Human Graft-Derived Mesenchymal Stromal Cells Potently Suppress Alloreactive T-Cell Responses

Abstract

After organ transplantation, recipient T cells contribute to graft rejection. Mesenchymal stromal cells from the bone marrow (BM-MSCs) are known to suppress allogeneic T-cell responses, suggesting a possible clinical application of MSCs in organ transplantation. Human liver grafts harbor resident populations of MSCs (L-MSCs). We aimed to determine the immunosuppressive effects of these graft-derived MSCs on allogeneic T-cell responses and to compare these with the effects of BM-MSCs. BM-MSCs were harvested from aspirates and L-MSCs from liver graft perfusates. We cultured them for 21 days and compared their suppressive effects with the effects of BM-MSCs on allogeneic T-cell responses. Proliferation, cytotoxic degranulation, and interferon-gamma production of alloreactive T cells were more potently suppressed by L-MSCs than BM-MSCs. Suppression was mediated by both cell–cell contact and secreted factors. In addition, L-MSCs showed ex vivo a higher expression of PD-L1 than BM-MSCs, which was associated with inhibition of T-cell proliferation and cytotoxic degranulation in vitro. Blocking PD-L1 partly abrogated the inhibition of cytotoxic degranulation by L-MSCs. In addition, blocking indoleamine 2,3-dioxygenase partly abrogated the inhibitive effects of L-MSCs, but not BM-MSCs, on T-cell proliferation. In conclusion, liver graft-derived MSC suppression of allogeneic T-cell responses is stronger than BM-MSCs, which may be related to in situ priming and mobilization from the graft. These graft-derived MSCs may therefore be relevant in transplantation by promoting allohyporesponsiveness.

Introduction

Mesenchymal stromal cells (MSCs) are rare nonhematopoietic cells, and reside in the bone marrow (BM) cavity. They are characterized by their ability to produce colony-forming unit-fibroblast, support of the hematopoietic microenvironment, promotion of bone formation, and adherence to plastic in vitro [1 –3]. Besides their presence in the BM, MSC-like cells are present throughout the body, occupying a perivascular location [4,5], and are critically involved in maintaining tissue homeostasis through antiapoptotic and tissue-supporting properties [6 –8].

It has been demonstrated that resting or naive MSCs are inherently capable of suppressing T-cell responses [9 –11]. Bone marrow MSCs (BM-MSCs) can suppress recipient alloreactive T-cell responses and thereby prevent graft rejection [12], suggesting a possible clinical application of MSCs in organ transplantation. Human liver grafts also contain MSCs (L-MSCs) [13,14] and these are mobilized from the liver graft during the transplantation procedure. MSCs express several cell surface receptors that enable them to sense the microenvironment and alter their phenotype accordingly [6,15]. Many studies have shown that it is the nature of the environmental cues that dictates the plasticity and—in the end—immunosuppressive capacity of MSCs (14–27). Most of the data that support this concept of MSC mobilization and immune priming are based on experimental mouse studies. However, because humans and mice differ in immunomodulatory pathways used by MSCs, there is an urgent need to assess the immunosuppressive function of human MSCs in the clinical setting of liver transplantation [16].

In addition, it is still unknown whether human graft-derived MSCs are able to suppress alloreactive T-cell responses as potently as their BM counterparts. Although the liver is an immunologically tolerogenic organ [17 –19], it is unknown whether MSCs in the healthy liver contribute to this tolerogenicity. The organ donation process is associated with a wide range of hemodynamic and inflammatory changes throughout the body [20,21] that may well affect the inherent immunosuppressive properties of MSCs in the liver graft. Inflammation affects pathways that can inhibit T-cell responses, such as the PD-L1/PD-1 pathway. PD-1 is an inhibitory receptor expressed on T cells, which inhibits T-cell responses after interaction with its ligand PD-L1 [22,23], which is expressed on antigen-presenting cells and MSCs. PD-L1 is known to be upregulated after exposure to inflammatory cytokines, such as interferon-gamma (IFN-γ) [24]. Another factor that is upregulated in MSCs in response to IFN-γ is indoleamine 2,3-dioxygenase (IDO) [25,26], which has been associated with the immunosuppressive capacity of MSCs. Therefore, these mechanisms may well contribute to the inhibition of alloreactive T cells by MSCs after liver transplantation. In addition, soluble factors secreted by MSCs, such as transforming growth factor-beta, hepatocyte growth factor [27], and prostaglandin E₂ (PGE₂) [9] may contribute to their capacity to inhibit alloreactive T-cell responses as well.

