Human Leukocyte Antigen-G5 Secretion by ... - Wiley Online Library

TISSUE-SPECIFIC STEM CELLS Human Leukocyte Antigen-G5 Secretion by Human Mesenchymal Stem Cells Is Required to Suppress T Lymphocyte and Natural Killer Function and to Induce CD4ⴙCD25highFOXP3ⴙ Regulatory T Cells ZOHAIR SELMANI,a ABDERRAHIM NAJI,b INES ZIDI,b BENOIT FAVIER,b EMILIE GAIFFE,a LAURENT OBERT,a,c CHRISTOPHE BORG,a PHILIPPE SAAS,a PIERRE TIBERGHIEN,a NATHALIE ROUAS-FREISS,b EDGARDO D. CAROSELLA,b FREDERIC DESCHASEAUXa a

Institut National de la Sante´ et de la Recherche Me´dicale U645, IFR 133, Universite de Franche-Comte, Etablissement Français du Sang Bourgogne-Franche-Comte, Besançon, France; bService de Recherches en HematoImmunologie, CEA-DSV-DRM, Hopital Saint-Louis, IUH, Paris, France; cDe´partement d’Orthopedie, CHU Jean Minjoz, Besançon, France Key Words. Immunosuppression • Human leukocyte antigen-G • Mesenchymal stem cells • Interleukin-10 • Regulatory T cells

ABSTRACT Adult bone marrow-derived mesenchymal stem cells (MSCs) are multipotent cells that are the subject of intense investigation in regenerative medicine. In addition, MSCs possess immunomodulatory properties with therapeutic potential to prevent graft-versus-host disease (GvHD) in allogeneic hematopoietic cell transplantation. Indeed, MSCs can inhibit natural killer (NK) function, modulate dendritic cell maturation, and suppress allogeneic T-cell response. Here, we report that the nonclassic human leukocyte antigen (HLA) class I molecule HLA-G is responsible for the immunomodulatory properties of MSCs. Our data show that MSCs secrete the soluble isoform HLA-G5 and that such secretion is interleukin-10-dependent. Moreover, cell contact between MSCs and allostimulated T

cells is required to obtain a full HLA-G5 secretion and, as consequence, a full immunomodulation from MSCs. Blocking experiments using neutralizing anti-HLA-G antibody demonstrate that HLA-G5 contributes first to the suppression of allogeneic T-cell proliferation and then to the expansion of CD4ⴙCD25highFOXP3ⴙ regulatory T cells. Furthermore, we demonstrate that in addition to their action on the adaptive immune system, MSCs, through HLA-G5, affect innate immunity by inhibiting both NK cell-mediated cytolysis and interferon-␥ secretion. Our results provide evidence that HLA-G5 secreted by MSCs is critical to the suppressive functions of MSCs and should contribute to improving clinical therapeutic trials that use MSCs to prevent GvHD. STEM CELLS 2008;26:212–222

Disclosure of potential conflicts of interest is found at the end of this article.

INTRODUCTION Adult bone marrow (BM) mesenchymal stem cells (MSCs) are multipotent cells. MSCs are isolated from BM by their capacities to adhere on culture flasks and to proliferate extensively. Depending on the culture conditions, MSCs are able to either retain their differentiation potential or to differentiate into osteoblasts, chondroblasts, adipocytes, stromacytes, or neurons [1– 4]. Expanded MSCs were identified as nonhematopoietic cells expressing CD90 and CD73 but not CD45 and CD14 antigens [3]. MSCs are of a particular interest in regenerative medicine, as well as in transplantation. After infusion, MSCs are able to home in to the injured tissues and still be detected 1 year postgraft [5]. Such infusions have been used to regenerate or to repair damaged tissues in both animal models and human clinical trials [6 –10]. MSCs also possess immunomodulatory properties capable of restraining allogeneic reactions. MSCs express human leukocyte antigen (HLA) class I molecules at low levels but do not

express either HLA class II antigens or CD80, CD86, CD40, or CD40L costimulatory molecules [11, 12]. Therefore, MSCs are not able to trigger T-cell activation [13]. Moreover, MSCs are capable, as third-party cells, of inhibiting allostimulated T-cell proliferation and expanding CD4⫹CD25⫹ T cells [14]. These properties have been reported to be mediated by secretion of soluble mediators, such as prostaglandin E2 (PGE2), transforming growth factor-␤1 (TGF-␤1), interleukin-10 (IL-10), and hepatocyte growth factor (HGF) [11, 13], as well as through indoleamine 2,3-dioxygenase (IDO) [14, 15]. MSCs also inhibit the differentiation of monocytes to dendritic cells [16] and modulate B-cell functions [17]. Finally, MSCs have been shown to suppress natural killer (NK) cytotoxicity [18]. In this respect, MSCs offer new perspectives to prevent rejection in human transplantation settings. To date immunosuppressive treatments using MSC infusions to prevent or treat graft-versus-host disease after allogeneic hematopoietic cell transplantation (HCT) is under investigation [19]. Also, MSCs may be used in both HCT and solid transplantation as a novel therapeutic approach to modulate immune rejection.

Correspondence: Frederic Deschaseaux, Ph.D., INSERM U645, 240 route de Dole, 25 000 Besançon, France. Telephone: 33-0-3-81-52-3300; Fax: 33-0-3-81-52-64-80; e-mail: [email protected] Received July 11, 2007; accepted for publication September 26, 2007; first published online in STEM CELLS EXPRESS October 11, 2007. ©AlphaMed Press 1066-5099/2007/$30.00/0 doi: 10.1634/ stemcells.2007-0554

STEM CELLS 2008;26:212–222 www.StemCells.com

Selmani, Naji, Zidi et al.

213

Such a therapeutic could be an alternative to immunosuppressive drugs. HLA-G, the nonclassic HLA class I molecule, shares with MSCs common inhibitory properties on immune cells. Thus, HLA-G alters various immune cell functions, such as NK celland cytotoxic T lymphocytes-mediated cytolysis [20, 21], allogeneic T-cell proliferation, and dendritic cells maturation [22, 23]. HLA-G differs from classic HLA-class I molecules since HLA-G has a low polymorphism and its expression is found in a restricted number of healthy tissues [24]. HLA-G was initially identified in cytotrophoblasts, where it plays a key role in maternal tolerance of the fetus [25, 26]. However, HLA-G expression can be upregulated in various tissues under “pathological” conditions, such as in tumors, as well as after solid organ transplantation [27, 28]. HLA-G can be expressed in seven different isoforms, including four membrane-bound (HLA-G1 to -G4) and three soluble (HLA-G5 to -G7) proteins, due to the alternative splicing of the HLA-G primary transcript (reviewed in [24]). Thus far, three HLA-G receptors have been identified: the killer immunoglobulin-like receptor (KIR2DL4/ CD158d), the leukocyte immunoglobulin (Ig)-like receptor (LILRB1/ILT-2/CD85j), and the LILRB2/ILT-4/CD85d receptor [29, 30]. KIR2DL4 expression is restricted to NK cells, and ILT-4 is myeloid lineage-specific, whereas ILT-2 is expressed by monocytes, dendritic cells, T cells, B cells, and NK cells. Based on the common immunomodulatory properties between HLA-G and MSCs and since HLA-G mRNA and HLA-G proteins were detected within human fetal MSCs [31], we investigated whether adult BM MSCs express HLA-G and whether HLA-G contributed to MSC-mediated immunomodulatory functions.

