Evidence for the Involvement of Galectin-3 in Mesenchymal Stem Cell Suppression of Allogeneic T-Cell Proliferation

Authors


M. Sioud, Department of Immunology, The Institute for Cancer Research, Rikshospitalet-Radiumhospitalet Medical center, Montebello N-310, Oslo, Norway. E-mail: mosioud@medisin.uio.no

Abstract

Human bone marrow-derived mesenchymal stem cells (MSC) are multipotent non-hematopoietic progenitors that have regulatory activity on immune cells. NOD- and Toll-like receptors (NLR, TLR) have several roles in immunity, including those relevant to pathogen recognition and shaping the course of immune responses by controlling gene expression. We have shown that these innate immune receptors are expressed by hematopoietic CD34+ progenitors and MSC. To uncover genes critical in MSC function, first we have used microarray to screen for potential transcripts whose levels are altered in response to NOD-1 and TLR-2 activation, and second we validated some candidate genes using real-time RT-PCR, Western blots and cellular assays. Amongst the altered genes, galectin-3 was upregulated at both mRNA and protein levels in response to TLR-2 activation. Interestingly, MSC secreted galectin-3, a protein known to modulate T-cell proliferation, gene expression, cell adhesion and migration. Knockdown of galectin-3 in MSC using small interfering RNA (siRNA) reduced the immunosuppressive effect of MSC on mixed lymphocyte cultures when compared to cells treated with an irrelevant siRNA (< 0.05). Collectively, the data emphasize a new role of galectin-3 in the immunomodulatory function of MSC and indicate that NOD signalling pathway is also functional in these cells.

Introduction

Mesenchymal stem cells (MSC), also known as marrow stromal cells, are a self-renewing population of multipotent cells present in bone marrow and many other adult tissues [1, 2]. Ex-vivo expanded MSC obtained from different species, including human have been shown to give rise to a variety of cell types including myocytes, adipocytes, fibroblasts, endothelial cells and osteoblasts [1, 2]. Moreover, they are capable of suppressing the activity of a broad range of immune cells, including T cells, antigen-presenting cells, natural killer cells and B cells [3, 4]. Recent studies have also shown that MSC infusion can reduce the incidence of graft-versus-host disease (GvHD) after allogeneic HSC transplantation in humans, and can be used to treat severe acute GvHD refractory to conventional immunosuppressive therapy [5, 6]. Although several studies were performed on the possible role of MSC in tissue regeneration and immunosuppression, the primary mechanisms involved in the MSC-mediated suppressive activity on immune cells and the role of MSC-derived stromal cells in normal lymphoid development are still partially unknown.

Given the role played by Toll-like receptors (TLR) in innate and adaptive immunity [7, 8], we have previously asked whether these receptors are expressed by hematopoietic CD34+ progenitor cells and MSC. We have shown that TLR and associated signalling adaptor molecules are expressed by CD34+ progenitors and TLR activation induced their differentiation into monocytes and dendritic cells capable of priming T cells [9, 10]. Similarly, mouse hematopoietic progenitors expressed functional TLR whose activation induced cell differentiation into monocytes and DCs [11]. Furthermore, we and others have reported on the expression of TLR by MSC [12–14]. Activation of TLR-3 and TLR-4 on MSC affected their immunosuppressive function on T cells, once more suggesting a novel role of TLR in stem cell function [13].

In addition to TLR, we have found that NOD-like receptors (NLR), a new family of intracellular bacterial sensors, are expressed by BM CD34+ progenitors [14]. Through their C-terminal LRR motif, NOD-1 and NOD-2 recognize active entities of bacterial peptidoglycan resulting in the activation of caspase and NF-κB signalling pathways. The core structure of the ligand recognized by NOD-1 is the peptidoglycan-specific dipeptide, γ-D-glutamyl-meso-diaminopimelic acid (iE-DAP) and NOD-2 recognizes the muramyldipeptide (MDP), representing the minimal motif of bacterial peptidoglycan able of activating NOD2 [15]. Given the significance of TLR and NLR in immunity and cell differentiation, in this study we explored the expression of NLR in MSC, the transcriptional response of MSC to NOD-1 and TLR-2 ligands and the ability of galectin-3, an identified candidate gene, to affect the inhibitory function of MSC on T-cell proliferation to alloantigens.