In the present study, we aimed to assess whether graft-derived MSCs can suppress alloreactive T-cell responses and, if so, whether they suppress more potently than their BM counterparts. To assess this, we isolated L-MSCs from liver perfusates obtained from donor livers and added these cells to mixed lymphocyte reactions (MLRs). We measured proliferation and effector function of alloreactive T cells in these cocultures in the presence or absence of L-MSCs. These outcomes were compared with those of cocultures with BM-MSCs.

Materials and Methods

Ethics statement

Liver perfusates and splenocytes were obtained from deceased donors following organ donation. The use of all human materials was approved by the Medical Ethical Committee of the Erasmus MC-University Medical Center, Rotterdam, and the study was performed in accordance with the amended Declaration of Helsinki.

Isolation and culture of MSCs

BM-MSCs were harvested from aspirated BM as previously described [28]. L-MSCs were isolated from perfusates of human liver grafts. Liver perfusates were collected from liver graft donors during the organ donation procedure by flushing the transplant liver with 1 L of University of Wisconsin preservation solution, followed by a second flush with 400 mL human albumin during the back table bench procedure, before implantation, as previously described [14]. Liver perfusates were collected and centrifuged (1500 rpm, 4°C; 10 min) to pellet cells. The cell pellet was resuspended in phosphate-buffered saline (PBS) and mononuclear cells (MNCs) were isolated using Ficoll-Hypaque density gradient separation. MNCs were counted and resuspended in culture medium consisting of alpha-minimum essential medium (MEM)/GLUTAMAX (Gibco), 2% fetal bovine serum (FBS), 1% antibiotic–antimycotic (Gibco) supplemented with 20 ng/mL of recombinant human epidermal growth factor (eBioscience) and 10 ng/mL of recombinant human fibroblast growth factor 2 (eBioscience), plated at 1.0×10⁵ cells/cm² in 10-cm dishes, and cultured at 37°C, 5% CO₂. After 3 days, the nonadherent fraction was removed and the medium was replaced by fresh medium. Cells were cultured until colonies became confluent and were harvested using a nonenzymatic solution (TrypLE; Gibco) and further expanded for later use. Early passage cells from BM were cultured and expanded in the same culture conditions as described for L-MSCs.

Immunophenotypic profile of MSCs in liver perfusates and BM-MSCs

To detect the presence of mobilized MSCs and hematopoietic stem/progenitor cells (HSCs) in liver perfusates from deceased donors, MNCs collected from the liver perfusates, as described above, were counted for viability and stained to detect and quantify MSCs. For HSCs, MNCs were stained with the following antibodies: lineage-FITC (BD Pharmingen), CD34-PE (BD Pharmingen), CD38-APC-Cy7 (eBioscience), CD45RA-PB (eBioscience), and CD90-APC (eBioscience). For MSCs, MNCs were stained with the following antibodies: CD45-PB (eBioscience), CD146-PE (eBioscience), CD44-APC-Cy7 (eBioscience), CD73-APC (eBioscience), CD105-PE (eBioscience), CD90-FITC (eBioscience), CD14-PB (eBioscience), CD19-APC-Cy7 (eBioscience), HLA-ABC-FITC (eBioscience), and HLA-DR-APC (eBioscience). The same antibodies were used to stain BM-MSCs to compare the immunophenotypic profile of L-MSCs and BM-MSCs. The immunophenotypic profile of L-MSCs and BM-MSCs was carried out on early passage (P3) cells. Before addition of antibodies, cells were treated with Fc block (Miltenyi). Cells were stained at 4°C in the dark for 30 min. Afterward, cells were washed, centrifuged, and resuspended in PBS before flow cytometric analysis. All cells were incubated with 7-aminoactinomycin D (7-AAD) (eBioscience) to discriminate dead/dying cells and debris. Samples were collected, and a minimum of 1 million cells were measured on a BD FACS Canto II and analyzed using FlowJo software (Tree Star). In the analyses, gates were based on fluorescence-minus-one controls.