MATERIALS

AND

METHODS

Cells and Cultures Mesenchymal Stem Cells. Adult human BM samples were collected from healthy volunteers undergoing orthopedic surgery following the ethical guidelines of the Jean Minjoz Hospital in Besançon, France. Mononuclear cells were plated in ␣-minimum essential medium supplemented with 10% fetal bovine serum (FBS) (StemCell Technologies, Vancouver, BC, Canada, http://www.stemcell. com) and 1% penicillin/streptomycin (Invitrogen Ltd., Paisley, U.K., http://www.invitrogen.com). On day 14, when 80% confluence was achieved, cells were split and expanded (passage 1). For all experiments, we used at least three different donors. Peripheral Blood Lymphocytes and NK Cells. Whole blood samples from healthy volunteers donors were provided by the Etablissement Français du Sang (E.F.S., Besançon, France). Approval for this study was obtained from the local ethics committee. Peripheral blood mononuclear cells were seeded on a T-175 culture flask and incubated at 37°C for 2 hours, and nonadherent cells were harvested and used as peripheral blood lymphocytes (PBL). NK cells from peripheral blood were selected using a non-NK cell depletion kit (Miltenyi Biotec, Bergisch Gladbach, Germany, http:// www.miltenyibiotec.com). NK cells were maintained in RPMI 1640 medium supplemented with 20% FBS (Invitrogen) and cultured in the presence of 100 U䡠ml⫺1 recombinant human interleukin-2 (rhIL-2) (R&D Systems Inc., Abingdon, U.K., http://www. rndsystems.com). B-EBV, JEG-3, M8, and K562 Cell Lines. B-EBV and K562 cell lines were cultured in RPMI 1640/10% FBS. Two cell lines were used as control of HLA-G expression, the human choriocarcinoma cell line JEG3 and M8-HLA-G5 (HLA-G5-transfected M8 melanoma cell line). The M8-pcDNA line (pcDNA mock-transfected M8 cell line) was used as a negative control [20]. JEG-3 cells were cultured in Dulbecco’s modified Eagle’s medium/10% FBS, and the

www.StemCells.com

M8-pcDNA and -HLA-G5 cells were cultured in RPMI 1640 medium/10% FBS and 50 ␮g䡠ml⫺1 hygromycin (Invitrogen).

MSC Multipotentiality Assay MSC multipotentiality was determined by differentiating MSCs into osteoblasts, chondroblasts, or adipocytic cells on eight-well chamber slides following procedures described previously [3, 32]. To evaluate the differentiation processes, cells were stained by alizarin red and Alcian Blue to provide evidence for osteoblastic and chondroblastic differentiation, respectively (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). Lipid droplets within adipocytic cells were labeled with Nile red oil solution (Sigma-Aldrich). Undifferentiated MSCs cultured in expansion medium were used as a negative control.

Real-Time Quantitative Reverse TranscriptionPolymerase Chain Reaction for HLA-G Primary and FOXP3 Transcripts Total RNA isolated from MSCs or PBL were extracted using the RNA extraction kit (Qiagen, Courtaboeuf, France, http://www1. qiagen.com) and converted to cDNA by standard methods using reverse transcriptase and random hexamers (Invitrogen). HLA-G Primary Transcripts. Real-time quantitative reverse transcription-polymerase chain reaction (RT-PCR) for HLA-G primary transcripts was performed as described previously [33]. The JEG-3 HLA-G mRNA content served as reference. FOXP3 Transcripts. Quantitative RT-PCRs for FOXP3 were performed using gene-specific probes and universal master mix (Qiagen) on an iCycler iQ thermocycler (Bio-Rad, Hercules, CA, http://www.bio-rad.com). Primers and dual-labeled fluorescent probes were designed using Primer Express software (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Primer pairs and related probes were as follows (sense, antisense, and probe, respectively): 5⬘-GCCCTTCTCCAGGACAGA-3⬘, 5⬘GCTGATCATGGCTGGGTTGT-3⬘, and 5⬘-[FAM]TTTCTGTCAGTCCACTTCACCAAGCCT [TAMRA]-3⬘; Abl (endogenous reference): 5⬘-TGGAGATAACACTCTAAGCATAACTAAAGGT-3⬘, 5⬘-GATGTAGTTGCTTGGGACCCA-3⬘, and 5⬘-[FAM]CCATTTTTGGTTTGGGCTTCACACCATT[TAMRA]-3⬘. FOXP3 mRNA expression was obtained by dividing the relative amount of FOXP3 mRNA for each sample by the relative amount of Abl mRNA from the same sample.

Western Blot Analysis Proteins were extracted, separated by 12% Tris-glycine-SDS-polyacrylamide gel electrophoresis, and subsequently transferred to nitrocellulose membrane (Amersham Biosciences, Saclay, France, http://www.amersham.com). HLA-G was detected with an antiHLA-G antibody, 1:1,000 (clone 4H84), followed by peroxidaseconjugated sheep anti-mouse IgG antibody (Sigma-Aldrich) and visualized by a chemiluminescence reaction using the ECL system (Amersham Biosciences).

Confocal Microscopy Analysis MSCs cultured in Labtek (Nunc, Rochester, NY, http://www. nuncbrand.com) were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 (Sigma-Aldrich). Staining was performed with Alexa 546-labeled phalloidin (Invitrogen) and antiHLA-G (4H84 clone) antibody (Ab) followed by fluorescein isothiocyanate (FITC)-labeled goat anti-mouse Ab (Beckman Coulter, Villepinte, France, http://www.beckmancoulter.com). The samples were mounted in Vectashield medium containing 4,6-diamidino-2phenylindole (Vector Laboratories, Burlingame, CA, http://www. vectorlabs.com). Images were acquired by confocal microscopy on a Carl Zeiss LSM 510 META confocal microscope (Carl Zeiss, Le Pecq, France, http://www.zeiss.com) with a Plan Apochromat ⫻63 N.A. 1.4 objective using the LSM510 software, version 3.2 (Carl Zeiss).

214

MSC-Mediated Immunosuppression Involves HLA-G5

Flow Cytometry Analysis Membrane and intracytoplasmic molecules were detected in (a) expanded MSCs, (b) MSCs stimulated by rhIL-10 (R&D Systems), (c) NK cells, and (d) PBL by using flow cytometry. For intracellular staining, cells were fixed and permeabilized. The following antibodies were used: FITC-conjugated mouse Ab directed against anti-HLA-G1/-G5 isoforms (clone MEM-G/9; Exbio, Vestec, Czech Republic, http://www.exbio.cz), mouse Ab specific to HLA-G5 (clone 5A6G7; Exbio), and other antibodies directed against CD markers conjugated with either FITC or phycoerythrin (PE) (Beckman Coulter or Diaclone, Besançon, France, http://www. diaclone.com/). PBL cocultured with MSCs during 6 days were stained for the intracellular FOXP3 expression using Alexa 488 FOXP3 Kit (clone 259D; Biolegend, San Diego, http://www. biolegend.com). Isotype controls were purchased from either Beckman Coulter, Diaclone, or Biolegend. Analyses were performed through FACSort using CellQuest software (BD Biosciences, San Diego, http://www.bdbiosciences.com).

Mixed Lymphocyte Reaction In a 96-well plate, isolated PBL (105 cells) were used as responding cells and were cocultured in a mixed lymphocyte reaction (MLR) with ␥-irradiated (75 Gy) allogeneic B-EBV cell line (5 ⫻ 104 cells) [34] as stimulating cells (B-EBVⴱ). MSCs were used as nonirradiated third-party cells (5 ⫻ 104 cells). T-cell proliferation was followed by detection of incorporated 5-bromo-2⬘-deoxyuridine (Delfia proliferation kit; PerkinElmer Life and Analytical Sciences, Wellesley, MA, http://www.perkinelmer.com) according to the manufacturer’s instructions. Depending on the experiment, to block HLA-G molecules or rhIL-10, neutralizing Abs directed to HLA-G (clone 87G; Exbio) and rhIL-10 (clone B-S10; Diaclone) were added at a concentration of 20 ␮g䡠ml⫺1 on day 0 and day 3. Irrelevant IgG2a and IgG1 Abs (Diaclone) were used as negative controls at the same concentration. For 6 days, 106 PBL were also cocultured with or without confluent MSCs in a 24-well plate separated or not by a semipermeable membrane (Transwell, 0.4 ␮m; Corning Enterprises, Corning, NY, http://www.corning.com). On day 6, PBL were recovered by agitation and analyzed by flow cytometry (93% ⫾ 2.5% CD3⫹ cells) and real-time quantitative RT-PCR. Cocultured MSCs were also analyzed by flow cytometry and Western blot. For each proliferation experiment, the strongest proliferation of PBL stimulated with B-EBVⴱ was regarded as 100%. All other values obtained were compared with this reference.

