TISSUE-SPECIFIC STEM CELLS

Reciprocal Interactions Between Human Mesenchymal Stem Cells

and cd T Cells Or Invariant Natural Killer T Cells

IGNAZIA PRIGIONE,aFEDERICA BENVENUTO,

b,cPAOLA BOCCA,

aLUCA BATTISTINI,

dANTONIO UCCELLI,

b,c,e

VITO PISTOIAa

aLaboratory of Oncology, Istituto di Ricovero e Cura a Carattere Scientifico G. Gaslini, Genoa, Italy; bDepartment

of Neurosciences, Ophthalmology and Genetics, and cCenter of Excellence for Biomedical Research, University of

Genoa, Genoa, Italy; dNeuroimmunology Unit, European Centre for Brain Research, Santa Lucia Foundation,

Rome, Italy; eAdvanced Biotechnology Center, Genoa, Italy

Key Words. Human mesenchymal stem cells • T-cell receptor cd1 lymphocytes • Invariant natural killer T cells

ABSTRACT

The immunomodulatory activities of human mesenchymalstem cells (MSCs) provide a rational basis for their appli-

cation in the treatment of immune-mediated diseases, suchas graft versus host disease and multiple sclerosis. The

effects of MSCs on invariant natural killer T (iNKT) andcd T cells, both involved in the pathogenesis of autoim-mune diseases, are unknown. Here, we investigated the

effects of MSCs on in vitro expansion of these unconven-tional T-cell populations. MSCs inhibited iNKT

(Va241Vb111) and cd T (Vd21) cell expansion from pe-ripheral blood mononuclear cells in both cell-to-cell con-tact and transwell systems. Such inhibition was partially

counteracted by indomethacin, a prostaglandin E2 inhibi-

tor. Block of indoleamine 2,3-deoxygenase and transform-ing growth factor b1 did not affect Va241Vb111 and

Vd21 cell expansion. MSCs inhibited interferon-c produc-tion by activated Va241Vb111 and impaired CD3-medi-

ated proliferation of activated Va241Vb111 and Vd21 Tcells, without affecting their cytotoxic potential. MSCs didnot inhibit antigen processing/presentation by activated

Vd21 T cells to CD41

T cells. In contrast, MSCs werelysed by activated Vd21 T cells through a T-cell receptor-

dependent mechanism. These results are translationallyrelevant in view of the increasing interest in MSC-basedtherapy of autoimmune diseases. STEM CELLS 2009;27:693–702

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

INTRODUCTION

Mesenchymal stem cells (MSCs) were first identified in the late1960s as a rare bone marrow population (0.01%–0.001% ofnucleated cells) [1]. MSCs have been subsequently detected incord blood, fetal liver, and amniotic fluid as well as in adiposetissue, cartilage, and other adult connective tissues [2]. MSCspossess multilineage potential [3, 4] and have been exploitedfor tissue repair and gene therapy applications [5, 6].

MSCs modulate in vitro effector functions and proliferationof T, B, dendritic, and natural killer (NK) cells [7–10]. There-fore, MSCs have been proposed for the treatment of immune-mediated diseases. Le Blanc et al. [11] first reported the suc-cessful in vivo use of MSCs for graft versus host disease(GvHD) treatment in an allogeneic bone marrow transplant re-cipient. Moreover, MSCs have been shown to improve alloge-neic bone marrow or cord blood engraftment [12–16]. Thedemonstration that systemic administration of MSCs amelio-rated experimental autoimmune encephalomyelitis, a model of

multiple sclerosis (MS) [17], recently prompted their use forthe treatment of MS and other autoimmune diseases [10, 18].

To date, no information is available about the interaction ofMSCs with natural killer T (NKT) and cd T cells, two uncon-ventional T-cell populations bridging innate and adaptive im-munity [19, 20]. NKT and cd T cells share protective and regu-latory immune functions because they are involved in defenseagainst infectious organisms, tumor rejection, autoimmune dis-ease pathogenesis [21–25], and maintenance of transplant toler-ance [26]. Regarding the latter issue, host invariant (i) NKTcells have been shown to promote abrogation of acute GvHD inallogeneic hematopoietic stem cell transplantation, whereas do-nor iNKT cells favor the development of acute GvHD [27, 28].

iNKT cells represent a unique T-cell subset expressing aninvariant T-cell receptor (TCR) a chain (Va14-Ja18 in mice;Va24-Ja18 in humans) paired with a limited TCR Vb chainrepertoire (Vb11 in humans) together with NK cell-relatedmarkers. iNKT cells recognize glycolipidic antigens in thecontext of the major histocompatibility complex (MHC) classI-like molecule CD1d [29, 30].

Author contributions: I.P. and F.B.: conception and design, collection of data, data analysis and interpretation, manuscript writing; P.B.:collection of data; L.B.: conception and design, data analysis and interpretation; A.U. and V.P.: conception and design, data analysis andinterpretation, manuscript writing, financial support. I.P. and F.B. contributed equally as first authors to this work. A.U. and V.P.contributed equally as last authors to this work.

Correspondence: Ignazia Prigione, Ph.D., Laboratorio Scientifico di Oncologia, IRCCS G. Gaslini, Largo G. Gaslini 5, 16148 Genoa,Italy. Telephone: 39 010 5636342; Fax: 39 010 3779820; e-mail: ignaziaprigione@ospedale-gaslini.ge.it Received July 18, 2008;accepted for publication December 9, 2008; first published online in STEM CELLS EXPRESS December 18, 2008. VC AlphaMed Press1066-5099/2009/$30.00/0 doi: 10.1634/stemcells.2008-0687

STEM CELLS 2009;27:693–702 www.StemCells.com

A restricted TCR V region repertoire also characterizes cdT cells. In humans, only three Vd genes and six Vc genes areexpressed. Vc9þVd2þ cells represent the predominant cd cellsubset in human peripheral blood, where they account for1%–5% of all CD3þ T cells [31].

Vc9þVd2þ T cells recognize small nonpeptidic antigenssuch as microbe-derived metabolites of the isoprenoid biosyn-thesis pathway and alkylamines [32, 33]. Several natural andsynthetic molecules, such as, isopentenylpyrophosphate (IPP),induce in vitro proliferation of Vc9þVd2þ cells [34]. Antigenrecognition by Vc9þVd2þ T cells does not require presenta-tion by classic or nonclassic (CD1a, 1b, 1c, or 1d) MHC mol-ecules [35].

In the present study we investigated whether human MSCscould influence the in vitro expansion of iNKT (hereafterdefined as Va24þVb11þ) and cd T cells (hereafter defined asVd2þ) and interfere with effector functions of activatedVa24þVb11þ and Vd2þ cells.