Differentiation assays

Early passage (P3) L-MSCs were placed in defined culture conditions and compared with early passage (P3) BM-MSCs. For osteogenic differentiation, cells were plated at 4,000 cells/cm² per well in a six-well dish (Corning) in regular culture medium as described above and placed in a humidified chamber with 5% CO₂ at 37°C. After 24 h, the medium was changed to osteogenic induction medium [alpha-MEM/GLUTAMAX, 10% FBS, 1% penicillin–streptomycin, 50 μM ascorbic acid (Sigma), 10 mM β-glycerophosphate (Sigma), 100 nM dexamethasone (Calbiochem)]. To induce differentiation, half of the medium was replaced by fresh medium every 3 days for a total of 21 days. After 21 days, cell cultures were fixed with 4% paraformaldehyde (Sigma) for 15 min, washed twice with PBS, and stained with 40 mM Alizarin Red S stain at pH=4.1 (Sigma) for 30 min to detect mineralization.

For adipogenic induction, 1×10⁵ cells were plated per well in a six-well dish in regular culture medium and placed in a humidified chamber with 5% CO₂ at 37°C. After 24 h, the wells were washed with PBS and fresh adipogenic maintenance medium [Dulbecco modified Eagle medium (DMEM)/low glucose (Gibco), 10 μg/mL human insulin (Invitrogen), 10% FBS, and 1% penicillin/streptomycin] was added. After 3 days, the medium was changed to adipogenic induction medium [DMEM/low glucose, 10 μg/mL human insulin, 100 μM indomethacin (Sigma), 0.5 mM 3-isobutyl-l-methylxanthine (Sigma), and 1 μM dexamethasone]. After 3 days, the medium was changed back to adipogenic maintenance medium. After an additional two rounds of maintenance–induction media changes, cells were incubated for an additional 7 days in adipogenic maintenance medium. After this, cells were fixed with 4% paraformaldehyde and stained with Oil Red O (Sigma) to detect formation of lipid droplets.

Immune priming of MSCs with IFN-γ plus tumor necrosis factor alpha

Passage 3 culture-expanded L-MSCs and BM-MSCs were plated at 1.0×10⁵ cells per well in a six-well dish in regular culture medium. Cells were grown for 48 h and afterward were growth arrested in serum-free media (RPMI 1640; Gibco) for 24 h. Following this, cells were treated with 10 ng/mL of recombinant human tumor necrosis factor alpha (TNF-α) and 10 ng/mL of recombinant human IFN-γ. After 24 h, the cell media were collected, frozen, and stored at −80°C until further use. The cells were stained with the following primary fluorescently conjugated antibodies: PD-L1-APC (eBioscience), CD73-FITC (eBioscience), CD45-PB (eBioscience), and HLA-DR-PE (eBioscience) in the presence of Fc blocking reagent (Miltenyi) at 4°C in the dark for 30 min. Cells were analyzed using a BD FACS Canto II Flow cytometer (BD Biosciences); 7-AAD was included to discriminate dead/dying cells and debris. A minimum of 10,000 events were collected and analyzed using FlowJo software (Tree Star).

MLRs in the presence or absence of MSCs

To test the suppressive capacity of MSCs on allogeneic T-cell responses, we performed MLRs, in which we stimulated carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled peripheral blood mononuclear cells (PBMCs) with allogeneic CD40 ligand-stimulated (CD40)-B cells in the presence or absence of L-MSCs or BM-MSCs. PBMCs were isolated from the blood of healthy blood bank donors using Ficoll-Hypaque density gradient centrifugation and cryopreserved at −135°C until further use. CD40-B cells, which were used as allogeneic stimulator cells, were expanded from organ donor splenocytes, as previously described [29]. For different experiments, different responder and stimulator combinations were used. Both PBMCs and CD40-B cells were thawed and recuperated overnight in B-cell medium [Iscove's modified Dulbecco's medium+10% human serum+1% penicillin/streptomycin (Gibco)+1% insulin/transferrin/selenium (Gibco)] at 37°C and 5% CO₂. PBMCs were labeled with 0.5 μM CFSE (Invitrogen) according to the manufacturer's instructions, and CFSE-labeled PBMCs (1×10⁵) were stimulated with 2×10⁵ irradiated (30 Gy) donor CD40-B cells in 96-well U-bottom plates in a final volume of 250 μL B-cell medium. In separate wells, 1×10⁴ or 2×10⁴ irradiated (30 Gy) liver graft-derived L-MSCs or BM-MSCs were added to the cocultures. PBMCs stimulated with 5 μg/mL phytohemagglutinin (Murex) were included as positive controls to assess their proliferative capacity. Each culture condition was performed in duplicate.