Suppression Assays Suppression assays used PBL derived from the primary MLR described above and were subjected to ␥-irradiation (25Gy). ␥-Irradiated PBL were then added to a second MLR as third-party cells at different ratios with responder allogeneic PBL and B-EBVⴱ stimulator cells. T-cell proliferation was measured on day 6 of MLR, as described above.

Enzyme-Linked Immunosorbent Assay Soluble HLA-G5 concentrations were measured in filtered (0.2 ␮m) supernatants (SN) from MSCs or MSCs cocultured in the presence of allostimulated T cells. SN from M8-HLA-G5 cells was used as a positive control. Briefly, enzyme-linked immunosorbent assay (ELISA) used the 5A6G7 Ab (Exbio) as the capture antibody and the pan-HLA class I W6/32 monoclonal antibody (mAb) (Exbio) as the detection antibody, as described previously [35]. Supernatants were also harvested from either cocultured MSCs/PBL or MSCs/NK cells. Then, rhIL-10 and interferon (IFN)-␥ were quantified using commercially available ELISA kits (IL-10 and IFN-␥ Elipair kits; Diaclone).

NK Cytotoxicity Assays NK Cytotoxicity Facing MSCs as Target. NK cells, previously activated with IL-2 (100 U䡠ml⫺1) for 24 hours, were cocultured with target cells for the subsequent 4 hours. NK cells were harvested, washed, and incubated with mouse Ab anti-human CD107b (BD Pharmingen, San Diego, http://www.bdbiosciences.com/); a sec-

Figure 1. Adult bone marrow mesenchymal stem cell (MSC) characterization. (A): MSCs were characterized for cell surface markers expression by flow cytometry. Isotype controls are presented as shaded histograms and analyzed markers as open histograms (n ⫽ 15). (B): To evaluate their multipotentiality, MSCs were cultured in specific differentiation media. Osteoblasts and chondroblasts were stained by alizarin red and Alcian Blue, respectively (magnification, ⫻50). Lipidic droplets were labeled by Nile red (magnification, ⫻600). MSCs maintained in expansion culture medium were subjected to the same staining processes (magnification, ⫻50) and considered as negative control. One representative experiment of 15 is shown.

ondary goat anti-mouse PE-conjugated antibody was then added (Beckman Coulter). Evaluation of NK cell degranulation (i.e., CD107b surface expression) was performed on NK cells cocultured for 4 hours with M8 cells, M8-HLA-G5 cells or MSCs. Analyses were performed through FACSort using CellQuest software (BD Biosciences). NK Presensitized by MSC Cytotoxicity Facing K562 as Target. NK cells were cocultured for 5 days on confluent MSCs following a protocol described previously [36]. NK cells were then recovered and incubated with the K562 cell line as target cells previously labeled by PKH67 green fluorescent linker (Sigma-Aldrich). After 4 hours, the cells were incubated with 5 ␮g䡠ml⫺1 propidium iodide (PI) and analyzed by flow cytometry. The percentages of PKH67 and PI double-positive cells were then evaluated by using a flow cytometer.

Statistical Analysis Values are given as mean ⫾ SEM. Statistical differences between means in each experiment were performed using an analysis of variance test. The difference was considered statistically significant when p ⬍ .05.

RESULTS Characterization of Human Adult BM MSCs Expanded cells were subjected to phenotypic characterization. The CD14, CD34, and CD45 hematopoietic markers were not detected by flow cytometry analyses, whereas the CD44, CD73, CD90, CD106, and CD166 molecules were observed (Fig. 1A).


215

Figure 2. HLA-G expression by bone marrow MSCs. MSCs were lysed to extract either mRNA or proteins. (A): Expression of HLA-G mRNA by MSCs. The HLA-G mRNA primary transcripts from MSCs were evaluated by real-time quantitative reverse transcription-polymerase chain reaction. Results are expressed as mean ⫾ SEM of three independent experiments. The JEG-3 cell line was used as positive control for HLA-G mRNA expression [33]. (B): Expression of HLA-G in MSC lysates. Western blot analyses were performed on cell lysates using a specific anti-HLA-G5 antibody (Ab) (clone 5A6G7) recognizing the intron-4 retaining sequence. A 37-kDa protein corresponding to the HLA-G5 isoform was detected in MSC extracts (upper panel). M8 and M8-HLA-G5 were used as negative and positive controls, respectively. Loading and transfer of equivalent protein levels were confirmed using anti-␤-actin Ab (lower panel) (n ⫽ 5). (C): Intracellular expression of HLA-G within MSCs. Flow cytometry analyses were performed using cells permeabilized and stained by either FITC-conjugated anti-HLA-G Ab (clone MEM-G/9, which recognizes both HLA-G1 and HLA-G5 isoforms; open histograms) or FITC-conjugated irrelevant Ab control (shaded histograms). M8 and M8-HLA-G5 cell lines were used as controls. On each representative histogram, R corresponds to the mean fluorescence intensity (MFI) ratio (i.e., the MFI obtained with anti-HLA-G Ab divided by the MFI obtained with irrelevant control). R is indicated on top right of each histogram (n ⫽ 8). (D): MSCs HLA-G expression assessment by confocal microscopy. MSCs were also cultured into chamber slides. Images were obtained from fixed and permeabilized MSCs. Actin microfilaments were stained with Alexa 546-labeled phalloidin (red). The nuclei were stained with 4,6-diamidino-2-phenylindole (blue). HLA-G expression was detected using a purified anti-HLA-G Ab (clone 4H84) and revealed by an FITC-labeled goat anti-mouse Ab. HLA-G intracellular staining was found predominantly around nuclei, as shown in 2 representative experiments of 10. (E): Soluble HLA-G5 detection in supernatants of MSC cultures. Soluble HLA-G5 concentration in supernatants was quantified by ELISA. M8 cells were used as negative control; each dot represents HLA-G5 concentration found in independent experiment (n ⫽ 8). (F): Decrease of intracellular HLA-G in MSCs according to passages in culture. Intracytoplasmic HLA-G expression analyzed by flow cytometry as shown in (C) and assessed by mean R values showed a regular decrease according to the number of passages. From passage 1 to passage 5, the mean R decreased 1.5-fold (⌬ ⫽ 1.5; n ⫽ 6). Abbreviations: ␮, micrometer; DAPI, 4,6-diamidino-2-phenylindole; FITC, fluorescein isothiocyanate; HLA, human leukocyte antigen; MSC, mesenchymal stem cell.

This phenotype is in agreement with those previously described for MSCs [8]. Moreover, MSCs were analyzed for their expression of cell surface molecules known to be crucial for eliciting an immune response. They were found to express an intermediate level of HLA class I (W6/32 Ab) and to be HLA class II-negative. The CD83 and CD50 molecules, as well as the CD80 and CD86 costimulatory molecules, were not found to be expressed on MSCs (Fig. 1A). Additional characterizations of expanded cells were also performed, which confirmed their ability to differentiate into osteochondroblasts and adipocytic cells (Fig. 1B). All these results confirm that expanded cells have the phenotype and the differentiation abilities observed for MSCs and can thus be considered adult MSCs.