Materials and methods

Reagents

The peptidoglycan-specific dipeptide, γ-D-glutamyl-meso-diaminopimelic acid (iE-DAP, a NOD1 ligand) and control peptide (iE-Lys) were purchased from InvivoGen (Toulouse, France) Pam3CS(K)4, and a TLR2 ligand was purchased from Calbiochem (La Jolla, CA, USA). Conjugated anti-CD14, anti-CD4 were purchased from DakoCytomation (Copenhagen, Denmark). Conjugated anti-CD34, anti-CD105, anti-CD106 and anti-NOD2 monoclonal antibody (2D9) were purchased from BD Biosciences (Franklin Lakes, NJ, USA). Anti-NOD1 polyclonal antibodies were purchased from Cell Signalling (Danvers, MA, USA). Total RNA isolation kit Trizol and cDNA synthesis kit were purchased from Invitrogen (San Diego, CA, USA) and GE Healthcare AS (Oslo, Norway), respectively. SYBR Green PCR Master Mix was purchased from Applied Biosystems (Foster City, CA, USA). An Illumina TotalPrep RNA Amplification Kit was purchased from Ambion (Austin, TX, USA). Expression arrays were purchased from Illumina (San Diego, CA, USA). Human VEGF monoclonal antibody (clone 26503, capture antibody), human VEGF 165 biotinylated affinity purified polyclonal antibodies (detection antibody) and the galectin ELISA kit were purchased from R&D systems (Abingdon, UK).

Isolation of MSC from bone marrow

MSC were isolated and expanded from bone marrow (BM) taken from iliac crest of adult volunteers with informed consent. Heparinized BM was mixed with double volume of phosphate-buffered saline, and mononuclear cells were prepared by gradient centrifugation (Lymphoprep). Subsequently, the cells were cultured in 75-cm2 flask at a concentration of 30 × 106 per 20 ml Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal calf serum (FCS). Cultures were incubated at 37 °C in a humidified atmosphere containing 5% CO2. After 48- to 72-h incubation, non-adherent cells were removed and adherent cells constituted the MSC cell population that was expanded. Cells were detached by a treatment with trypsin and EDTA (GibcoBRL, Grand Island, NY, USA) and replated at a density of 106 cells/75 cm2 flask. These cells were verified for positive staining for CD105 and CD106, and are negative for CD14, CD34 and CD4 markers.

Cell staining

MSC were detached using trypsin/EDTA, resuspended in complete medium and placed at 37 °C for 2 h. Subsequently, cell aliquots (5 × 105) were incubated on ice with conjugated monoclonal antibodies against CD34, CD14, CD4, CD105 and CD106. The cells were then washed with PBS containing 1% FCS and 0.1% sodium azide (FACS buffer) for 1 h at 4 °C, resuspended in 300 μl of FACS buffer and then analysed by flow cytometry. The data were analysed with CellQuest software (Becton Dickinson, San Jose, CA, USA).

Cell treatments and cytokine assays

MSC were seeded in a 6-well plate at 5 × 103/cm2 in DMEM containing 10% FCS. After overnight incubation, the medium was replaced with DMEM supplemented with 10% FCS with or without TLR2 [Pam3CS(K)4, 10 μg/ml] or NOD1 ligand (iE-DAP, 10 μg/ml). After 18 h of incubation, culture supernatants were collected and cytokine levels were measured by ELISA according to the manufacturer’s instructions.

