical role in MSC-mediated immunosuppression and that cell separation in Transwells also abolished the immunosuppressive effect [1, 6]. We therefore examined the expression profile of chemokines and chemokine-related genes in human MSCs stimulated by supernatant from PBMC cultured with or without OKT3 antibody stimulation, using a human chemokine and chemokine receptor PCR array kit. The expression pattern of these genes was similar to that of mouse MSCs reported previously [6]. CXCR3 family genes (CXCL9, CXCL10, and CXCL11), the T-cell-specific chemokines, were the most strongly induced. Also induced were CXCL3, -6, and -8, which are ligands for CXCR1 and CXCR2, chemokine receptors found on polymorphonuclear leukocytes, mast cells, monocytes, and macrophages (Table 2). Importantly, although IDO is significantly induced by IFNγ alone, the greatest induction of CXCR3 family chemokines still required IFNγ and TNFα (Fig. 4B), which may explain why TNFα is required for maximal immunosuppression by human MSCs.

thumbnail image

Figure 4. Induction of CXCR3 family chemokines requires both IFNγ and TNFα. (A): Human MSCs (5 × 104 per well) were co-cultured with fresh human PBMCs at a 1:10 ratio in the presence of OKT3 (1 μg/ml) in Transwells (0.4-μm pore membrane) and proliferation was assayed after 72 hours. Each cell type was either placed together in one chamber where they were in contact with each other (“contact”) or separated in the two chambers of the Transwell (“no contact”). Data shown are mean ± SD. Representative of three experiments. ∗∗∗, p < .001. (B): Human MSCs (passage five) were supplemented with the indicated combinations of recombinant human cytokines (20 ng/ml each). Twelve hours later, CXCL9, CXCL10, and CXCL11 mRNA were assayed by real-time polymerase chain reaction and normalized to the level of GAPDH mRNA, defined as one arbitrary unit. Results are representative of four experiments. Abbreviations: IFNγ, interferon-γ; IL, interleukin; MSCs, mesenchymal stem cells; PBMCs, peripheral blood mononuclear cells; TNFα, tumor necrosis factor-α.

Download figure to PowerPoint

Table 2. Induction of chemokines and related genes in mesenchymal stem cells treated with OKT3-activated human peripheral blood mononuclear cells supernatant (GAPDH defined as 1 × 107 units)
inline image

Chemotaxis Is Critical for Effective Immunosuppression by Human MSCs

To demonstrate the role of the chemokine-mediated T-cell chemotaxis in human MSC-induced immunosuppression, we examined whether human MSCs are chemotactic for activated human T cells using the ChemoTx system. As shown in Figure 5A, brisk chemotaxis of human T-cell blasts could be induced by secreted products of human MSCs stimulated with OKT3-activated human PBMC supernatant (Sup) or by IFNγ combined with TNFα, whereas treatment with IFNγ alone was ineffective (TNFα alone also had no effect, data not shown). This chemotaxis was largely dependent on chemokine receptor CXCR3 since anti-CXCR3 abrogated T-cell chemotaxis in response to MSCs (Fig. 5A).

thumbnail image

Figure 5. Inflammatory cytokine-induced immunosuppression by human MSCs requires chemotaxis. (A): Chemotaxis assays were performed using the NeuroProbe ChemoTx Chemotaxis System with a 5-μm pore polycarbonate membrane separating the upper and lower chambers. Culture supernatant from MSCs stimulated with IFNγ, IFNγ + TNFα (20 ng/ml each), or OKT3-activated human peripheral blood mononuclear cells supernatant (“Sup”, 1:2 dilution) was added to the lower chambers, and activated human T-cell blasts + interleukin-2 (IL-2) (200 U/ml) were added to the upper chambers along with or without anti-human CXCR3 (10 μg/ml). After a 3-hour incubation, the cells that migrated through the pores and into the lower chambers were quantitated using the MTT assay and a chemotaxis index was calculated. (B): To measure the effect of chemotaxis on T-cell proliferation, human MSCs were added to the lower chambers with or without IFNγ and TNFα (20 ng/ml each) for 24 hours. Human T-cell blasts were then added to the upper chambers with IL-2 and incubated for 1 hour and proliferation was determined. Results are representative of three experiments. ∗∗, p < .01; ∗∗∗, p < .001. Abbreviations: IFNγ, interferon-γ; MSCs, mesenchymal stem cells; TNFα, tumor necrosis factor-α.