In separate experiments, we studied the roles of PD-L1, IDO, and PGE₂ in inhibition of allogeneic T-cell responses by MSCs by adding anti-PD-L1 mAb (10 μg/mL; eBioscience) to block PD-L1; 1-methyl-dl-tryptophan (1MT; 250 μM; Sigma-Aldrich) to block IDO; and EP1-3 receptor blockers (10 μM; AH6809; ITK Diagnostics) and EP4 receptor blocker (20 μM; AH23848; Sigma-Aldrich) to block PGE₂. In addition, to test the role of PGE₂, L-MSCs and BM-MSCs were pretreated overnight with 5 μM of indomethacin (I7378; Sigma Aldrich). Indomethacin (5 μM) was also present during the 5 days of coculturing.

After 5 days of culture at 37°C and 5% CO₂, cell-free supernatant was collected, frozen, and stored at −20°C for later analysis, and cells were stained for cell viability, using the LIVE/DEAD^® Fixable Dead Cell Stain Kit (Invitrogen), according to the manufacturer's protocol. Cells were then stained with anti-CD3-PerCP-Cy5.5 (BD Biosciences), anti-CD4-APC-H7 (BD Biosciences), anti-CD8-efluor450 (eBioscience) to distinguish T-cell subsets, and anti-CD19-horizonV500 (BD Biosciences) to exclude B cells from analysis. Cytotoxic degranulation was detected by addition of CD107a-APC (eBioscience) during the last 15 h of the cocultures. Cells were analyzed for proliferation, using CFSE dilution patterns, and for the phenotype on a BD FACS Canto II Flow cytometer (BD Biosciences). For analysis of phenotypic markers, we used FACSDiva software (Becton Dickinson) and precursor frequencies (PFs) of alloreactive CD4⁺, and CD8⁺ T cells were calculated using ModFit LT^® software (Verity Software House), as previously described [29]. From duplicate assays, average PFs were calculated.

IFN-γ production was measured in the culture supernatants by standard enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions (Human IFNγ CytoSet^™; Invitrogen). TNF-α production was also measured in the culture supernatants by ELISA (human TNF-α Ready-Set-Go!; eBioscience).

Polymerase chain reaction

Early passage (P3) liver perfusate-derived and BM-MSCs were plated at 1.0×10⁵ cells per well in a six-well dish (Corning) in full growth medium. After 24 h, cells were growth arrested in serum-free media (RPMI) for 24 h. Following this, 700 μL of Qiazol lysis buffer (Qiagen) was added to each well and cells were harvested. RNA was isolated using a Qiagen miRNeasy mini kit (Qiagen) and quantified using a Nanodrop ND-1000. A total of 300 ng was used to make cDNA using an iScript cDNA synthesis kit from Bio-Rad (Bio-Rad Laboratories), and 15 ng cDNA was used per real-time quantitative polymerase chain reaction (RT-qPCR). The expression of prostaglandin-endoperoxide synthase 1 and 2 (PTGS1 and PTGS2) was quantified using RT-qPCR with a SensiMix Plus SYBR Kit (BioLine) according to the manufacturer's instructions. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a household gene for normalization of gene expression. RT-qPCR was performed using the following primers: PTGS1-forward CGCCAGTGAATCCCTGTTGTT, PTGS1-reverse AAGGTGGCATTGACAAACTCC; PTGS2-forward CTGGCGCTCAGCCATACAG, PTGS2-reverse CGCACTTATACTGGTCAAATCCC; GAPDH-forward: AAGGTCGGAGTCAACGGATTT, GAPDH-reverse: ACCAGAGTTAAAAGCAGCCCTG. The fold change in mRNA was determined using the ΔCt method.