MATERIALS AND METHODS

MSC Culture and Expansion

Human bone marrow aspirates were obtained from healthy donorsafter informed consent during explants for allogeneic hematopoi-etic stem cell transplantation. MSCs were isolated in culture byplastic adherence as described previously [36]. In brief, MSCswere cultured in human MesenCult basal medium enriched withits specific human supplement (Stem Cell Technologies, Vancou-ver, BC, Canada, http://www.stemcell.com). At each passage cellswere harvested with 0.05% trypsin and 0.02 EDTA (EuroClone,Milan, Italy, http://www.euroclonegroup.it) and seeded at the den-sity of 7 � 105cells in a 75-cm2 flask. The immunophenotype ofMSCs was evaluated by flow cytometry, and cells were used forexperimental purposes upon achievement of the typicalCD45�CD34�CD14�CD73þCD44þCD105þ phenotype.

The ability of MSCs to differentiate into adipogenic andosteogenic cells was assessed as described [3]. In brief, MSCswere cultured in adipogenic medium (Cambrex, Walkersville,MD, http://www.cambrex.com) for 3 weeks as well as in osteo-genic medium (Cambrex) for 2 weeks. Histochemical analysis ofcell layers was then performed by Oil Red Oil staining (Bioptica,Milan, Italy, http://www.bioptica.co.uk) for adipogenic differen-tiation and alizarin red dye staining for osteogenic differentiation(Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com). Arepresentative experiment of adipogenic and osteogenic differen-tiation of our MSC populations is shown in supporting informa-tion Figure 1. Staining with May-Grunwald Giemsa was per-formed for undifferentiated MSCs.

Conditioned media from MSC cultures at 70% of confluencewere obtained upon culturing cells in RPMI 1640 þ 10% fetalbovine serum (FBS) for 48 hours. Culture supernatants were thencollected and stored at �20�C until tested.

Lymphocyte Culture

Peripheral blood mononuclear cells (PBMCs) were obtained fromhealthy donors after informed consent and separated by Ficolldensity gradient centrifugation (Histopaque 1077; Sigma-Aldrich).Cells were resuspended in RPMI 1640 supplemented with L-glu-tamine, penicillin-streptomycin, nonessential amino acids (Euro-Clone), and 10% FBS (Gibco, Invitrogen Corporation, Carlsbad,CA, http://www.invitrogen.com) when used for experiments withTCR Vd2þ cell or 10% pooled human AB sera (complete me-dium) for Va24þVb11þ cells experiments. To expand TCR Vd2þ

cell lines, PBMCs were stimulated with IPP (30 lM; Sigma-Aldrich) in the presence of 50 U/ml recombinant (r) human (h)interleukin (IL)-2 (Proleukin; Chiron Italia s.r.l., Milan, Italy,http://www.chiron.com) [37]. Proliferating T cells were main-

tained in IL-2-containing medium. The immunophenotype of pro-liferating T cells was checked weekly using anti-TCR Vd2þ andCD3 monoclonal antibodies (mAbs). Cultures containing higherthan 95% TCR Vd2þ CD3þ cells were used for further experi-ments, at least 2 weeks after IPP stimulation.

To expand TCR Va24þVb11þ cell lines, PBMCs were stimu-lated with a-galactosylceramide (a-GalCer) (100 ng/ml; Axxora,LLC, San Diego, CA, http://www.axxora.com) in the presence of50 U/ml rhIL-2 [29]. After 10 days, TCR Va24þ cells were posi-tively selected using anti-Va24 mAb and goat anti-mouse IgG-conjugated microbeads (Miltenyi Biotec, Milan, Italy, http://www.miltenyibiotec.com). These cells were cultured in the pres-ence of irradiated feeder cells (50 Gy) and rhIL-2 for 14 days.Then TCR Vb11þ cells were sorted using anti-Vb11 mAb andgoat anti-mouse IgG-conjugated microbeads (Miltenyi Biotec).Cells were further cultured in the presence of rhIL-2 and irradi-ated feeder cells. The immunophenotype was assessed using anti-TCR Va24, anti-TCR Vb11, and CD3 mAbs. Cell lines contain-ing higher than 95% CD3þVa24þVb11þ cells were used for fur-ther experiments, at least 10 days after the last stimulation.

For coculture experiments, PBMCs or in vitro expanded Vd2þ

and Va24þVb11þ cells were plated with MSC at different ratios(1:1, 4:1, and 10:1). Allogeneic MSC populations from differentdonors were used. Vd2þ or Va24þVb11þ cell lines were culturedfor 48 hours in the presence of MSCs in complete medium withrIL-2. To analyze the effect of MSCs on the in vitro expansion ofVd2þ or Va24þVb11þ cells, PBMCs were stimulated with IPP ora-GalCer in the presence or in the absence of MSCs.

In some coculture experiments addressing the role of indole-amine 2,3-dioxygenase (IDO), prostaglandin E2 (PGE2), andTGF-b1 in MSC immunomodulation, the following specific inhib-itors were used: 1 mM 1-methyltryptophan (Sigma-Aldrich) forIDO, 5 lM indomethacin (Liometacen; Nile Pharmaceuticals,Cairo, Egypt) for PGE2, or 10 lg/ml anti-TGF-b1 mAb (R&DSystems, Inc., Minneapolis, MN, http://www.rndsystems.com).Control experiments showed that 1 mM 1-methyltryptophan or in-domethacin did not alter Vd2þ or Va24þVb11þ cell expansionfrom PBMCs in the absence of MSCs (data not shown).

After 7–10 days of culture, immunophenotypic analyses wereperformed to enumerate the proportions of CD3þ Va24þVb11þ

or Vd2þ T cells, respectively. In some experiments, PBMCs werestained with 5,6-carboxyfluorescein diacetate succinimidyl ester(CFSE ) (Molecular Probes, Invitrogen, Milan, Italy, http://probes.invitrogen.com) before being cocultured with MSC [38]. Inbrief, PBMCs undergoing in vitro expansion under specific stimuli(IPP and a-GalCer for Vd2 and Va24b11, respectively) werestained with CFSE 1.7 lM for 15 minutes at 37�C, washed withRPMI 1640 þ 10% FCS, counted, and then plated in cocultureexperiments. Transwell experiments were also performed withflat-bottom microtiter plates (Millipore, Milan, Italy, http://www.millipore.com), seeding MSCs in the upper chamber and PBMCsin the lower chamber.