Download figure to PowerPoint

The role of chemotaxis in the inhibition of T-cell proliferation was also examined using the ChemoTx system. Human MSCs in the lower wells were treated with IFNγ, TNFα, or their combination, and T-cell blasts (with IL-2) were added to the upper wells. Thus, chemokines produced by MSCs would induce T-cell migration into the lower wells, where IDO secreted from MSCs should deplete tryptophan, thus inhibiting T-cell proliferation, as determined by 3H-thymidine incorporation. We found that the proliferation of T-cell blasts was significantly inhibited by MSCs in the presence of IFNγ (Fig. 5B). Addition of TNFα to IFNγ further inhibited T-cell proliferation (Fig. 5B), whereas TNFα alone had no effect (data not shown). Anti-CXCR3 significantly reversed the incremental inhibition induced by TNFα, but did not affect that induced by IFNγ alone (Fig. 5B). These data further support a role for chemotaxis in human MSC-mediated immunosuppression, especially involving the T-cell-specific CXCR3 family chemokines.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

The immunomodulatory properties of bone marrow-derived mesenchymal stem cells have generated enormous interest in the scientific community for their potential application in novel immunosuppressive therapies. However, the molecular mechanisms underlying such active immunosuppression remain controversial. Recent investigations have focused on two regulatory molecules: IDO and NO [6, 10, 11, 26, 27]. In our previous studies, we demonstrated a critical role for NO in mouse MSC-mediated immunosuppression. In the present study, however, we found that NO is not involved in human MSC-mediated immunosuppression. Instead, we show that MSCs from humans and non-human primates employ IDO as their major effector molecule of T-cell inhibition. The species-associated difference is also supported by the differential expression of IDO and iNOS in human and mouse MSCs: in human MSCs, inflammatory cytokines induce the expression of IDO, but not iNOS, whereas in mouse MSCs the opposite is true. Interestingly, we found that MSCs derived from rhesus monkey also employ IDO to suppress T-cell responses. Finally, although the effector mechanisms differ between mouse and human MSCs, they both depend on the induction of T-cell chemokines to achieve effective immunosuppression by MSCs. These findings may help explain the discrepancies in previous attempts to identify the molecular mechanisms of MSC-mediated immunosuppression and advance our understanding of immunosuppression by MSCs, thus allowing better utilization of these special adult stem cells in clinical practice.

One of the most remarkable successful clinical applications of MSCs is their effective treatment of acute steroid-refractory GvHD in humans [3]. In a recent clinical phase II study of 55 patients with acute GvHD (grade 2-4), Le Blanc and colleagues reported that infusion of 1.4 × 106 MSCs per kilogram of body weight led to complete remission in 55% of patients and partial or complete responses in 71% of patients [2]. Importantly, these cells functioned in a non-HLA restricted manner and no side effects were found during or immediately after MSC infusion. To improve this cell-based GvHD treatment, immunological studies addressing the molecular mechanisms of MSC-mediated immunosuppression in vivo are needed. In our previous study using a mouse model, we showed that MSCs exhibit a strong therapeutic effect on severe GvHD, and that this response is dependent on inflammatory cytokines and NO in vivo [6]. Human MSCs, while utilizing a different effector molecule for immunosuppression, still share the most important characteristic of mouse MSCs: inducibility by inflammatory cytokines. Thus, the mouse model is still a valid system for better defining the underlying molecular mechanisms and developing MSC-based treatments for human GvHD.