Allogeneic mixed leucocyte reaction

Human peripheral blood mononuclear cells (PBMCs) were prepared by density gradient centrifugation (Lymphoprep) from buffy coats obtained from healthy adult donors. Cells were washed and then resuspended in RPMI-1640 medium containing 10% fetal calf serum (FCS) and antibiotics. To study the effect of MSC on T-cell activation, mixed lymphocyte reaction (MLR) assays were performed in the presence of irradiated allogeneic MSC. The cells were cocultured in 96-well U-bottom microtiter plates for 5 days. T-cell proliferation was evaluated by incubating cells with [3H]-thymidine for additional 16 h. Cells were harvested, and 3H- thymidine uptake was measured. All experiments were run in triplicate.

Western blots

Total protein lysates (30–60 μg) were resolved on 10% SDS–polyacrylamide gels and subsequently transferred to nitrocellulose by electrophoresis. Membranes were blocked with 5% non-fat dried milk in PBS containing 0.1% Tween overnight. Subsequent to washing, membranes were incubated with antibodies against the selected proteins, followed by HRP-conjugated rabbit or mouse secondary antibodies. Antibody–protein complexes were visualized after exposure to X-ray film by enhanced chemiluminescence reagent. To control for protein loading, the blots were stripped and reprobed with anti-β actin polyclonal antibodies (Santa Cruz Biotech, Santa Cruz, CA, USA).

Gene expression analysis

MSC (3 × 106 cells per sample) were treated with TLR-2 [Pam3CS(K)4; 10 μg/ml] or NOD-1 ligand (iE-DAP, 10 μg/ml) for 18 h. Subsequently, they were harvested and total RNA was prepared from controls and treated cells. Each treatment was performed in triplicates, and cells were collected prior total RNA preparation. Control cells were treated with a control peptide (iE-Lys). Total RNA (500 ng per sample) was used to generate complementary biotin UTP-labelled DNA using the Illumina TotalPrep RNA Amplification Kit. Around 1.5 μg of labelled transcripts were used for hybridization to an array according to the Illumina Sentrix humanref-6 beadchip protocol. Following hybridization, the samples were washed and scanned with a BeadArray Reader (Illumina). Expression values were extracted and normalized by the BeadStudio software.

Transfection

Freshly isolated human monocytes were transfected with siRNA using the BTX electroporation apparatus as described previously [16]. Briefly, 2–5 μg of siRNA were mixed with 5 × 106 cells in 0.5 ml RPMI medium, and then the cells were transferred to 4-mm BTX cuvettes and pulsed at 500 V for 2 ms. After electroporation, the cells were diluted in 2.5 ml of prewarmed medium, and incubated at 37 °C in 5% CO2. Gal-3 expression was assayed with Western blots 18 h post-transfection time. The target sequences of the used siRNA are the following: siRNA-1 5′-GCUCCAUGAUGCGUUAUCU-3′; siRNA-2 5′-GAGAGUCAUUGUUUGCAAU-3′; siRNA-3 5′-GUCUGGGCAUUCUGAUGUU-3′; Control siRNA 5′-UUGAUGUGUUUAGUCGCUA-3′.

RT- and real-time-PCR

Total RNA was isolated from MSC using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Complementary DNA was synthesized from 5 μg of total RNA using the first-strand cDNA synthesis kit and oligo-dT primer in 15 μl volume according to manufacturer’s (GE Healthcare). PCR was conducted in 50 μl on 1/30 on the cDNA using 2.5 units of Tap polymerase. RT-PCR products were separated on 1.5% agarose gels, visualized by staining with SYBR® safe DNA gel stain (Invitrogen) and photographed using the 2UV Transilluminator BioDoc-ItTM Imaging system (AH diadognostic).