The effect of MSCs on activated Va24þVb11þ or Vd2þ T-cell proliferation in response to polyclonal stimulation was inves-tigated. In brief, cells were seeded in the presence of OKT3 hy-bridoma supernatant (1:50 final dilution; American Type CultureCollection [ATCC], Manassas, VA, http://www.atcc.org) and irra-diated (50 Gy) autologous PBMCs with or without MSCs. After2 days of culture, cells were pulsed with 0.5 lCi [3H]thymidinefor 18 hours (5 Ci/mmol specific activity; GE Healthcare EuropeGmbH, Milan, Italy, http://www.gehealthcare.com) and then har-vested, and thymidine incorporation was measured.

For some experiments, cd T cells were freshly purified fromPBMCs by negative selection using the TCR cdþ T-cell isolation kit(Miltenyi).The effect of MSCs on purified cd T-cell proliferation inresponse to IPP was investigated. To this end, freshly isolated cd Tcells were stimulated with IPP in the presence of rIL-2 with or with-out MSCs. After 48 hours the frequency of Vd2þ cells in the activephases of the cell cycle was assessed by double staining with anti-Vd2 and anti-Ki67 mAbs and flow cytometric analysis.

694 Interaction of MSCs with iNKT and cd T Cells

Coculture experiments using Va24þVb11þ and Vd2þ celllines were also performed with two different fibroblasts cell lines:a human skin-derived cell line kindly provided by M. C. Mingari(Department of Experimental Medicine, University of Genoa,Genoa, Italy) and G. Pietra (Istituto Nazionale per la Ricerca sulCancro, Genoa, Italy) and the CRL-2106 human fibroblast cellline (ATCC). Both fibroblast cell lines were cultured in RPMI1640 þ 10% FCS and propagated upon trypsin-EDTA treatmentsimilarly to MSC cultures. In brief, Va24þVb11þ and Vd2þ celllines were cultured in IL-2-containing medium in the absence orpresence of fibroblasts at a 4:1 ratio. MSCs were added or not tothese cultures at a 1:1 fibroblast-MSC ratio. After 6 days of cul-

ture, the frequency of Va24þ and Vd2þ cells in the active phasesof the cell cycle was assessed by flow cytometry upon doublestaining with anti-Va24 or anti-Vd2 and anti-Ki67 mAbs.

Flow Cytometry

The following monoclonal antibodies were used: CD34 fluores-cein isothiocyanate (FITC), CD73 phycoerythrin (PE), and CD44FITC (BD Biosciences, San Jose, CA, http://www.bdbiosciences.-com), CD45 PE-Cy5 (Serotec Ltd., Oxford, U.K., http://www.ser-otec.com), CD105 (Ancell, Bayport, MN, http://www.ancell.com), anti-TCR Vd2 PE, and CD3 PE-Cy5/antigen-

Figure 1. MSCs inhibit Va24þVb11þ and Vd2þ cell expansion through the release of soluble factors. PBMCs from healthy donors were cocul-tured for 7-10 days with allogeneic MSCs in the presence of a-galactosylceramide (a-GalCer) or isopentenylpyrophosphate (IPP), and recombi-nant human interleukin-2 (rhIL-2). Experiments were performed both in cell-to-cell contact (A, B, left) and in transwell systems (A, B, right).(A): Proportion of Va24þVb11þ in the gated CD3þ cells. Results are expressed as mean percent � SD obtained from four independent experi-ments. (B): Proportion of CD3þ Vd2þ cells expressed as mean percent � SD obtained from four independent experiments. **, p < .01; *, p <.05. (C, D): PBMCs were stained with CFSE (1.7 lM) and stimulated with a-GalCer (100 ng/ml) (C) or IPP (30 lM) (D) in the presence ofrhIL-2 with or without MSC. After 6 days cells were harvested and cell proliferation was assessed by flow cytometry displaying CFSE dilutionof gated Va24þ (C) and Vd2þ (D) cells. Histogram plots from one representative experiment of three are shown. Abbreviations: CFSE, 5,6-car-boxyfluorescein diacetate succinimidyl ester; MSC, mesenchymal stem cell; PBMC, peripheral blood mononuclear cell; R1, undivided population.

Prigione, Benvenuto, Bocca et al. 695

www.StemCells.com

presenting cell (APC)/FITC (BD Pharmingen, San Diego, CA,http://www.bdbiosciences.com/index_us.shtml); anti-TCR Vb11FITC and anti-TCR Va24 PE (Immunotech, Marseille, France,http://www.beckmancoulter.com/products/pr_immunology.asp);CD80 FITC, CD86 FITC, and CD1d PE (BD Pharmingen).Cells were stained with specific monoclonal antibodies for30 minutes in the dark at 4�C, washed once with Dulbecco’sphosphate-buffered saline (D-PBS) (Sigma-Aldrich), and ana-lyzed by flow cytometry.

Flow cytometric analysis of apoptosis was performed byAnnexin V staining using the recombinant human Annexin VFITC kit (Bender MedSystems Diagnostics GmbH, Vienna, Aus-tria, http://www.bendermedsystems.com) as described elsewhere[39]. For intracellular cytokine staining, Va24þVb11þ or Vd2þ Tcells cultured alone or with MSCs for 48 hours were stimulatedwith phorbol 12-myristate 13-acetate (20 ng/ml) and ionomycin(250 ng/ml) in the presence of Brefeldin A (5 lg/ml) (Sigma-Aldrich) for 4 hours at 37�C and 5% CO2. Cells were surface-stained with anti-TCR Vd2 PE or anti-TCR Va24 PE mAbs for 30minutes in the dark at 4�C, fixed, and permeabilized with Cytofix/Cytoperm solutions (BD Pharmingen). Cell were then incubatedwith anti-IFN-c FITC or anti-tumor necrosis factor (TNF)-a FITC(BD Biosciences) for 30 minutes at 4�C in the dark, washed, andresuspended in D-PBS with 1% FBS for cytofluorimetric analysis.

Intracellular expression of Ki67 was analyzed using Cytofix/Cytoperm solutions and staining with anti-Ki67 FITC mAb fromDakoCytomation (Glostrup, Denmark, http://www.dakocytoma-tion.com). Flow cytometric analyses were performed by FACSCa-libur or FACSCanto cytometers (BD Biosciences), and data wereanalyzed by CellQuest or DIVA software (BD Biosciences),respectively.

Cytotoxicity Assays

Cytotoxicity assays were performed using the 4-hour 51Cr releasetest [40]. Va24þVb11þ and Vd2þ T-cell lines were cultured inrhIL-2-containing medium, either alone or in the presence ofMSCs for 48 hours, collected, and then used as effectors. TheJurkat T lymphoma cell line, either unpulsed or pulsed with a-GalCer, or the Daudi B lymphoma cell line was used as a target.Effector cells were seeded in duplicate at different concentrationsin V-bottom plates together with 5 � 103 51Cr-labeled targetscells in a final volume of 200 ll. Results were expressed as per-cent specific lysis [40].