Statistics

Data were analyzed with GraphPad Prism 5.0 software and expressed as mean±standard error of the means. To test whether data were normally distributed, the Kolmogorov–Smirnov test was used. Differences between groups were analyzed using the paired t-test (normally distributed paired data), Wilcoxon signed rank test (nonnormally distributed paired data), or Mann–Whitney test (unpaired data). One-sample t-test was performed to test differences between experimental conditions and control conditions consisting of one sample. P-values<0.05 were considered significant.

Results

Immunophenotypic and differentiation potential of MSC-like cells mobilized from liver grafts are similar to those of BM-MSCs

To explore whether MSCs mobilize from human liver grafts after tissue injury in vivo, we collected liver graft perfusates during flushing of liver grafts at the time of transplantation. These livers have been exposed to hemodynamic, ischemic, and inflammatory changes during the donation process. From these liver perfusates, we recovered a population of cells that adhered to plastic and were amenable to expansion, such as BM-MSCs (Fig. 1A, B). In addition, these graft-derived cells could well differentiate into adipocytes and osteocytes (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/scd). It is likely that these cells originate from the graft parenchyma and not from residual donor blood as MSCs could not be recovered from donor blood cells [14,30]. To confirm whether these graft-derived cells were of mesenchymal origin, we compared their phenotype with that of BM-MSCs. This analysis revealed that liver perfusate-derived cells lacked the hematopoietic markers, CD45, HLA-DR, CD14, and CD19 (representative flow cytometric plots in Fig. 1C), while they expressed the common mesenchymal markers, CD146, CD44, CD73, CD90, and CD105 (Fig. 1D). This phenotype was similar to that of BM-MSCs (Fig. 1C, D). When combined, these data suggest that MSC-like cells in perfusates represent genuine graft-derived liver MSCs.

FIG. 1.

Detection of mesenchymal stromal-like cells (MSCs) in the perfusates from human liver grafts. (A) Representative phase-contrast images of plastic adherent cells recovered from the mononuclear cell fraction of liver perfusates from liver grafts and from bone marrow (BM). (B) Expansion of plastic adherent cells obtained from the perfusates of liver grafts over the course of 21 days. P1–3: passage 1–3. (C) Representative flow cytometric plots demonstrating that culture-expanded plastic adherent cells recovered from liver perfusates lacked the hematopoietic markers, CD45, HLA-DR, CD14, and CD19. (D) Representative flow cytometric plots demonstrating that the immunophenotypic profile of culture-expanded plastic adherent cells recovered from liver perfusates was similar to that of BM-MSCs based on CD45^-CD146⁺CD44⁺CD73⁺CD90⁺. Gates are based on fluorescence minus one in control conditions.