Adult BM MSCs Express Soluble HLA-G5 Proteins As shown in Figure 2A, adult MSCs expressed HLA-G mRNA primary transcripts at a level reaching that of at least the half of HLA-G transcripts found in the JEG-3 choriocarcinoma cell line. This line was used as a positive control and is known to express a high level of HLA-G mRNA [33]. MSCs were then evaluated for their expression of HLA-G proteins. Western blot analyses using either the 4H84 Ab (recognizing all HLA-G isoforms) or www.StemCells.com

the 5A6G7 Ab (specific for the HLA-G5 isoform) showed that the 37-kDa HLA-G5 protein was the unique HLA-G isoform detected in the MSC extracts (Fig. 2B). The M8-HLA-G5transfected cells were used as HLA-G5-positive control [20]. The expression of HLA-G5 by MSCs was confirmed by intracellular flow cytometry analyses using either 5A6G7 or MEMG/9 Ab (Fig. 2C). Confocal microscopy revealed an intracellular staining on cultured MSCs permeabilized and stained with an anti-HLA-G FITC-conjugated Ab (clone 4H84). Fluorescence intensity was more pronounced in perinuclear localization, strongly suggesting de novo HLA-G protein synthesis (Fig. 2D). Otherwise, since HLA-G5 exists as a soluble isoform, its basal level expression in MSC supernatants was quantified by ELISA. Results showed that supernatants from 80% confluent monolayers of MSCs, obtained after 3 days of culture, contained 7.5 ng䡠ml⫺1 soluble HLA-G5 proteins, whereas no HLA-G was detected within cultures of HLA-G negative M8 cell line (Fig. 2E). Finally, in the course of in vitro expansion of MSCs, we observed a decrease of HLA-G expression. From passage 1 to passage 5, cells divided 10 ⫾ 0.5 times, which corresponded to a 1.5-fold decrease of intracytoplasmic fluorescence intensity (i.e., intracellular HLA-G) (Fig. 2F), whereas HLA-G5 content within supernatants was stable (data not shown). Finally, we

216


Figure 3. Inhibition of NK functions by MSCs through HLA-G5. (A): NK cells were cocultured for 4 hours with MSCs or with either M8 or M8-HLA-G5 as controls. NK cell lytic activity against the cell targets was measured using cell surface CD107b molecule (LAMP-2) expression on NK cells as assessed by flow cytometry. Neutralization of HLA-G5 partially but significantly restored the expression of CD107b on NK cells cocultured with MSCs and M8-HLA-G5 cells (n ⫽ 4). (B, C): K562 cells, labeled by the PKH67 fluorescent tracer, were used as target cells and cocultured with NK cells as cytotoxic effectors. NK cells were preincubated with either an anti-HLA-G neutralizing antibody (clone 87G) or an irrelevant antibody. The number of cells positive for PKH67 and PI (dead cells) was determined using dot plot analyses from flow cytometry. Results are reported on histograms (n ⫽ 3). A representative dot plot figure for each experiment with its values (in upper right square) is shown in (C). (D): Interferon-␥ secretions were quantified by enzyme-linked immunosorbent assay within NK cell culture supernatants from experiments reported in Figure 3A and 3B (n ⫽ 6). Abbreviations: HLA, human leukocyte antigen; MSC, mesenchymal stem cell; NK, natural killer; PI, propidium iodide.

performed experiments to determine whether MSCs express the inhibitory receptors for HLA-G, namely ILT-2, ILT-4, and KIR2DL4. None of these receptors was detected on MSCs (data not shown).

HLA-G5 Protects MSCs, as Well as Neighboring Cells, from NK Cytolysis and Inhibits IFN-␥ Secretion by NK Cells HLA-G expression has been described as protecting target cells from NK cell-mediated cytolysis [20]. Thus, we determined whether MSCs could inhibit NK function through HLA-G5. Two different experiments were performed to study (a) the direct MSCs escape from NK cytolysis, and (b) the ability of MSCs to protect neighboring cells from NK lytic activity. The first experiment consisted of the detection of the CD107b protein on the surface of NK cells facing target cells, e.g., M8 (negative control), M8-HLA-G5 (positive control), or MSCs. CD107b is expressed only on the inner membrane of lytic

vesicles containing perforin and granzyme in resting NK cells. It is detected at the cell surface during the NK degranulation. The detection of CD107b has previously been used successfully as a functional marker reflecting the cytolytic activity of NK cells [37]. The M8 and M8-HLA-G5 cells were used as controls since we previously showed with 51Cr release experiments that M8 can be efficiently lysed by the NKL cell line, whereas M8-HLA-G5 was protected [20]. In agreement with our previous findings, CD107b was expressed on NK cells facing M8 but not on NK cells facing M8-HLA-G5 (Fig. 3A). When NK cells were coincubated with MSCs, no CD107b cell surface expression was detected. In contrast, an increased expression of CD107b was observed when MSCs were previously treated with the neutralizing anti-HLA-G Ab (Fig. 3A), thus demonstrating that HLA-G5 molecules are involved in MSC-mediated inhibition of NK cytolysis. A second experiment was carried out to investigate the protecting effect of HLA-G5 secreted by MSCs on neighboring cells by decreasing NK cell lytic ability. The experiment


217

Figure 4. Inhibition of allogeneic T-cell proliferation via secretion of HLA-G5 by bone marrow MSCs requires a close contact between T cells and MSCs. MSC ability to inhibit T-cell proliferation was assessed in MLR, where PBL were cocultured with ␥-irradiated allogeneic B-EBV cells (B-EBVⴱ) in the presence or in the absence of MSCs or MSC SN. Proliferation of allogeneic responder cells (PBL) was assessed by the incorporation of 5-bromo-2⬘deoxyuridine. For each experiment, the strongest proliferation observed for T cells stimulated with allogeneic B-EBVⴱ was regarded as 100%. All other values obtained were compared with this reference. (A): MSCs inhibit allogeneic T-cell proliferation when added as third-party cells in MLR. Neutralizing anti-HLA-G antibody (Ab) (␣-HLA-G) or an irrelevant control Ab (isotype) was added to determine the role of HLA-G in the inhibition of T-cell proliferation by MSCs (n ⫽ 6). (B): A close contact between T cells and MSCs is required to inhibit T-cell alloproliferation. Inhibition of allogeneic proliferation mediated by HLA-G5 was tested in contact versus noncontact conditions between allogeneic T cells (PBL) and MSCs. For noncontact experiments, we used either MSC-conditioned media (SN MSCs) or semipermeable membranes (Transwell) to separate cells. Neutralizing anti-HLA-G Ab was used to evaluate the HLA-G contribution in these different conditions. (C–E): The expression of HLA-G5 protein was then quantified in MSCs recovered from MLR (C, D), as well as in SN of MLR (E). (C, D): Intracytoplasmic HLA-G5 in MSCs recovered from MLR was measured by using flow cytometry for 6 days. (C): Mean ⫾ SEM of mean fluorescence intensity (MFI) ratio pooled from six independent experiments. (D): Raw data of one representative experiment out of six with its MFI ratio values (R). (E): The concentrations of soluble HLA-G5 within SN were quantified by enzyme-linked immunosorbent assay at different times of culture and are provided as kinetic curves (n ⫽ 5). Abbreviations: h, hours; HLA, human leukocyte antigen; MLR, mixed lymphocyte reactions; MSC, mesenchymal stem cell; PBL, peripheral blood lymphocytes; SN, supernatants.

consisted of NK cells cultured for 5 days on MSCs (presensitized NK cells) and then tested for their cytolytic ability against K562 target cell line (Fig. 3B, 3C). NK cytolysis facing K562 cells was followed by the detection of PI⫹ dead cells among the PKH67⫹-labeled K562 cells. Freshly activated NK cells, cocultured with K562 cells, led to approximately 80% of K562 being double-positive for PI and PKH67 after 4 hours of incubation. When NK cells were presensitized with MSCs and used as effector cells against the K562 cell line, the percentage of PI⫹PKH67⫹ cells decreased to 30% (Fig. 3B, 3C). These data support the inhibitory effects of MSCs on NK cytolysis. When the neutralizing antiHLA-G Ab was added to the presensitization experiments, the number of double-positive cells increased to 70%, attesting that HLA-G5 is the key molecule protecting neighboring target cells from NK cell cytolysis. www.StemCells.com

Furthermore, HLA-G5 soluble molecules secreted by MSCs over the experiments described above inhibited the NK cellmediated secretion of IFN-␥. This fact was supported by the observation that the inhibition of IFN-␥ secretion by NK cells was reversed by adding anti-HLA-G Ab during NK presensitization (Fig. 3D).