IDO breaks down the essential amino acid tryptophan, which not only depletes tryptophan but also leads to the formation of immunomodulatory tryptophan metabolites along the kynurenine pathway, thus leading to a local tolerogenic microenvironment [28]. Both of these contribute to its immunosuppressive effect. Recently, IDO has been implicated as a normal, endogenous mechanism of peripheral tolerance and immunosuppression in a number of settings, such as maternal tolerance of the fetus, immunoprivilege of the cornea, and immunosuppression inside tumors [17, 18]. The importance of IDO in immune tolerance has been demonstrated in vivo by using 1-methyl-L-tryptophan (1-MT) to block IDO. Administration of 1-MT prevents CTLA4-Ig-induced tolerance for islet cell allografts and systemic tolerance for foreign antigens induced by antigen introduction into the anterior chamber of the eye in animal models [29, 30]. 1-MT also causes exacerbation of autoimmune and allergic reactions [31]. In pre-clinical studies, 1-MT shows synergy with chemotherapy in the treatment of some tumors [32]. In contrast, ectopic overexpression of IDO by gene transfer results in suppression of immune responses to allogeneic transplants [33]. Many cell types have been found to express IDO in response to inflammation. However, certain subsets of APCs seem to be inherently more responsive to the induction of IDO expression, such as a subset of plasmacytoid DCs or CD8α+ splenic DCs [34]. IDO expression by antigen-presenting cells has been shown to be upregulated by LPS, CpG dinucleotides, and inflammatory cytokines, which are frequently found at sites of inflammation [28, 30]. Vaccine adjuvants, such as Mycobacterium bovis and Listeria monocytogenes, are also potent inducers of IDO in vivo [35]. Thus, various types of proinflammatory stimuli have the potential to induce counter-regulatory IDO, and one of the main pathways of IDO induction has been found to be the IFNγ-STAT1 pathway [36]. However, in an in vivo study using mice deficient in IFNγ and TNFα, systemic IDO expression was largely dependent on TNFα, rather than IFNγ, indicating that there still exists an IFNγ-independent pathway for the induction of IDO expression [37]. Further elucidation of the mechanisms controlling IDO expression in human MSCs is needed.

We have found that mouse MSC-mediated immunosuppression is executed through NO produced by the MSCs and that MSCs derived from iNOS-deficient mice have impaired immunosuppressive capacity [6]. We report here that iNOS does not play a major role in immunosuppression mediated by human MSC. The mechanisms of iNOS regulation in MSCs, however, are unknown in both the human and mouse systems. Differences in the expression of iNOS between human macrophages and mouse macrophages have been demonstrated both in vitro and in vivo [38–40]. Mouse macrophages synthesize a large amount of NO from L-arginine, but human macrophages do not. It is important to note, however, that some human cells do express iNOS in response to activation: for example, human hepatocytes and human smooth muscle cells indeed express iNOS in a manner similar to mouse macrophages [41, 42].

IFNγ is critical for inducing immunosuppression by both mouse and human MSCs. In mouse MSCs, induction of the effector molecule, NO, requires the presence of both IFNγ and either TNFα or IL-1. The same cytokine combinations are also required for induction of T-cell-specific chemokines. These in turn promote T-cell migration into proximity with MSCs, where high concentrations of NO exist [6]. In human MSCs, IFNγ alone is sufficient to induce IDO expression, and addition of TNFα or IL-1 has no effect. As in the mouse system, optimal immunosuppression by human MSCs also depends on chemokines, whose induction requires IFNγ in combination with TNFα; IFNγ alone is not sufficient. Therefore, like that for mouse MSCs, maximal immunosuppression by human MSCs still requires the combination of IFNγ and TNFα. We believe that further investigation of the molecular mechanisms controlling the expression of iNOS, IDO, and chemokines will provide important information for better clinical application of human MSCs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

This work was supported by grants from New Jersey Commission on Science and Technology (NJCST-2042-014-84), Chinese Academy of Sciences, National Institutes of Health (AI057596 and AI50222), and the National Space Biomedical Research Institute (IIH00405) supported by the National Aeronautics and Space Administration through Cooperative Agreement NCC 9-58. The authors also thank Dr. Daniel J. Medina for providing human MSC lines.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES

The authors indicate no potential conflicts of interest.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  9. REFERENCES