L-MSCs suppress alloreactive T cells better than BM-MSCs

Previously, it has been shown that BM-MSCs can suppress proliferation of alloprimed T cells [9]. We compared the suppressive capacity of L-MSCs with that of BM-MSCs in MLRs, in which CFSE-labeled PBMCs were stimulated with allogeneic splenocyte-derived CD40-B cells as strong antigen-presenting cells, as described previously [29]. The level of T-cell proliferation was determined using CFSE dilution patterns, from which we calculated PFs of responding T cells, as shown in Fig. 2A. We found that both CD4⁺ and CD8⁺ T cells proliferated in response to stimulation with allogeneic CD40-B cells over a 5-day period. PFs in the control conditions varied from ∼2% to 10% because different responder and stimulator cell combinations were used for different experiments. To test the suppressive capacity of L-MSCs and BM-MSCs, we added L-MSCs or BM-MSCs to the cocultures at an MSC/PBMC ratio of 1:10 or 1:5. With addition of L-MSCs (n=12, different donors) at a ratio of 1:5, proliferation of alloreactive CD4⁺ T cells was nine-fold lower and of CD8⁺ T cells was 10-fold lower than in control culture without L-MSCs. This difference was significantly different (P<0.0001). The antiproliferative effect of BM-MSCs (n=6, different donors) was also evident, causing a four-fold lower proliferation of both CD4⁺ and CD8⁺ T cells than in control cultures without BM-MSCs (P<0.0001). Interestingly, L-MSCs suppressed CD8⁺ T-cell proliferation significantly better than BM-MSCs (Fig. 2B; P<0.05). This superior suppressive effect of L-MSCs was also observed for CD4⁺ T cells, although not statistically significantly (Fig. 2B, P=0.068). With the addition of MSCs at the MSC/PBMC ratio of 1:10, we found the same trend toward a stronger antiproliferative effect of L-MSCs than BM-MSCs, although the suppression was less strong and differences between L-MSCs and BM-MSCs were stronger at a ratio of 1:5. Apparently, there is a dose-dependent inhibitive effect of MSCs on T-cell proliferation. In addition, L-MSCs inhibited cytotoxic degranulation, as measured with surface expression of CD107a on proliferated T cells, of both alloreactive CD4⁺ and CD8⁺ T cells significantly better than BM-MSCs (P<0.01, Fig. 2C, D). Finally, L-MSCs reduced IFN-γ production significantly (P<0.01), while BM-MSCs did not. This difference between BM-MSCs and L-MSCs was statistically significant (Fig. 2E, P<0.05). TNF-α production was strongly inhibited by both L-MSCs and BM-MSCs with no significant difference between L-MSCs and BM-MSCs (data not shown).

FIG. 2.

Graft-derived MSC, from human liver grafts (L-MSCs), suppression of alloreactive T-cell responses is stronger than BM-MSCs. Carboxyfluorescein diacetate succinimidyl ester (CFSE)-labeled peripheral blood mononuclear cells (PBMCs) were cultured for 5 days with allogeneic splenocyte-derived CD40-ligand-stimulated-B cells (CD40-B cells). Mixed lymphocyte reactions (MLRs) were performed in the presence and absence of either L-MSCs (n=12) or BM-MSCs (n=6). CFSE dilution patterns of CD4⁺ and CD8⁺ T cells were analyzed and precursor frequencies (PFs) were calculated with ModFit^® software. (A) Representative ModFit-derived CFSE dilution patterns of CD4⁺ (left panel) and CD8⁺ (right panel) T cells after 5 days of stimulation with allogeneic CD40-B cells with and without MSCs. (B) Scatter dot plots showing PFs of alloreactive CD4⁺ and CD8⁺ T cells in the MLR without MSCs (control MLR, dotted line) and in the MLRs with L-MSCs versus BM-MSCs. Data are depicted as mean±standard error of the means (SEM). (C) In parallel, flow cytometric analysis of CD107a (y-axis) on CD4⁺ (upper panel) and CD8⁺ (lower panel) T cells in the MLRs was performed. Representative flow cytometric plots are shown from control MLR, MLRs with L-MSCs, and MLRs with BM-MSCs. (D) Scatter dot plots showing % proliferated CD107a⁺ CD4⁺ and CD8⁺ T cells in the control MLR (dotted line) and in the MLRs with L-MSCs versus BM-MSCs. Data are depicted as mean±SEM. (E) In parallel, samples of the same PBMCs were analyzed for interferon-gamma (IFN-γ) production upon stimulation with CD40-B cells. Five days after beginning the MLR, supernatants were harvested and IFN-γ levels were measured using enzyme-linked immunosorbent assay. Scatter dot plots showing IFN-γ production in MLR (dotted line) and in MLRs with L-MSCs versus BM-MSCs. Differences between suppressive effects of L-MSCs versus BM-MSCs were tested using Mann–Whitney test. *P<0.05, **P<0.01.