HLA-G5 Is Involved in the MSC-Mediated Inhibition of Allostimulated T Cells To determine whether MSC immunomodulatory properties toward T cells could be mediated by HLA-G5, MLR experiments were carried out using MSCs as a third-party cell with or without neutralizing anti-HLA-G antibody. MSCs suppressed T-cell alloproliferation by 70%. Such suppression was prevented by blocking HLA-G5 with specific mAb (Fig. 4A). These results demonstrate that HLA-G5 is directly involved in

218


Figure 5. HLA-G5 and IL-10 function in a dependent manner during an allogeneic challenge. MSCs were incubated with either 10 or 50 ng䡠ml⫺1 rhIL-10. After 6 days of cultures, MSCs and MSC supernatants were recovered to quantify both intracytoplasmic (A) and released soluble (B) HLA-G5 by flow cytometry and enzyme-linked immunosorbent assay (ELISA), respectively (n ⫽ 5). (C): During the cocultures containing allogeneic PBL (responder cells), B-EBVⴱ cells (irradiated stimulator cells) and MSCs (third-party cells), IL-10 concentration within supernatants were quantified by ELISA on day 6 (n ⫽ 8). Supernatants of cocultures containing the neutralizing anti-HLA-G antibody (Ab) (␣-HLA-G) or an irrelevant control Ab (isotype) were also evaluated. Each dot corresponds to an independent experiment. (D): The role of IL-10 was also evaluated in mixed lymphocyte reactions by adding a neutralizing anti-IL-10 Ab (␣-IL-10) into the cocultures. An anti-HLA-G (␣-HLA-G) Ab was also used alone or concomitantly with the anti-IL-10 Ab. For each evaluation, three different experiments were performed using MSCs from three different donors (n ⫽ 3). Abbreviations: HLA, human leukocyte antigen; IL, interleukin; MSC, mesenchymal stem cell; PBL, peripheral blood lymphocytes; rhIL, recombinant human interleukin.

MSC-mediated inhibition of T-cell alloproliferation. Since HLA-G5 is a soluble isoform, we investigated the immunosuppressive properties of MSC-conditioned media versus cell:cell contact. MSCs were used as third-party cell in MLR either directly with allostimulated PBL or separated by a semipermeable membrane (Transwell) allowing only soluble factors from MSCs to diffuse into the chamber where the alloreaction (PBL⫹B-EBVⴱ) was conducted. Results showed that MSCs inhibited allogeneic T-cell proliferation, whether they were in close contact with alloreactive T cells or not (Fig. 4B). In accordance with this observation, addition of supernatant from MSC cultures to MLR also inhibited allogeneic T-cell proliferation. This inhibition involved HLA-G5 since addition of an anti-HLA-G blocking Ab restored T-cell alloproliferation. However, the maximal inhibition was observed when MSCs were added directly to the MLR (2.3-fold higher; p ⬍ .05), demonstrating that MSC:T-cell contact played a critical role in MSCmediated suppressive function. This latter inhibition could be prevented by neutralizing HLA-G5, thus leading us to hypothesize that MSC:T-cell contacts possibly enhance HLA-G5 secretion by MSCs, which in turn would induce a more pronounced inhibition of T-cell responses.

HLA-G5 Secretion by MSCs Is Rapidly but Transiently Enhanced After Cell:Cell Contact with Allostimulated T Cells To confirm this hypothesis, we measured soluble HLA-G5 production by MSCs in the presence of allostimulated T cells. Flow cytometric analysis revealed a decrease of intracellular HLA-G5 content within MSCs after 6 days of coculture (R ⫽ 8.3 ⫾ 2.8 to 3.9 ⫾ 0.7; p ⬍ .02; n ⫽ 6) (Fig. 4C, 4D). To examine whether such a decrease was associated with a con-

comitant secretion of soluble HLA-G5 molecules, HLA-G5 concentrations were determined by ELISA in supernatants during the allogeneic reaction. We observed that HLA-G5 secretion was enhanced after 24 hours MLR (14.1 ⫾ 1.7 ng䡠ml⫺1) and slowly decreased until day 4 (9.4 ⫾ 1.2 ng䡠ml⫺1) (Fig. 4E). Such upregulation of HLA-G5 was not observed when PBL were separated from MSCs by a semipermeable membrane (Transwell) (Fig. 4E). Thus, these results demonstrate that close contact interactions between MSCs and T cells in an alloreactive context leads to transient increase of soluble HLA-G5 secretion by MSCs.

HLA-G5 and IL-10 Are Involved in an Interdependent Manner in MSC-Mediated Immunosuppression Interleukin-10 has previously been shown to upregulate the expression of HLA-G on monocytes and trophoblasts [38]. We therefore investigated whether IL-10 could increase HLA-G5 expression by MSCs. For this purpose, rhIL-10 was added at different concentrations to MSC cultures. Intracellular expression and release of soluble HLA-G5 were evaluated by flow cytometry and ELISA, respectively. As shows Figure 5A, rhIL-10 enhanced HLA-G5 intracellular content in MSCs. Furthermore, HLA-G5 concentration was increased in supernatants from rhIL-10-treated MSC cultures (Fig. 5B). Both effects evolved in an IL-10 dose-dependent manner (Fig. 5A, 5B). Since IL-10 plays a role in HLA-G5 MSCs expression, we then investigated in turn whether IL-10 levels in supernatants of cocultures consisting of T cells and MSCs are affected by HLA-G5. These supernatants contained significantly higher levels of IL-10 (75 ⫾ 5 pg䡠ml⫺1) compared with those consisting of T cells alone (20 ⫾ 2 pg䡠ml⫺1; p ⬍ .05). Similar results were


219

Figure 6. HLA-G5 secreted by bone marrow MSCs expands CD4⫹CD25highFOXP3⫹ regulatory T cells within an allostimulated T-cell population. (A, B): On day 6 of mixed lymphocyte reactions (MLR) containing PBL, B-EBV cells and MSCs in the presence of irrelevant antibody (Ab) (isotype) or HLA-G neutralizing Ab (␣-HLA-G), nonadherent cells were harvested and stained with anti-CD4 and anti-CD25 Abs and then analyzed by flow cytometry. As controls, we used PBL and PBL ⫹ B-EBVⴱ cells in the absence of MSCs (Ø). In (A), dot plots from a representative experiment are shown. (B): Percentages ⫾ SEM of CD4⫹CD25high among CD3⫹ T cells from six experiments are shown. (C): FOXP3 mRNA content of the recovered nonadherent cells obtained from MLR on day 6 was determined by real-time quantitative reverse transcription-polymerase chain reaction. Results are expressed as mean ⫾ SEM of normalized FOXP3 expression (n ⫽ 6). (D): Recovered cells after MLR were stained by anti-CD4 phycoerythrin and anti-FOXP3 Alexa 488 Abs. Percentage of CD4⫹FOXP3⫹ cells was then quantified by flow cytometry (n ⫽ 6). Abbreviations: HLA, human leukocyte antigen; MSC, mesenchymal stem cell; PBL, peripheral blood lymphocytes.