Cytotoxic activity of Va24þVb11þ and Vd2þ T-cell lineswas assessed also against 51Cr-labeled allogeneic MSCs. In someexperiments, Vd2þ T cells were incubated with anti-NKG2DmAb (R&D Systems, Inc.) or anti-TCR Vc9 mAb (clone 7A5;Pierce Endogen, Rockford, IL, http://www.piercenet.com) or bothmAbs before being added to 51Cr-labeled MSCs.

Antigen Presentation Assays

In vitro expanded Vd2þ T cells were cultured overnight in thepresence of Candida albicans (2 � 106/ml autoclaved bodies) inrhIL-2-containing medium, either alone or in the presence ofMSCs at different Vd2þ T cell/MSC ratios. The negative controlwas represented by Vd2þ T cells cultured alone in rhIL-2-con-taining medium. Vd2þ T cells were then collected, extensivelywashed, and centrifuged on a Ficoll density gradient to removeextracellular Candida bodies. Vd2þ T cells were irradiated with12 Gy and used to stimulate freshly isolated autologous CD4þ Tlymphocytes. The latter cells were purified from frozen aliquotsof PBMCs as follows. PBMCs were first depleted of monocytesby centrifugation on a 50.6% Percoll density gradient, and CD4þ

cells were then sorted using an anti-CD4 mAb (Oxford Biomar-keting, Oxford, U.K.) and goat anti-mouse IgG-conjugatedmicrobeads (Miltenyi Biotec).

Antigen presentation assays were performed in U-bottom 96-well plates culturing antigen-pulsed Vd2þ T cells (as APC) to-gether with CD4þ T lymphocytes (responders) at a 1:4 cell ratioin complete medium (10% human serum) without exogenous

rhIL-2 for 6 days. Proliferation was assessed by [3H]thymidineincorporation in the last 18 hours of culture.

Statistical Analysis

The Mann-Whitney U test was applied to define the statisticalsignificance between groups. p < .05, p < .01, or p < .001 wasconsidered statistically significant.

RESULTS

Effects of MSC on Vd21 and Va241Vb111

Cell Expansion

Because of the low frequency of Vd2þ T cells andVa24þVb11þ T cells in the peripheral blood from healthyindividuals (2.3% and 0.3%, respectively; mean value inPBMCs from 10 healthy donors), these cell populations wereexpanded in culture from PBMCs for 7-10 days with or with-out MSCs at 1:1, 4:1, and 10:1 ratios, in the presence of IPPfor Vd2þ cells or a-GalCer for Va24þVb11þ cells.

Expansion of both Va24þVb11þ and Vd2þ cells was sig-nificantly inhibited by coculture with MSCs at all ratios tested(Fig. 1A, 1B, respectively, left panels; p < .01 and p < .05,respectively). We next investigated Vd2þ and Va24þVb11þ

expansion from PBMCs in the presence of MSCs in a trans-well system. Again, the expansion of both cell populationswas significantly inhibited at all cell ratios (p < .05), indicat-ing that soluble factors released by MSCs were sufficient tofully impair Va24þVb11þ and Vd2þ expansion (Fig. 1A, 1B,right panels).

Figure 1C and 1D shows a representative experiment inwhich a-GalCer-driven expansion of Va24þVb11þ cells andIPP-driven expansion of Vd2þ T cells, respectively, in the ab-sence or presence of MSCs were assessed by CFSE staining.The inhibition of cell proliferation at day þ6 is shown asnoncomplete dilution of CFSE of Va24þVb11þ or Vd2þ

gated cells into cell progenies. MSC-mediated inhibition ofproliferation was not due to the induction of target cell apo-ptosis as assessed by Annexin V staining of Va24þVb11þ

and Vd2þ gated cells (supporting information Fig. 2A, 2B,respectively).

Because the above results had been obtained using PBMCsas the source of Vd2þ and Va24þVb11þ cells, the possibilityexisted that inhibition of the expansion of these unconventionalT-cell populations was the consequence of indirect effects ofMSC on other cell types. To investigate this issue, cd T cellswere purified from PBMCs by negative selection and tested forIPP-induced cell proliferation by Ki67 staining in the absenceor presence of MSCs at the same ratios indicated above. Purifi-cation of Va24þVb11þ cells was not performed because oftheir paucity in peripheral blood. As shown in Figure 2, MSCsvirtually abolished Vd2þ cell proliferation at 1:1 and 1:4 ratios,with a partial recovery at a 1:10 ratio.

MSC-Mediated Suppression of Va241Vb111

and Vd21 Cell Expansion Involves PGE2

The role of soluble factors in the MSC-mediated inhibition ofVa24þVb11þ and Vd2þ cell expansion was next investigated.Because it has been demonstrated that PGE2, IDO, and TGF-b1, among many others, play an important role in MSC-medi-ated immunosuppressive functions [7, 41], we performedcoculture experiments in the presence of inhibitors of thesefactors. Figure 3 shows that MSC-mediated inhibition ofVa24þVb11þ and Vd2þ cell expansion was significantlycounteracted by indomethacin, an inhibitor of PGE2 synthesis

696 Interaction of MSCs with iNKT and cd T Cells

(p < .05). In contrast, inhibition of IDO and TGF-b1 wasineffective (Fig. 3). These data indicate that PGE2 plays a rel-evant role in the inhibition of both Va24þVb11þ and Vd2þ

cell expansion.

Effect of MSC on Effector Functions of ActivatedVa241Vb111 and Vd21 Cell Lines

To evaluate whether MSC can also affect effector functionsof in vitro expanded and activated unconventional T cells, weinvestigated the proliferative responses of Va24þVb11þ andVd2þ cell lines to a polyclonal stimulus. As shown in Figure4A, left panel, proliferation of Va24þVb11þ cells to CD3mAb was significantly reduced by MSCs at all ratios tested(p < .05), whereas a significant inhibition (p < .05) of Vd2þ

cell proliferation (Fig. 4 A, right panel) was detected only atthe 1:1 ratio.

We subsequently determined whether MSCs could inhibitthe potential alloreactive proliferation of Va24þVb11þ andVd2þ cell lines cocultured with allogeneic fibroblasts at a 4:1ratio and IL-2 (Fig. 4 B). MSCs were added to the cultures atthe same concentration as fibroblasts. In three differentexperiments, no alloreactive responses of Va24þVb11þ orVd2þ cells to fibroblasts was detected (see the representative

experiment in Fig. 4B). However, MSCs were found todampen the IL-2-dependent proliferation of Va24þVb11þ andVd2þ cell lines in the presence of skin fibroblasts. Such inhi-bition was not detected in cultures containing the CRL-2106fibroblast cell line that possessed a constitutive suppressiveactivity on Va24þVb11þ and Vd2þ cell proliferation. Thislatter finding is consistent with previous reports [42, 43].