The following primers were used to amplify the investigated genes:

  • NOD-1 forward, 5′-GTACGTCACCAAAATCCTGGA-3′;

    • reverse, 5′-CAGTCCCCTTAGCTGTGATC-3′;

  • NOD-2 forward,5′-CTGGCAAAGAACGTCATGCTA-3′;

    • reverse, 5′-CCTGGGATTGAATCTTGGGAA-3′;

  • VEGFA forward, 5′-GAGGAGGAAGAAGAGAAGGAAG-3′;

    • reverse, 5′-TTGGCATGGTGGAGGTAGAG-3′;

  • GAL-3 forward, 5′-CTGAGTAGCGGGAAGTGCGGTA-3′;

    • reverse, 5′-CAGGCCATCCTTGAGGGTTTGG-3′;

  • EPHB-1* forward, 5′-CAGGAAACGGGCTTATAGCA-3′;

    • reverse, 5′-CTCAGCCAGGTACTTCATGC-3′;

  • Gal-3* forward 5′-CTTCCCCTTGATCAGCTCCA-3′;

    • reverse, 5′-CTGGGCCTTTTGGTGAAAGG-3;

  • VEGFA* forward 5′-CTCGGGCCGGGGAGGAAGA-3′

    • reverse 5′- GCAGGGCACGACCGCTTACC-3′

  • SQSTM* (P62) forward, 5′-CTCTGGCGGAGCAGATGAGGA-3′;

    • reverse, 5′-CCAGCCGCCTTCATCAGAGA-3′;

  • NOTCH-1* forward, 5′-AGCTCGTCCCCGCATTCCAA-3′;

    • reverse, 5′-AGGCAGGTGATGCTGGTGGA-3′;

  • CXCL-10* forward, 5′-CAAGCCAATTTTGTCCACGT-3′;

    • reverse, 5′-GTAGGGAAGTGATGGGAGAG-3′;

  • DGCR-8* forward, 5′-TCATGCATCGTGCACCACAG-3′;

    • reverse, 5′-CTGCACCACTGTCCACAGTC-3′;

  • IRAK-2*, forward 5′-GGCCCCAGCGTGTCAGCATC-3′

    • reverse 5′-AGCTGCCCCACCCGGATGAA-3′

  • TRAF-7*, forward 5′-GCGGTGTCCCAACAACCCCA-3′

    • reverse, 5′-AGCGGTCATCCGTCTGCTGC-3′

  • β actin forward, 5′-ATCTGGCACCACACCTTCTAC-3′;

    • reverse, 5′-CGTCATACTCCTGCTTGCTGATC-3′.

In addition to standard RT-PCR, gene expression was analysed by real-time RT-PCR using specific primers for the selected genes and SYBR Green PCR Master Mix (Applied Biosystems). For each sample, comparative threshold (Ct) difference between control and treated cells were calculated. The fold difference for each gene was calculated using the delta-delta Ct method [17]. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal reference gene. Primers indicated with asterisks were used in real-time RT-PCR.

Statistical analysis

Statistical significance was determined by a two-tailed unpaired Student’s t-test. P values of <0.05 were considered to indicate statistical significance.

Results

Expression of NOD1 by BM MSC

We initially evaluated the expression of NOD-1 and NOD2- in human BM-derived MSC by RT-PCR. As shown in Fig. 1A, the in vitro expanded BM MSC showed a homogenous cell population with fibroblast like cells. In addition, they were uniformly negative for markers of the haematopoietic lineage, including CD34, CD14 and CD4, and positive for CD105 (endoglin) and CD106 (vascular cell adhesion molecule 1) (Fig. 1B). RT-PCR analysis revealed the transcription of NOD-1, but not NOD-2 gene (Fig. 1C, as a representative example). To further support the RT-PCR data, protein extracts from MSC were analysed by Western blots using a monoclonal antibody against NOD-1. Consistent with the RT-PCR data, MSC expressed NOD1 protein (Fig. 1D).

Figure 1.

 MSC express NOD1. (A) A light picture of MSC showing spindle-shaped fibroblast morphology after culture expansion ex vivo. (B) Representative flow cytometry analysis of cultured MSC with monoclonal antibodies against CD34, CD14, CD4, CD105, and CD106. (C) Expression of NOD-1, NOD-2, and β actin (Ac) in MSC was determined by RT-PCR. Western blot analysis of NOD1 expression (D). Results are representative of 1 of 3 donors.