observed in immunosuppression assays when T cells were stimulated by allogeneic B-EBV cells (155 ⫾ 7 pg䡠ml⫺1 with MSCs vs. 46 ⫾ 7 pg䡠ml⫺1 without MSCs; p ⬍ .05). Interestingly, addition of the anti-HLA-G neutralizing Ab to these cocultures dramatically decreased IL-10 concentrations within supernatants (56 ⫾ 4 pg䡠ml⫺1; p ⬍ .05) compared with cocultures containing an irrelevant Ab (158 ⫾ 11 pg䡠ml⫺1; p ⬍ .05) (Fig. 5C). Taken together, these results demonstrate that HLA-G5 and IL-10 molecules are closely linked and appear regulated through an amplification feedback loop. In addition to their linked expression, both molecules are required for MSC-mediated immunosuppression. Indeed, blocking either IL-10 or HLA-G5 by specific Abs was able to reverse the MSC-mediated inhibition of allogeneic T-cell proliferation. When IL-10 and HLA-G5 were blocked concomitantly, a more pronounced alloproliferation restoration was observed (Fig. 5D), thus establishing that both molecules acted in synergy.

HLA-G5 Secretion by MSCs Is Required to Expand CD4ⴙCD25highFOXP3ⴙ Regulatory T Cells Previous studies have shown that MSCs are able to generate CD4⫹CD25high regulatory T cells [14]. Here, we asked whether HLA-G5 is involved in such generation. For this purpose, cocultures containing MSCs and allostimulated T cells were performed for 6 days. Allostimulated T cells were then recovered and analyzed for CD4, CD25, and FOXP3 expression. As observed previously [14], the population of CD4⫹CD25high T cells increased significantly among allostimulated T cells when MSCs were added to MLR (12.6% ⫾ 2.5%) compared with allostimulated T cells without MSCs (6.9% ⫾ 1.4%; p ⬍ .05) (Fig. 6A, 6B). This effect was not due to the addition of a third-party cell since addition of an irrelevant cell line (e.g., M8 cells) instead of MSCs did not enhance the percentage of www.StemCells.com

CD4⫹CD25high cells (data not shown). Most importantly, addition of an anti-HLA-G Ab resulted in a significant decrease in the percentage of CD4⫹CD25high T cells (7.3% ⫾1%; p ⬍ .05) (Fig. 6A, 6B), clearly showing that HLA-G5 is required to expand CD4⫹CD25high T cells within the allostimulated T-cell population. Moreover, quantitative RT-PCR showed a significant upregulation of FOXP3 mRNA in allostimulated T cells cocultured with MSCs compared with conditions where MSCs were omitted (FOXP3/Abl ⫽ 1.6 ⫾ 0.3 vs. 0.6 ⫾ 0.2, respectively; p ⬍ .05). Such FOXP3 mRNA upregulation was prevented when neutralizing anti-HLA-G Ab was added to the cultures (FOXP3/Abl ⫽ 0.8 ⫾ 0.2) (Fig. 6C). These results were confirmed by analyzing FOXP3 protein expression within CD4⫹ T cells by intracellular flow cytometry (Fig. 6D).

CD4ⴙCD25highFOXP3ⴙ T Cells Expanded via HLA-G5 Secreted by MSCs Exhibit Regulatory Functions To demonstrate whether T cells generated in the presence of MSCs and containing an enhanced CD4⫹CD25highFOXP3⫹ population exhibit regulatory functions, we first evaluated their ability to respond to a secondary allogeneic stimuli. Results showed that such T cells were hyporesponsive to allogeneic stimuli (Fig. 7A). Next, we analyzed their immunosuppressive function by adding them as a third-party cell in a suppression assay. These T cells significantly suppressed allogeneic stimuli (Fig. 7B). By contrast, T cells generated with MSCs in the presence of anti-HLA-G neutralizing Ab prevented their suppressive functions (Fig. 7B). These latter data are in accordance with results described above (Fig. 6B) showing that CD4⫹CD25high T cells were not expanded when anti-HLA-G neutralizing antibody was added to cocultures. Moreover, allostimulated T cells generated with MSCs, but separated by a


220

Figure 7. Suppressive functions of allostimulated T cells after coculture with MSCs. (A–D): PBL generated during a primary mixed lymphocyte reactions (MLR) consisting of PBL allostimulated with irradiated B-EBVⴱ cells in the absence (PBLØ) or in the presence (PBLMSCs) of MSCs were then tested for their ability to proliferate after a secondary stimulation (n ⫽ 3) (A), and to modulate an allogeneic proliferative response from a secondary MLR in a suppression assay (B–D). (B): Suppressor assay was carried out with PBLMSCs from primary MLR containing either neutralizing anti-human leukocyte antigen (HLA)-G antibody (Ab) (PBLMSCs⫹␣-HLA-G) or control Ab (PBLMSCs⫹isotype) (n ⫽ 3). (C): From cocultures where MSCs were in close contact with PBL (PBLMSCs) or separated by semipermeable membrane (PBLMSCs [Transwell]), PBL were recovered and tested in a suppression assay (n ⫽ 3). (D): Dose-response assays were performed to assess the suppressive functions of allostimulated T cells generated in the presence of MSCs. Allostimulated T cells obtained from cocultures with MSCs but physically separated (MSC Transwell) were used as a control (n ⫽ 3). Abbreviations: MSC, mesenchymal stem cell; PBL, peripheral blood lymphocytes.

semipermeable membrane (Transwell), were significantly less able (⌬ ⫽ 2.2) to inhibit T-cell alloproliferation (Fig. 7C, 7D). Furthermore, the immunosuppression level induced by PBL previously cocultured directly on MSCs was dose-dependent (Fig. 7D).

DISCUSSION MSCs are known as multipotent cells with immunosuppressive properties [39]. Several molecules expressed by MSCs, such as PGE2, HGF, and IDO, have been described as modulating T-cell activation, NK cytotoxicity, and dendritic cell maturation [11, 14 –16]. Here, we demonstrate for the first time that HLA-G5 secreted by MSCs contributes to the immunosuppressive functions of MSCs. As described previously in maternofetal tolerance and tumor escape, HLA-G5 constitutes a potent molecule allowing MSCs to induce immunosuppression. HLA-G gene transcription by fetal and adult MSCs has been reported previously [31]. However, the protein was found only within fetal cells. In this report, we demonstrate the expression of HLA-G proteins within the cytoplasm of MSCs, as well as in the supernatants from MSC culture media. We also observed that neutralizing HLA-G molecules partially restored T-cell proliferation in response to allogeneic stimuli, demonstrating that HLA-G5 expressed by MSCs played a key role in their immu-

nomodulatory properties. HLA-G5 secreted by MSCs was found to be greatly increased during allogeneic challenge, while at the same time intracellular HLA-G5 MSCs content was significantly decreased. Otherwise, we identify IL-10 as a key cytokine upregulating both expression and secretion of HLA-G5 by MSCs. This observation is in agreement with previous reports showing that HLA-G expression is enhanced in IL-10-treated monocytes [38]. In turn, HLA-G5 has been described previously as inducing a Th2 profile consisting of IL-10 secretion [40, 41]. In addition to their linked expression, we demonstrate that both IL-10 and HLA-G5 are required for full MSC-mediated immunosuppression and that both molecules act in synergy. IL-10 and HLA-G5 expression appear regulated according to an amplification feedback loop with, as a result, an efficient suppression of an allogeneic challenge. Recently, MSCs were described to induce CD4⫹CD25⫹ T cells [14, 42]. Here, we demonstrate the ability of MSCs to expand functional CD4⫹CD25highFOXP3⫹ regulatory T cells, and we show for the first time that HLA-G5 secreted by MSCs is absolutely required for this purpose. We demonstrate that MSCs mainly expressed the HLA-G5 soluble isoform. Interestingly, HLA-G5 was found previously, as specifically expressed by fetal and adult erythroid progenitors residing within BM [43]. This finding may indicate a selective expression of the soluble HLA-G5 isoform by cells located within this tissue. Thus, such expression could contribute to the high frequency of regulatory T cells within BM [44]. Expansion of CD4⫹CD25highFOXP3⫹ T cells is mediated through