We subsequently addressed the effects of MSCs on cyto-kine production (Fig. 5A, 5B). IFN-c production was signifi-cantly downregulated by MSCs in Va24þVb11þ cells (Fig. 5A, left panel, p < .05 at 1:1 and 4:1 ratios), whereas it wasnot affected in Vd2þ cells (Fig. 5B, left panel). Likewise,TNF production by both unconventional T-cell populationswas not inhibited by MSCs (Fig. 5A, 5B, right panels).

Next we investigated the capability of MSCs to alter thecytotoxic potential of activated Va24þVb11þ and Vd2þ

T-cell lines. Jurkat cells, pulsed with a-GalCer, and Daudicells, were used as targets for Va24þVb11þ and Vd2þ T-cellpopulations, respectively. Figure 6A shows the results of acytotoxicity assay representative of the four performed withVa24þVb11þ (left panel) or Vd2þ T cells (right panel). Asapparent, coculture of the latter cell fractions with MSCs atdifferent ratios did not significantly affect cytotoxic activity.

Figure 2. MSCs inhibit isopentenylpyrophosphate (IPP)-driven proliferation of freshly purified peripheral cd T cells. Circulating cd T cellswere freshly isolated as untouched cells by negative selection. Cells were stimulated with IPP (30 lM) in the presence of recombinant humaninterleukin-2 with or without MSCs for 48 hours and then collected and double stained with anti-Vd2 and anti-Ki67 monoclonal antibodies. His-tograms show the cytofluorimetric analysis of intracellular Ki67 expression in the gated Vd2þ cells. The percentage of Ki67 positive cells is indi-cated in each histogram. Abbreviations: Ctr, control; FITC, fluorescein isothiocyanate; MSC, mesenchymal stem cell.

Figure 3. Role of prostaglandin E2 in MSC-mediated inhibition of Va24þVb11þ and Vd2þ cell expansion. Peripheral blood mononuclear cells(PBMCs) from healthy donors were stimulated with a-galactosylceramide or isopentenylpyrophosphate in the presence of recombinant humaninterleukin-2 without or with MSCs at a 10:1 PBMC/MSC ratio for 7-10 days. Indom, 1-M-Trp, and a neutralizing anti-TGF-b1 monoclonal anti-body alone or in combinations were added in some wells. Results are expressed as percentage of Va24þVb11þ or Vd2þ cells in PBMCs cocul-tured with MSC in the presence or in the absence of Indom, 1-M-Trp, and anti-TGF-b1, with respect to the percentage of Va24þVb11þ or Vd2þ

cells expanded without MSC (100%). Data are expressed as mean percentage � SD obtained from five independent experiments. *, p < .05.Abbreviations: indom, indomethacin; MSC, mesenchymal stem cell; TGFb1, transforming growth factor-b1; Trp, tryptophan.

Prigione, Benvenuto, Bocca et al. 697

www.StemCells.com

Figure 4. MSCs modulate the proliferation of Va24þVb11þ and Vd2þ cell lines. (A): Va24þVb11þ (left) and Vd2þ (right) cell lines werestimulated with anti-CD3 monoclonal antibody (mAb) in the presence and in the absence of MSCs at different ratios for 72 hours. Proliferationwas assessed by [3H]thymidine incorporation in the last 18 hours. Left and right panels show the proliferative responses of Va24þVb11þ

and Vd2þ cells, respectively. Data are expressed as mean count per minutes � SD obtained from three independent experiments. *, p < .05.(B): In vitro expanded Va24þVb11þ (left) and Vd2þ (right) cells were cultured in recombinant interleukin-2-containing medium, alone orin the presence of two different fibroblast cell lines (human skin fibroblasts and CRL-2106 American Type Culture Collection cell line) withor without MSCs. After 6 days cells were harvested, and cell proliferation was assessed by double staining with anti-Va24- or anti-Vd2-PEand anti-Ki67-fluorescein isothiocyanate mAb followed by cytofluorimetric analysis. The percentage of Ki67-positive cells in the gatedVa24þ (left) or Vd2þ cells (right) is shown. Abbreviations: CTR, cell proliferation without stimulus; NKT, natural killer T; MSC, mesenchy-mal stem cell.

Figure 5. Effects of MSCs on IFN-c and TNF production by activated Va24þVb11þ and Vd2þ cell lines. Va24þVb11þ and Vd2þ cell lineswere cultured alone or in the presence of MSCs at different ratios for 48 hours. Afterward, cells were stimulated with phorbol 12-myristate 13-ac-etate þ ionomycin in the presence of Brefeldin A for 4 hours and analyzed for cytokine production by intracellular staining and flow cytometry.The proportions of Va24þVb11þ (A) and Vd2þ (B) cells producing IFN-c (left) and TNF (right) are shown. Results are expressed as mean per-centage � SD obtained from five independent experiments. *, p < .05. Abbreviations: IFN-c, interferon-c; MSC, mesenchymal stem cell; TNF,tumor necrosis factor.

698 Interaction of MSCs with iNKT and cd T Cells

MSCs Are Lysed by In Vitro Expanded Vd21

T-Cell Lines Through a TCR-Dependent Mechanism

The susceptibility of MSCs to unconventional T cell-mediatedcytotoxic activity was next investigated. In five differentexperiments, MSCs were not lysed by activated Va24þVb11þ

cells, consistently with the absence of CD1d, that is, therestriction element for Va24þVb11þ cell-mediated antigenrecognition, from the MSC surface (data not shown). In con-trast, Vd2þ T cells lysed MSCs (Fig. 6B, left). These resultswere consistently obtained by testing four Vd2þ T-cell linesagainst different MSC populations.

In the attempt to identify the receptor-ligand interactionsinvolved in the Vd2þ cell-operated killing of MSCs, we per-formed antibody-mediated blocking experiments. Vd2þ Tcells were preincubated with anti-Vc9 mAb, anti-NKG2DmAb or both before the addition of target cells. Figure 6B,right panel, shows the results of a representative experimentout of three performed with different MSC and Vd2þ T-celllines. mAb-mediated masking of TCR cd resulted in stronginhibition of lysis, whereas blocking of NKG2D had minimaleffects. Inhibition of cytotoxicity observed in the presence ofboth anti-TCR cd and anti-NKG2D mAbs was similar to thatdetected when the former mAb was tested alone (Fig. 6B,right panel). These results indicate that lysis of MSC by acti-vated Vd2þ cells is TCR-dependent.