Screening of candidates genes that are under the NOD1 signalling pathway

NOD1 senses the iE-DAP dipeptide which is found in peptidoglycan of all gram-negative and certain gram-positive bacteria whereas NOD-2 recognizes the muramyl dipeptide (MDP) structure found in almost all bacteria [17]. First, we have used microarray to screen for potential transcripts whose levels may be affected by NOD-1 activation. Cells were treated overnight with iE-DAP dipeptide, a specific ligand for NOD-1. We also evaluated the response to Pam3CS(K)4, a prototypic TLR-2 ligand. Gene expression was normalized to cells treated with a control peptide (iE-Lys). Around 800 and 200 genes were altered by TLR2 and NOD-1 ligands, respectively. Amongst the altered genes, VEGFA, NOTCH-1, TRAF-7, DGCR-8, EPHB-1 receptor, CD9, SQSTM-1, CXCL-10, IRF-7 and galectin-3 (Gal-3) were significantly changed in response to NOD-1 and TLR-2 signalling.

To validate the microarray data, initially, a set of primers specific for human vascular endothelial growth factor A (VEGFA), Gal-3, and EPHB-1 receptor (EPHB1) were used in reverse transcription (RT-PCR) analyses to establish their expression in MSC. VEGF-A is called just VEGF because it is the most important VEGF members. In agreement with the array data, Fig. 2A shows the upregulation of VEGF and Gal-3, and downreglation of EPH B1 receptor in response to TLR-2 or NOD-1 ligand. A set of upregulated and downregulated genes were also assessed by real-time RT-PCR (Fig. 2B). Almost all analysed genes were significantly altered in response to TLR-2 or NOD-1 activation. The upregulation of Gal-3 and DGCR-8 was also validated by Western blots using specific antibodies (Fig. 3A and B).

Figure 2.

 Validation of the microarray data using RT-PCR. (A) Total RNA prepared from stimulated and stimulated MSC was subjected to RT-PCR and amplified with VEGFA, Gal-3 or EPBH-1-specific primers. (B) Relative expression levels of certain genes in MSC in response to NOD-1 and TLR-2 activation. The relative levels of mRNA for each gene were generated from duplicate quantitative RT-PCR reactions following normalization to GAPDH. The bars show the relative expression of each gene vs GAPDH in cells treated for 18 h with control peptide, NOD-1 ligand or TLR-2 ligand. *< 0.05; **< 0.01 (stimulated versus unstimulated cells).

Figure 3.

 Western blot analysis. The expression of Gal-3 (A) and SQSTM1 (B) in stimulated and unstimulated MSC was assessed by Western blots. (C) The autoradiographs were quantified by densitometry. (D) MSC secrete VEGF and Gal-3 proteins. The cells were stimulated with TLR2 ligand (Pam3Cys) or NOD1 ligand (iE-DAP dipeptide) for 18 h. Subsequently, VEGF and Gal-3 proteins were measured in culture supernatants by ELISA. The results represent the means ± SD of triplicate wells. *< 0.05; **< 0.01; ***< 0.001 (stimulated versus unstimulated cells).

Gal-3 is secreted by BM MSC

Gal-3 is a member of a large family of β-galactoside-binding animal lectins [18]. It is expressed in a variety of tissues and cell types, and is localized mainly in the cytoplasm, although, depending on the cell types and proliferative states, a significant amount of this lectin can be detected in the nucleus, on the cell surface or in the extracellular environment [18]. Therefore, in the next experiment we evaluated Gal-3 levels in culture supernatants by ELISA (Fig. 3D). BM MSC constitutively secreted Gal-3 and VEGF. The production of Gal-3 significantly increased when the cells were stimulated with TLR-2 ligand but not NOD-1 ligand (< 0.05). In contrast, both ligands increased the VEGF levels (Fig. 3D).