Selmani, Naji, Zidi et al. HLA-G5 secretion by MSCs and required a preliminary direct cell:cell contact between T cells and MSCs. Several reports have failed to observe HLA-G-mediated increase of CD4⫹CD25highFOXP3⫹ regulatory T cells. However, HLA-G has been described as inducing CD3⫹CD4low and CD3⫹CD8low immunosuppressive T cells [45]. Discrepancy with our results is probably due to the distinct HLA-G⫹ cells used to generate suppressive T cells. HLA-G5 is here provided by MSCs that, as we demonstrated, need direct contact with T cells to increase HLA-G expression. Furthermore, MSCs also express and secrete several other factors contributing to regulatory T cell induction, such as PGE2, TGF-␤, and HGF. Previous studies have reported the ability of alloreactive T cells to express HLA-G proteins [46]. In our experiments, we were unable to detect HLA-G expression on the CD3⫹ T-cell populations, as well as on B-EBV cells. Furthermore, addition of anti-HLA-G neutralizing antibodies in our MLR containing allogeneic T cells and irradiated B-EBV cells had no significant effect on T-cell alloproliferation (data not shown). Thus, all effects displayed by HLA-G in our study can clearly be attributed to HLA-G5 proteins expressed and secreted by MSCs. In addition to the key role of HLA-G5 secreted by MSCs on adaptive immunity, we demonstrated an important function of HLA-G5 on innate immunity. Indeed, we show that MSCs impair NK cells to secrete the perforin/granzyme-containing granules and to lyse target cells. However, NK cells have been described previously as able to lyse MSCs even if IL-2-activated NK cell proliferation was inhibited [18]. Here, we used another protocol where NK cells were cultured on MSCs for 5 days before assessing their lytic potential. In this case, our results are in accordance with data found by Sotiropoulou et al. [36]. In addition, we described for the first time that HLA-G5 expressed by MSCs inhibits IFN-␥ secretion by NK cells. Several other factors participating in MSC-induced immunomodulation, such as IDO, have been described previously [15, 47]. We were unable to detect the kynurenine metabolite in our cocultures (data not shown), thus strongly suggesting that IDO production did not play a significant role in our experiments. Furthermore, addition of tryptophan or of IDO inhibitor to allostimulated T cells cocultured with MSCs has been reported to not restore T-cell proliferation [48, 49]. Otherwise, Aggarwal and Pittenger observed a negative effect of the PGE2 secreted by MSCs on T-cell proliferation that was reversed by addition of an inhibitor of cyclooxygenase-2 (Cox-2) [14]. The authors observed a transient secretion since they were unable to detect PGE2 after 5 days of culture. As our experiments were all

221

performed on day 6, the PGE2 effect should be negligible compared with the effects mediated by HLA-G molecules. Accordingly, a previous report showed a low PGE2 production by MSCs on day 5 of culture, and addition of a PGE2 inhibitor (indomethacin) did not restore the alloproliferative response [50]. We also added the Cox-2 antagonist SC-236 into allostimulated T-cell-MSC cocultures and found no reversion of the immunosuppression, as well as no effect on HLA-G expression (data not shown). However, potential relationships among IDO, PGE2, and HLA-G molecules remain to be evaluated.

CONCLUSION In conclusion, we demonstrate that HLA-G5 expressed and secreted by MSCs contributes to (a) the suppression of the NK lytic activity and IFN-␥ secretion, (b) the direct inhibition of allogeneic T-cell responses, (c) the increase of IL-10 concentration in the alloreaction microenvironment, and (d) the expansion of CD4⫹CD25highFOXP3⫹ regulatory T cells. Such properties require first a direct cell:cell contact between alloreactive T cells and MSCs (supplemental online Fig. 1). Overall, our results provide evidence that HLA-G5 secreted by MSCs is critical to the suppressive functions of MSCs. Such novel findings should contribute to improving clinical therapeutic trials using MSC infusion for immunomodulation purposes.

ACKNOWLEDGMENTS We thank Charlene Lasgi, Thomas Domet, Christophe Ferrand, Anne Duperrier, Patricia Ame-Thomas, and Jean-Paul RemyMartin for technical assistance. Z.S. was supported by a grant from the Conseil Scientifique de l’Etablissement Français du Sang. A.N. was supported by a grant from the French Commissariat a` L’Energie Atomique. This work was supported by grants from the Ligue Contre le Cancer (Doubs) and the Fondation pour la Transplantation (ET-040615). Z.S. and A.N. contributed equally to this study.

DISCLOSURE

8

2

3 4 5 6 7

Deschaseaux F, Charbord P. Human marrow stromal precursors are alpha 1 integrin subunit-positive. J Cell Physiol 2000;184:319 –325. Friedenstein AJ, Gorskaja JF, Kulagina NN. Fibroblast precursors in normal and irradiated mouse hematopoietic organs. Exp Hematol 1976; 4:267–274. Pittenger MF, Mackay AM, Beck SC et al. Multilineage potential of adult human mesenchymal stem cells. Science 1999;284:143–147. Woodbury D, Schwarz EJ, Prockop DJ et al. Adult rat and human bone marrow stromal cells differentiate into neurons. J Neurosci Res 2000; 61:364 –370. Devine SM, Cobbs C, Jennings M et al. Mesenchymal stem cells distribute to a wide range of tissues following systemic infusion into nonhuman primates. Blood 2003;101:2999 –3001. Bruder SP, Kurth AA, Shea M et al. Bone regeneration by implantation of purified, culture-expanded human mesenchymal stem cells. J Orthop Res 1998;16:155–162. Horwitz EM, Gordon PL, Koo WK et al. Isolated allogeneic bone marrow-derived mesenchymal cells engraft and stimulate growth in

www.StemCells.com

CONFLICTS

The authors indicate no potential conflicts of interest.

REFERENCES 1

OF POTENTIAL OF INTEREST

9 10 11 12 13

14 15

children with osteogenesis imperfecta: Implications for cell therapy of bone. Proc Natl Acad Sci U S A 2002;99:8932– 8937. Pittenger MF, Martin BJ. Mesenchymal stem cells and their potential as cardiac therapeutics. Circ Res 2004;95:9 –20. Quarto R, Mastrogiacomo M, Cancedda R et al. Repair of large bone defects with the use of autologous bone marrow stromal cells. N Engl J Med 2001;344:385–386. Wakitani S, Goto T, Pineda SJ et al. Mesenchymal cell-based repair of large, full-thickness defects of articular cartilage. J Bone Joint Surg Am 1994;76:579 –592. Di Nicola M, Carlo-Stella C, Magni M et al. Human bone marrow stromal cells suppress T-lymphocyte proliferation induced by cellular or nonspecific mitogenic stimuli. Blood 2002;99:3838 –3843. Majumdar MK, Keane-Moore M, Buyaner D et al. Characterization and functionality of cell surface molecules on human mesenchymal stem cells. J Biomed Sci 2003;10:228 –241. Bartholomew A, Sturgeon C, Siatskas M et al. Mesenchymal stem cells suppress lymphocyte proliferation in vitro and prolong skin graft survival in vivo. Exp Hematol 2002;30:42– 48. Aggarwal S, Pittenger MF. Human mesenchymal stem cells modulate allogeneic immune cell responses. Blood 2005;105:1815–1822. Meisel R, Zibert A, Laryea M et al. Human bone marrow stromal cells