Effects of MSC on Antigen Processingand Presenting Function of Vd21 Cells

Brandes et al. [44] demonstrated that activated Vd2þ cells actas professional APCs for CD4þ T cell-mediated responses.We next investigated whether MSCs could affect the APC

function of human Vd2þ T cells. In vitro expanded Vd2þ Tcells were pulsed overnight with Candida albicans bodies inthe presence or absence of MSC and then used as APCs forautologous CD4þ T lymphocytes.

Figure 7A shows pooled results from four differentexperiments indicating that antigen-pulsed Vd2þ T cellsinduced proliferation of CD4þ T cells and that the APC func-tion of the former cells was unaffected by incubation withMSC at three different ratios.

Accordingly, expression of the costimulatory moleculesCD80 and CD86 on activated Vd2þ T cells pulsed overnightwith antigen in the presence or absence of MSC was compa-rable (Fig. 7B). Taken together these findings suggest thatMSC did not affect the APC activity of Vd2þ T cells.

DISCUSSION

The functional properties and anatomic localization of iNKTand cd T cells put them on the border between innate andadaptive immunity. iNKT cells share protective and regula-tory functions with cd T cells [22–25].

Previous studies have shown that in vitro expanded MSCsexert broad-spectrum immunoregulatory functions on cells ofinnate immunity, such as dendritic cells, NK cells, and neu-trophils, and cells of adaptive immunity, that is, T and B cells[8–10, 45]. The aim of this study was to investigate the inter-actions between Va24þVb11þ and Vd2þ T cells and MSCsexpanded in vitro from human bone marrow in view of theincreasing interest for the use of MSCs in the treatment ofimmune-mediated disorders [18].

Figure 6. MSCs do not affect the cytotoxic activity of Va24þVb11þ and Vd2þ cells but are recognized and lysed by Vd2þ cells through a T-cell receptor-dependent mechanism. (A): Va24þVb11þ and Vd2þ cell lines were cultured alone (n) or in the presence of MSCs at differentratios (l, 1:1; ~, 4:1; ^, 10:1) for 48 hours. Afterward, cells were tested against target cells at different E:T ratios using standard 51Cr releaseassays. Left shows the cytotoxic activity of Va24þVb11þ cells against Jurkat cells pulsed with a-galactosylceramide. Right shows the cytotoxicactivity of Vd2þ cells against Daudi cells. One representative experiment of four is shown. (B): Left, Vd2þ cell lines were tested for cytotoxicactivity against allogeneic MSCs. Results are expressed as mean percentage of specific lysis � SD from three independent experiments. Right,Vd2þ cells were preincubated with medium (no mAb) or with an isotype control mouse IgG1 (isotype control), anti-NKGD2 mAb, anti-TCRVc9 mAb, or a combination of both mAbs before the addition of 51Cr-labeled MSCs. Results from one representative experiment of the three per-formed are shown. Abbreviations: CTR, control; MSC, mesenchymal stem cell.

Prigione, Benvenuto, Bocca et al. 699

www.StemCells.com

Primary MSCs are scanty cells localized in niches thatmay prevent encounters with immunocompetent cells [46].However, differentiated cells of the mesenchymal lineage,such as fibroblasts and chondrocytes, that are widely distrib-uted in the body, share with MSCs the ability to suppress pro-liferation of conventional T cells expressing ab TCRs [42, 43,47]. Thus, information gained using in vitro cultured MSCsmay be extended to their progeny. In this respect, we haveshown here that proliferation of Va24þVb11þ and Vd2þ cellswas damped by a fibroblast cell line.

Here, we show for the first time that MSCs abolish invitro proliferation of resting peripheral blood Va24þVb11þ

and Vd2þ cells stimulated by specific ligands through therelease of soluble factors. The proliferative responses of acti-vated Va24þVb11þ and Vd2þ cell lines to polyclonal stimuliwere also inhibited by MSCs but at a lower extent comparedwith the corresponding resting cell fractions. Accordingly, themajor effector functions of activated Va24þVb11þ and Vd2þ

cell lines, that is, cytokine production and cytotoxic activity,were only partially affected by incubation with MSCs.

PGE2 has been here identified as a relevant MSC-derivedsoluble inhibitor of Va24þVb11þ and Vd2þ cell proliferation.Previous studies have shown the involvement of PGE2 in dif-ferent immunoregulatory activities of MSC that express con-stitutively cyclooxygenase-1 and -2 [41, 45, 48]. Whereasmany other molecules may be involved in the MSC-mediatedmodulation of unconventional T-cell activities [48, 49], ourdata indicate that TGF-b1 and IDO do not play any role.

Previously it was shown that IL-2- or IL-15-activatedhuman NK cells lyse MSCs through the engagement of thecytotoxicity-activating receptors NKG2D and DNAM-1 onthe NK cell surface with the respective pairs of ligandsMICA/ULB-P3 and PVR/nectin-1 on the MSC surface [50].Because activated Va24þVb11þ and Vd2þ cells share cyto-toxic activity with NK cells, we investigated whether thesecells could lyse MSCs. Indeed, Vd2þ cells but notVa24þVb11þ cells were cytotoxic to MSCs. Blocking ofNKG2D on Vd2þ cells had minimal effects on MSC lysis. Incontrast, preincubation of Vd2þ cells with a blocking anti-TCR cd mAb resulted into strong inhibition of MSC lysis,suggesting the TCR-mediated recognition of an unknown en-dogenous ligand on target cells.

A novel function of human-activated Vd2þ cells is theirability to serve as professional APCs for naive CD4þ T-cellresponses [51]. Here we asked whether MSCs that have previ-ously been found to impair the antigen-presenting functionsof another professional APC, that is, the myeloid dendriticcell [52, 53], could exert a similar effect on Vd2þ cells.These experiments demonstrated that antigen processing andpresentation by the latter cells, as well as the expression ofthe CD80 and CD86 costimulatory molecules, were unaf-fected by incubation with MSCs.

An apparent paradox emerging from this study is, on oneside, the ability of MSCs to inhibit Vd2þ cell proliferationand, on the other side, their killing by activated Vd2þ cells.This seeming contradiction could be explained by the lower

Figure 7. The antigen-presenting function of activated Vd2þ T cells is unaffected by MSCs. (A): Activated Vd2þ cells were pulsed overnightwith Candida albicans bodies as antigen (cd p) in the absence (�) or in the presence of MSC at different ratios and used as antigen-presentingcells for autologous CD4þ lymphocytes (see Materials and Methods). Control cultures were Vd2þ (ctr cd) and CD4þ cells (ctr CD4) culturedalone. Proliferation was assessed by [3H]thymidine incorporation after 6 days of culture. Mean counts per minute � SD from four independentexperiments are shown. (B): Activated Vd2þ cells were pulsed overnight with C. albicans (cd p) in the absence or in the presence of MSC at a1:1 ratio and analyzed for CD80 and CD86 surface expression. Results are expressed as percentage of positive cells. One representative experi-ment out three performed is shown. Abbreviations: Ctr, control; MSC, mesenchymal stem cell.