Gene silencing of Gal-3 in MSC

Previous studies have suggested the possible involvement of Gal-3 in diverse physiological and pathological processes, including pre-mRNA splicing, neoplastic transformation and immune response [18]. Gal-3 is also reported to play a negative role in T-cell activation, a process that requires clustering of a threshold number of T-cell receptor at the site of antigen presentation [19, 20]. Based on these early findings, we investigated the potential effect of Gal-3 gene silencing in MSC on T-cell proliferation to alloantigens. To identify effective siRNA against Gal-3, we visually examined the sequence of Gal-3 mRNA and selected 3 targeting sites. The silencing potency of the designed siRNA was tested in freshly isolated human monocytes (Fig. 4A). All the 3 siRNA inhibited Gal-3 expression with siRNA-3 being the most effective. At a concentration of 2 μg, the silencing efficiency was around 99% when compared to control cells.

Figure 4.

 Selection of Gal-3 siRNA and the effects of Gal-3 gene silencing on MSC function. (A) Human monocytes were transfected with various siRNA concentrations using the BTX electroporation method [19]. Subsequent to transfection, the cells were incubated at 37 °C for 18 h and then the expression of Gal-3 was assessed by Western blots using anti-Gal-3 monoclonal antibody (Santa Gruz). Control cells were transfected with irrelevant siRNA targeting β galactosidase. (B) Downregulation of IDO gene expression by siRNA in MSC. Around 3 × 106 MSC cells were transfected with 2 μg siRNA-3 using BTX electroporation method and then processed as in A. (C) VEGF and Gal-3 levels in culture supernatants. (D) Representative examples of mixed lymphocyte reactions in the presence of MSC. Allogeneic irradiated mesenchymal stem cells (MSC) from one donor were added to mixed lymphocyte reactions (MLR). The MLR assay consists of mixing PBMC1 and PBMC2 from two different donors (75 × 104 each), which exhibited a good response in the initial testing. The cells were cultured in 96-well U-bottom plates and harvested on day 6. Cell proliferation was measured by [3H]-thymidine incorporation. The proliferative response was inhibited when the cells were co-cultured with a third-party allogeneic MSC (3 × 104 cells). Gene silencing of Gal-3 reduced the MSC inhibitory effect on MLR. The results represent the means ± SD of triplicate wells.

Having demonstrated that siRNA-3 is effective in human monocytes, next we assessed its silencing potency in MSC (Fig. 4B and C). The designed siRNA resulted in nearly 94% (±3%) reduction in intracellular protein levels, and around 95% (±4%) reduction in the secreted protein when compared to cells transfected with control siRNA. In contrast, depletion of Gal-3 has no significant effect on either β actin or VEGF expression, thus confirming the specificity of the designed siRNA-3.

Inhibition of Gal-3 expression reduced MSC inhibitory function on T-cell proliferation

To uncover the potential effects of Gal-3 knockdown on MSC function, we asked whether MSC-expressing Gal-3 could have an effect on the proliferation of lymphocytes in response to alloantigens. To this end, we first determined the cell concentration that gave a significant inhibition and found that suppression can be achieved after the addition of approximately 10–50 000 MSC to mixed lymphocyte cultures. Second, we tested lymphocyte response in the presence of 30 000 allogeneic MSC that have been transfected with either siRNA-3 against Gal-3 or control siRNA. In these experiments, peripheral blood mononuclear cells from donor 1 (PBMC1) were incubated with PBMC from a responder donor 2 (PBMC2) in the absence or presence of irradiated “third-party” MSC. In contrast to Gal-3 expressing MSC, knockdown of Gal-3 resulted in less immunosuppressive effect on T-cell proliferation (Fig. 4D, < 0.05, as a representative example).