Soluble factors secreted by L-MSCs suppress CD4⁺, but not CD8⁺, T-cell proliferation, and CD4⁺ and CD8⁺ cytotoxic degranulation

To assess whether the inhibition of L-MSCs on T-cell responses was cell–cell contact independent, MLRs were performed in the presence of conditioned medium (CM) derived from MSC cultures. We found that the CM from L-MSCs suppressed the proliferation of alloreactive CD4⁺ T cells, while CM from BM-MSCs did not (Fig. 3A, P<0.001). CM from L-MSCs also suppressed proliferation of CD8⁺ T cells significantly (P<0.01), but this effect was not significantly different from the suppressive effects of CM from BM-MSCs (Fig. 3A). In addition, CM from L-MSCs suppressed the cytotoxic degranulation capacity of alloreactive CD4⁺ T cells and CD8⁺ T cells significantly better than CM from BM-MSCs (Fig. 3B, P<0.001 and P<0.01, respectively). CM from L-MSCs was not able to suppress IFN-γ production or TNF-α production (data not shown). Taken together, these data suggest that the immunomodulatory potential of L-MSCs is at least partially mediated by secreted factors.

FIG. 3.

Graft-derived L-MSC suppression of alloreactive T-cell responses is stronger by soluble factors than BM-MSCs. Supernatants from MSCs [conditioned medium (CM)] were added to the MLR, and at the end of 5 days, alloreactive T-cell responses were measured. (A) Scatter dot plots showing PFs of alloreactive CD4⁺ and CD8⁺ T cells in the MLR without CM (control MLR, dotted line) and in the MLRs with CM from L-MSCs versus CM from BM-MSCs. Data are depicted as mean±SEM. (B) Scatter dot plots showing % proliferated CD107a⁺ CD4⁺ and CD8⁺ T cells in the control MLR (dotted line) and in the MLRs with CM L-MSCs versus CM BM-MSCs. Data are depicted as mean±SEM. **P<0.01, ***P<0.001.

L-MSCs express higher levels of PD-L1 than BM-MSCs, which is associated with suppression of alloreactive T-cell responses

We wondered which mechanism could cause the superior immunosuppressive capacity of L-MSCs in comparison with their BM-MSC counterparts. An important coinhibitory pathway in (alloreactive) T-cell responses is the PD-1/PD-L1 pathway [31]. We found that L-MSCs have higher baseline expression levels of PD-L1 than BM-MSCs (Fig. 4A, C, P<0.05), which could be a reason for the higher suppressive capacity of L-MSCs. Proinflammatory cytokines have been described to upregulate PD-L1. Since proinflammatory cytokines increase during the organ donation process, it is possible that during donation, MSCs in the liver become primed and thereby upregulate PD-L1. To assess the effect of proinflammatory cytokine priming on PD-L1 expression and suppressive capacity, we exposed L-MSCs and BM-MSCs to TNF-α and IFN-γ ex vivo and measured their PD-L1 expression and suppressive function. After priming, we indeed found a strongly upregulated PD-L1 expression on both L-MSCs and BM-MSCs (Fig. 4B, C).

FIG. 4.

Graft-derived L-MSCs express higher levels of PD-L1 than BM-MSCs; PD-L1 is upregulated after ex vivo priming with proinflammatory cytokines and associated with suppression of alloreactive T-cell responses. (A) Representative flow cytometric plots demonstrating baseline expression of PD-L1 on L-MSCs and BM-MSCs. (B) Representative flow cytometric plots demonstrating induced expression of PD-L1 following ex vivo priming with proinflammatory cytokines, recombinant human-interferon-gamma (rh-IFN-γ)/tumor necrosis factor alpha (TNF-α), at 10 ng/mL for 24 h. (C) Scatter dot plot showing percentages of live CD45⁻ cells expressing PD-L1 at baseline or following stimulation with rh-IFN-γ/TNF-α, quantified using flow cytometry (L-MSCs, n=16; BM-MSCs, n=6). (D) Correlation between baseline PD-L1 expression on the MSCs added to the MLRs and PFs of CD4⁺ and CD8⁺ T cells in the MLRs, and correlation between PD-L1 and cytotoxic degranulation in the MLRs. Correlation was tested using Spearman's rank correlation coefficient. To test whether the suppressive effects of L-MSCs and BM-MSCs could be abrogated by blocking PD-L1, a PD-L1 blocking antibody was added to the MLRs. (E) Scatter dot plots showing % proliferated CD107a⁺ CD4⁺ and CD8⁺ T cells in the control MLR (dotted line); in the MLRs with L-MSCs versus L-MSCs+anti-PD-L1 and in the MLRs with BM-MSCs versus BM-MSCs+anti-PD-L1. Data are depicted as mean±SEM. Differences between MLRs with and without anti-PD-L1 were tested using paired t-test. *P<0.05, **P<0.01.