222

16 17 18

19 20 21

22

23 24 25 26

27 28 29 30 31 32

33

inhibit allogeneic T-cell responses by indoleamine 2,3-dioxygenasemediated tryptophan degradation. Blood 2004;103:4619 – 4621. Jiang XX, Zhang Y, Liu B et al. Human mesenchymal stem cells inhibit differentiation and function of monocyte-derived dendritic cells. Blood 2005;105:4120 – 4126. Corcione A, Benvenuto F, Ferretti E et al. Human mesenchymal stem cells modulate B-cell functions. Blood 2006;107:367–372. Spaggiari GM, Capobianco A, Becchetti S et al. Mesenchymal stem cell-natural killer cell interactions: Evidence that activated NK cells are capable of killing MSCs, whereas MSCs can inhibit IL-2-induced NKcell proliferation. Blood 2006;107:1484 –1490. Le Blanc K, Rasmusson I, Sundberg B et al. Treatment of severe acute graft-versus-host disease with third party haploidentical mesenchymal stem cells. Lancet 2004;363:1439 –1441. Riteau B, Rouas-Freiss N, Menier C et al. HLA-G2, -G3, and -G4 isoforms expressed as nonmature cell surface glycoproteins inhibit NK and antigen-specific CTL cytolysis. J Immunol 2001;166:5018 –5026. Rouas-Freiss N, Marchal RE, Kirszenbaum M et al. The alpha1 domain of HLA-G1 and HLA-G2 inhibits cytotoxicity induced by natural killer cells: Is HLA-G the public ligand for natural killer cell inhibitory receptors? Proc Natl Acad Sci U S A 1997;94:5249 –5254. Lila N, Rouas-Freiss N, Dausset J et al. Soluble HLA-G protein secreted by allo-specific CD4⫹ T cells suppresses the allo-proliferative response: A CD4⫹ T cell regulatory mechanism. Proc Natl Acad Sci U S A 2001;98:12150 –12155. Ristich V, Liang S, Zhang W et al. Tolerization of dendritic cells by HLA-G. Eur J Immunol 2005;35:1133–1142. Carosella ED, Moreau P, Le Maoult J et al. HLA-G molecules: From maternal-fetal tolerance to tissue acceptance. Adv Immunol 2003;81: 199 –252. McMaster MT, Librach CL, Zhou Y et al. Human placental HLA-G expression is restricted to differentiated cytotrophoblasts. J Immunol 1995;154:3771–3778. Rouas-Freiss N, Goncalves RM, Menier C et al. Direct evidence to support the role of HLA-G in protecting the fetus from maternal uterine natural killer cytolysis. Proc Natl Acad Sci U S A 1997;94: 11520 –11525. Le Rond S, Azema C, Krawice-Radanne I et al. Evidence to support the role of HLA-G5 in allograft acceptance through induction of immunosuppressive/regulatory T cells. J Immunol 2006;176:3266 –3276. Lila N, Carpentier A, Amrein C et al. Implication of HLA-G molecule in heart-graft acceptance. Lancet 2000;355:2138. Colonna M, Navarro F, Bellon T et al. A common inhibitory receptor for major histocompatibility complex class I molecules on human lymphoid and myelomonocytic cells. J Exp Med 1997;186:1809 –1818. LeMaoult J, Zafaranloo K, Le Danff C et al. HLA-G up-regulates ILT2, ILT3, ILT4, and KIR2DL4 in antigen presenting cells, NK cells, and T cells. FASEB J 2005;19:662– 664. Gotherstrom C, West A, Liden J et al. Difference in gene expression between human fetal liver and adult bone marrow mesenchymal stem cells. Haematologica 2005;90:1017–1026. Gindraux F, Selmani Z, Obert L et al. Human and rodent bone marrow mesenchymal stem cells that express primitive stem cell markers can be directly enriched by using the CD49a molecule. Cell Tissue Res 2007; 327:471– 483. Moreau P, Mouillot G, Rousseau P et al. HLA-G gene repression is

34

35 36 37 38 39 40 41

42

43 44 45 46

47 48 49 50

reversed by demethylation. Proc Natl Acad Sci U S A 2003;100: 1191–1196. Sauce D, Bodinier M, Garin M et al. Retrovirus-mediated gene transfer in primary T lymphocytes impairs their anti-Epstein-Barr virus potential through both culture-dependent and selection process-dependent mechanisms. Blood 2002;99:1165–1173. Rebmann V, Lemaoult J, Rouas-Freiss N et al. Report of the Wet Workshop for Quantification of Soluble HLA-G in Essen, 2004. Hum Immunol 2005;66:853– 863. Sotiropoulou PA, Perez SA, Gritzapis AD et al. Interactions between human mesenchymal stem cells and natural killer cells. STEM CELLS 2006;24:74 – 85. Betts MR, Brenchley JM, Price DA et al. Sensitive and viable identification of antigen-specific CD8⫹ T cells by a flow cytometric assay for degranulation. J Immunol Methods 2003;281:65–78. Moreau P, Adrian-Cabestre F, Menier C et al. IL-10 selectively induces HLA-G expression in human trophoblasts and monocytes. Int Immunol 1999;11:803– 811. Uccelli A, Pistoia V, Moretta L. Mesenchymal stem cells: A new strategy for immunosuppression? Trends Immunol 2007;28:219 –226. Kanai T, Fujii T, Kozuma S et al. Soluble HLA-G influences the release of cytokines from allogeneic peripheral blood mononuclear cells in culture. Mol Hum Reprod 2001;7:195–200. Kanai T, Fujii T, Unno N et al. Human leukocyte antigen-G-expressing cells differently modulate the release of cytokines from mononuclear cells present in the decidua versus peripheral blood. Am J Reprod Immunol 2001;45:94 –99. Maccario R, Podesta M, Moretta A et al. Interaction of human mesenchymal stem cells with cells involved in alloantigen-specific immune response favors the differentiation of CD4⫹ T-cell subsets expressing a regulatory/suppressive phenotype. Haematologica 2005; 90:516 –525. Menier C, Rabreau M, Challier JC et al. Erythroblasts secrete the nonclassical HLA-G molecule from primitive to definitive hematopoiesis. Blood 2004;104:3153–3160. Zou L, Barnett B, Safah H et al. Bone marrow is a reservoir for CD4⫹CD25⫹ regulatory T cells that traffic through CXCL12/CXCR4 signals. Cancer Res 2004;64:8451– 8455. Rouas-Freiss N, Naji A, Durrbach A et al. Tolerogenic functions of human leukocyte antigen G: From pregnancy to organ and cell transplantation. Transplantation 2007;84(Suppl):S21–S25. Le Rond S, Le Maoult J, Creput C et al. Alloreactive CD4⫹ and CD8⫹ T cells express the immunotolerant HLA-G molecule in mixed lymphocyte reactions: In vivo implications in transplanted patients. Eur J Immunol 2004;34:649–660. Hwu P, Du MX, Lapointe R et al. Indoleamine 2,3-dioxygenase production by human dendritic cells results in the inhibition of T cell proliferation. J Immunol 2000;164:3596–3599. Gieseke F, Schutt B, Viebahn S et al. Human multipotent mesenchymal stromal cells inhibit proliferation of PBMCs independently of IFNgammaR1 signaling and IDO expression. Blood 2007;110:2197–2200. Tse WT, Pendleton JD, Beyer WM et al. Suppression of allogeneic T-cell proliferation by human marrow stromal cells: Implications in transplantation. Transplantation 2003;75:389 –397. Moreau P, Faure O, Lefebvre S et al. Glucocorticoid hormones upregulate levels of HLA-G transcripts in trophoblasts. Transplant Proc 2001; 33:2277–2280.

See www.StemCells.com for supplemental material available online.