Discussion

In addition to the expression of certain TLR, this study shows that MSC also express NOD-1. Unlike TLR, NLR consist of soluble proteins that survey the cytoplasm for signs that advertise the presence of intracellular invaders [15]. By screening the expression profiles in response to NOD-1 and TLR-2 synthetic ligands, we have identified a set of genes that were altered subsequent to overnight activation of MSC. More significantly, we have found that Gal-3 can modulate MSC immunosuppressive activity on T-lymphocyte proliferation, a new finding that merits some attention.

Amongst the upregulated genes, the p62 (also known as sequestosome 1) (SQSTM1) is an adaptor protein that has a role in inflammation, neurogenesis, osteoclastogeneis, adipogenesis and T-cell differentiation [21]. Our data indicated that p62 is induced by TLR-2 and NOD-1 activation at both mRNA and protein levels. Elucidating the pathways that control p62 levels in MSC will add another layer of detail to our understanding of the cell differentiation cascades in which p62 is involved. In addition to p62, VEGF and CXCL-10 were upregulated in response to NOD-1 and TLR-2 signalling. Human MSC released VEGF in response to TLR-2 and NOD-1 ligands as a potentially beneficial paracrine response. It will be interesting to investigate which mechanisms are involved in VEGF upregulation and secretion in MSC. Notably, previous studies have suggested a direct contribution of MSC to the blood vessel formation, as differentiation of MSC into endothelial cells has been demonstrated [22, 23].

In contrast to NOD-1, TLR-2 signalling also upregulated the expression of several important genes such as interleukin-1 receptor-associated kinase 2 (IRAK-2), involved in TLR signalling, NOTCH-1 and Gal-3 involved in innate and adaptive immunity. Notably, Notch pathway is highly conserved in evolution and is generally involved in cell fate decisions during cell differentiation [24]. A recent study showed that the inhibition of Notch signalling in MSC can hinder their suppressive activity on T-cell proliferation [13]. In addition to binding to glycan structures that are expressed by host cells, galectins can also recognize β-galactoside carbohydrates that are common structures on many pathogens [25], and therefore they are considered as a soluble pathogen recognition receptor. Within the immune system, galectins are expressed by virtually all immune cells, either constitutively or in an inducible fashion [17]. Also, they can be expressed by a spectrum of normal and tumour cells. As found in this study, Gal-3 is constitutively expressed by MSC and upregulated in response to TLR-2 ligation. Of note, high levels of Gal-3 protein are found in MSC culture supernatants; thus, it may participate in extra cellular matrix (ECM)-cell interactions and modulation of surrounding immune cells. Results from knockdown experiments showed that the immunosuppressive effects of MSC on T cells was lower than that from cells expressing Gal-3, suggesting a possible involvement of Gal-3 in MSC immunosuppressive function. This observation would fit with the demonstrated inhibitory effect of Gal-3 on T-cell proliferation [19, 20]. Also, a more recent study showed that tumour-associated Gal-3 contributes to tumour immune escapes by inhibiting the function of tumour-reactive T cells [26].

Some studies demonstrated that the MSC immunoregulatory properties are at least in part mediated by the production of cytokines, such TGF-β and hepatocyte growth factors [27]. However, other studies implicated prostaglandin E2 and interleukin 6 in MSC-mediated inhibition [28]. The identification of genes that regulate MSC inhibitory function will increase our understanding of the immunosuppressive properties of MSC and their therapeutic applications in the field of solid organ transplant and/or graft-versus-host disease (GVHD), a major complication of hematopoietic stem cell transplantation. Further studies of galectin expression and secretion by MSC under diverse culture conditions and differentiation pathways may reveal new immunological functions of these molecules.

Acknowledgment

This work was supported by in part by grants from the Norwegian Cancer Society and the gene therapy programme at the Norwegian Radium Hospital to Mouldy Sioud. We thank Lina Cekaite for performing the microarray screening experiments, Tommy Karlsen for providing some MSC and Anne Dybwad for reading the manuscript. The authors declare no conflict of interest.

Ancillary