700 Interaction of MSCs with iNKT and cd T Cells

susceptibility of the latter cells to MSC-mediated inhibition ofcell proliferation and by the high MSC/Vd2þ cell ratio neces-sary to achieve such inhibition. Therefore, functional impair-ment of activated Vd2þ cells by MSCs in vivo may be partialand insufficient to impair the cytotoxic activity of Vd2þ cellsversus MSCs. In this scenario, a microenvironment enrichedin proinflammatory cytokines such as TNF and IFN-c wouldpotentiate the inhibitory activity of MSC on Va24þVb11þ

and Vd2þ cells in vivo through the enhancement of PGE2production by the former cells [7, 41], whereas anti-inflamma-tory conditions would promote MSC elimination by Vd2þ

cells. This piece of information may be relevant in view ofthe clinical application of MSC-based cellular therapy for thetreatment of graft versus host disease and, possibly, autoim-mune diseases such as multiple sclerosis and rheumatoid ar-thritis [18, 54, 55].

ACKNOWLEDGMENTS

We are grateful to Dr. Francesca Gualandi (Dipartimento diEmatologia, Ospedale S. Martino, Genoa, Italy) for kindly sup-

plying the bone marrow samples. This study was supported bygrants from Progetto Integrato Ministero della Salute 2006‘‘Microambiente tumorale: ruolo nella progressione neoplasticaed effetti sulle difese dell’ospite. Identificazione di nuovi bersa-gli per lo sviluppo di terapie innovative’’ to V.P. FondazioneCarige, Genoa and Fondazione Maria Vilma e Bianca Querci toV.P; Ministero della Salute, Progetto di Ricerca Finalizzata 2006‘‘Caratterizzazione delle proprieta di immuno-modulazionedelle cellule staminali mesenchimali e possibile applicazione neltrattamento delle malattie autoimmuni’’ to V.P. and A.U.;MIUR–PRIN 2008: ‘‘Studio delle potenzialita terapeutiche dellecellule staminali neurali e mesenchimali nelle malattie degener-ative e infiammatorie del sistema nervoso centrale e periferico’’to A.U.; Fondazione CARIGE to A.U.; Fondazione Cariplo toV.P. and A.U. and Italian Foundation for Multiple Sclerosis:‘‘Stem Cell Project 2007-2009’’ to A.U.

DISCLOSURE OF POTENTIAL CONFLICTS

OF INTEREST

The authors indicate no potential conflicts of interest.

REFERENCES

1 Friedenstein AJ, Chailakhjan RK, Lalykina KS. The development offibroblast colonies in monolayer cultures of guinea-pig bone marrowand spleen cells. Cell Tissue Kinet 1970;3:393–403.

2 da Silva Meirelles L, Chagastelles PC, Nardi NB. Mesenchymal stemcells reside in virtually all post-natal organs and tissues. J Cell Sci2006;119:2204–2213.

3 Pittenger MF, Mackay AM, Beck SC et al. Multilineage potential ofadult human mesenchymal stem cells. Science 1999;284:143–147.

4 Porada CD, Zanjani ED, Almeida-Porad G. Adult mesenchymal stemcells: a pluripotent population with multiple applications. Curr StemCell Res Ther 2006;1:365–369.

5 Horwitz EM, Prockop DJ, Fitzpatrick LA et al. Transplantability andtherapeutic effects of bone marrow-derived mesenchymal cells in chil-dren with osteogenesis imperfecta. Nat Med 1999;5:309–313.

6 Ozawa K, Sato K, Oh I et al. Cell and gene therapy using mesenchy-mal stem cells (MSCs). J Autoimmun 2008;30:121–127.

7 Krampera M, Cosmi L, Angeli R et al. Role for interferon-c in theimmunomodulatory activity of human bone marrow mesenchymalstem cells. Stem Cells 2006;24:386–398.

8 Nauta AJ, Fibbe WE. Immunomodulatory properties of mesenchymalstromal cells. Blood 2007;110:3499–3506.

9 Uccelli A, Moretta L, Pistoia V. Immunoregulatory function of mesen-chymal stem cells. Eur J Immunol 2006;36:2566–2573.

10 Uccelli A, Moretta L, Pistoia V. Mesenchymal stem cells in healthand disease. Nat Rev Immunol 2008;8:726–736.

11 Le Blanc K, Rasmusson I, Sundberg B et al. Treatment of severeacute graft-versus-host disease with third party haploidentical mesen-chymal stem cells. Lancet 2004;363:1439–1441.

12 Koc ON, Gerson SL, Cooper BW et al. Rapid hematopoietic recoveryafter coinfusion of autologous-blood stem cells and culture-expandedmarrow mesenchymal stem cells in advanced breast cancer patientsreceiving high-dose chemotherapy. J Clin Oncol 2000;18:307–316.

13 Noort WA, Kruisselbrink AB, in’t Anker PS et al. Mesenchymal stemcells promote engraftment of human umbilical cord blood-derivedCD34þ cells in NOD/SCID mice. Exp Hematol 2002;30:870–878.

14 Le Blanc K, Samuelsson H, Gustafsson B et al. Transplantation ofmesenchymal stem cells to enhance engraftment of hematopoieticstem cells. Leukemia 2007;21:1733–1738.

15 Ball LM, Bernardo ME, Roelofs H et al. Cotransplantation of ex vivoexpanded mesenchymal stem cells accelerates lymphocyte recoveryand may reduce the risk of graft failure in haploidentical hematopoi-etic stem-cell transplantation. Blood 2007;110:2764–2767.

16 Le Blanc K, Frassoni F, Ball L et al. Mesenchymal stem cells fortreatment of steroid-resistant, severe, acute graft-versus-host disease: aphase II study. Lancet 2008;371:1579–1586.

17 Zappia E, Casazza S, Pedemonte E et al. Mesenchymal stem cellsameliorate experimental autoimmune encephalomyelitis inducing T-cell anergy. Blood 2005;106:1755–1761.

18 Uccelli A, Frassoni F, Mancardi G. Stem cells for multiple sclerosis:promises and reality. Regen Med 2007;2:7–9.

19 Godfrey DI, Kronenberg M. Going both ways: immune regulation viaCD1d-dependent NKT cells. J Clin Invest 2004;114:1379–1388.

20 Holtmeier W, Kabelitz D. cd T cells link innate and adaptive immuneresponses. Chem Immunol Allergy 2005;86:151–183.

21 Zocchi MR, Poggi A. Role of cd T lymphocytes in tumor defense.Front Biosci 2004;9:2588–2604.

22 Bach JF, Bendelac A, Brenner MB et al. The role of innate immunityin autoimmunity. J Exp Med 2004;200:1527–1531.

23 Battistini L, Caccamo N, Borsellino G et al. Homing and memory pat-terns of human cd T cells in physiopathological situations. MicrobesInfect 2005;7:510–517.

24 Bonneville M, Scotet E. Human Vc9Vd2 T cells: promising new leadsfor immunotherapy of infections and tumors. Curr Opin Immunol2006;18:539–546.

25 Yamamura T, Sakuishi K, Illes Z et al. Understanding the behavior ofinvariant NKT cells in autoimmune diseases. J Neuroimmunol 2007;191:8–15.

26 Pillai AB, George TI, Dutt S et al. Host NKT cells can prevent graft-versus-host disease and permit graft antitumor activity after bone mar-row transplantation. J Immunol 2007;178:6242–6251.

27 Lan F, Zeng D, Higuchi M et al. Host conditioning with total lymph-oid irradiation and antithymocyte globulin prevents graft-versus-hostdisease: the role of CD1-reactive natural killer T cells. Biol BloodMarrow Transplant 2003;9:355–363.

28 Morris ES, MacDonald KP, Rowe V et al. NKT cell-dependent leuke-mia eradication following stem cell mobilization with potent G-CSFanalogs. J Clin Invest 2005;115:3093–3103.

29 Kawano T, Cui J, Koezuka Y et al. CD1d-restricted and TCR-medi-ated activation of Va14 NKT cells by glycosylceramides. Science1997;278:1626–1629.

30 Brigl M, Brenner MB. CD1: antigen presentation and T cell function.Annu Rev Immunol 2004;22:817–890.

31 O’Brien RL, Roark CL, Jin N et al. cd T-cell receptors: functionalcorrelations. Immunol Rev 2007;215:77–88.

32 Bukowski JF, Morita CT, Brenner MB. Human cd T cells recognizealkylamines derived from microbes, edible plants, and tea: implica-tions for innate immunity. Immunity 1999;11:57–65.

33 Constant P, Davodeau F, Peyrat MA et al. Stimulation of human cdT cells by nonpeptidic mycobacterial ligands. Science 1994;264:267–270.

34 Tanaka Y, Morita CT, Tanaka Y et al. Natural and synthetic non-peptideantigens recognized by human cd T cells. Nature 1995;375:155–158.

35 Morita CT, Beckman EM, Bukowski JF et al. Direct presentation ofnonpeptide prenyl pyrophosphate antigens to human cd T cells. Immu-nity 1995;3:495–507.

36 Corcione A, Benvenuto F, Ferretti E et al. Human mesenchymal stemcells modulate B-cell functions. Blood 2006;107:367–372.

37 Cipriani B, Knowles H, Chen L et al. Involvement of classical andnovel protein kinase C isoforms in the response of human Vc9Vd2 Tcells to phosphate antigens. J Immunol 2002;169:5761–5770.

Prigione, Benvenuto, Bocca et al. 701

www.StemCells.com

38 Glennie S, Soeiro I, Dyson PJ et al. Bone marrow mesenchymal stem cellsinduce division arrest anergy of activated T cells. Blood 2005;105:2821–2827.

39 Benvenuto F, Ferrari S, Gerdoni E et al. Human mesenchymal stem cellspromote survival of T cells in a quiescent state. Stem Cells 2007;25:1753–1760.

40 Prigione I, Chiesa S, Taverna P et al. T cell mediated immuneresponses to Toxoplasma gondii in pregnant women with primary tox-oplasmosis. Microbes Infect 2006;8:552–560.

41 Aggarwal S, Pittenger MF. Human mesenchymal stem cells modulateallogeneic immune cell responses. Blood 2005;105:1815–1822.

42 Haniffa MA, Wang XN, Holtick U et al. Adult human fibroblasts arepotent immunoregulatory cells and functionally equivalent to mesen-chymal stem cells. J Immunol 2007;179:1595–1604.

43 Jones S, Horwood N, Cope A et al. The antiproliferative effect ofmesenchymal stem cells is a fundamental property shared by all stro-mal cells. J Immunol 2007;179:2824–2831.

44 Brandes M, Willimann K, Moser B. Professional antigen-presentationfunction by human cd T cells. Science 2005;309:264–268.

45 Rasmusson I. Immune modulation by mesenchymal stem cells. ExpCell Res 2006;312:2169–2179.

46 Shiozawa Y, Havens AM, Pienta KJ et al. The bone marrow niche:habitat to hematopoietic and mesenchymal stem cells, and unwittinghost to molecular parasites. Leukemia 2008;22:941–950.

47 Bocelli-Tyndall C, Barbero A, Candrian C et al. Human articularchondrocytes suppress in vitro proliferation of anti-CD3 activated pe-ripheral blood mononuclear cells. J Cell Physiol 2006;209:732–734.

48 Spaggiari GM, Capobianco A, Abdelrazik H et al. Mesenchymal stemcells inhibit natural killer-cell proliferation, cytotoxicity, and cytokineproduction: role of indoleamine 2,3-dioxygenase and prostaglandin E2.Blood 2008;111:1327–1333.

49 Sotiropoulou PA, Perez SA, Gritzapis AD et al. Interactions betweenhuman mesenchymal stem cells and natural killer cells. Stem Cells2006;24:74–85.

50 Spaggiari GM, Capobianco A, Becchetti S et al. Mesenchymal stemcell-natural killer cell interactions: evidence that activated NK cellsare capable of killing MSCs, whereas MSCs can inhibit IL-2-inducedNK-cell proliferation. Blood 2006;107:1484–1490.

51 Moser B, Brandes M. cd T cells: an alternative type of professionalAPC. Trends Immunol 2006;27:112–118.

52 Jiang XX, Zhang Y, Liu B et al. Human mesenchymal stem cells in-hibit differentiation and function of monocyte-derived dendritic cells.Blood 2005;105:4120–4126.

53 Jung YJ, Ju SY, Yoo ES et al. MSC-DC interactions: MSC inhibitmaturation and migration of BM-derived DC. Cytotherapy 2007;9:451–458.

54 Dazzi F, Horwood NJ. Potential of mesenchymal stem cell therapy.Curr Opin Oncol 2007;19:650–655.

55 Tyndall A, Walker UA, Cope A et al. Immunomodulatory proper-ties of mesenchymal stem cells: a review based on an interdiscipli-nary meeting held at the Kennedy Institute of RheumatologyDivision, London, U.K., 31 October 2005. Arthritis Res Ther 2007;9:301.

See www.StemCells.com for supporting information available online.

702 Interaction of MSCs with iNKT and cd T Cells