We wondered whether the superior suppressive effects of L-MSCs were mediated by their higher level of PD-L1 expression. We first looked at the correlation between PD-L1 expression on MSCs and the corresponding T-cell responses in the MLRs and found that a higher PD-L1 expression on MSCs was strongly correlated with lower proliferation of both CD4⁺ and CD8⁺ T cells in the MLRs. In addition, higher PD-L1 expression levels were also strongly correlated with lower cytotoxic degranulation capacities of both CD4⁺ and CD8⁺ T cells in the MLRs (Fig. 4D). We then added an anti-PD-L1 antibody to the MLRs to test whether blockade of the PD-1/PD-L1 interaction abrogated the suppressive effects of the L-MSCs. We found that anti-PD-L1 did not abrogate the suppressive effects of L-MSCs on proliferation and IFN-γ production or TNF-α production of T cells (data not shown), but significantly abrogated the effects on CD4⁺ and CD8⁺ T-cell cytotoxic degranulation (Fig. 4E, P<0.05 and P<0.01, respectively).

Immunosuppressive effects of MSCs partly involve IDO, but not PGE₂

In a previous study of our research group [14], we performed a gene expression profiling to compare the molecular phenotypes of L-MSCs and BM-MSCs. These data showed that L-MSCs and BM-MSCs were very similar, but a small number of genes were differentially expressed between L-MSCs and BM-MSCs. In the present study, we focused on the genes that were differentially expressed to investigate whether these could provide insight into the differences in immunomodulatory effects of L-MSCs and BM-MSCs. One of the genes that were significantly differentially expressed was PTGS1, which encodes for cyclooxygenase-1 (COX-1). COX-1 is the key enzyme in prostaglandin biosynthesis [32]. Prostaglandins are produced by MSCs and have previously been described to inhibit T-cell responses [9]. Since PTGS1 was expressed 20-fold higher in L-MSCs than in BM-MSCs, differences in prostaglandin production between the two types of MSCs could well explain the immunomodulatory differences that we found. We first confirmed the difference in PTGS1 gene expression between L-MSCs and BM-MSCs using PCR (Fig. 5A, P<0.001). We then investigated whether culturing MSCs in the presence of COX-1 inhibitor, indomethacin, affected the immunomodulatory function of L-MSCs or BM-MSCs in the cocultures with PBMCs. However, we found no effects of this (Fig. 5B). In addition, we determined whether blockers of the prostaglandin receptors, EP1,2,3, and 4, abrogated the inhibitive effects of CM from L-MSCs or BM-MSCs. However, we were not able to show higher rates of T-cell proliferation after addition of the blockers (Fig. 5C).

FIG. 5.

Effects of blocking prostaglandin E2 (PGE₂) on the inhibition of T-cell responses by L-MSCs and BM-MSCs. (A) Relative expression of prostaglandin-endoperoxide synthase 1 (PTGS1) in L-MSCs (n=10) versus BM-MSCs (n=5). (B) Scatter dot plots showing PFs of alloreactive CD4⁺ and CD8⁺ T cells in the MLR without MSCs (control MLR, dotted line) and in the presence and absence of L-MSCs (n=3) versus BM-MSCs (n=2) pretreated with and in the presence of cyclooxygenase-1 (COX-1) inhibitor, indomethacin. Data are depicted as mean±SEM. (C) Scatter dot plots showing PFs of alloreactive CD4⁺ and CD8⁺ T cells in MLRs with CM from L-MSCs (n=6) versus CM from BM-MSCs (n=3) in the presence and absence of blockers of PGE₂ receptors, EP1–4. Data are depicted as mean±SEM. ***P<0.001.

Another suppressive mechanism described for MSCs is the production and activity of IDO [25,26]. After addition of 1MT to block IDO, we found significant abrogation of the immunosuppressive effects of L-MSCs, but not BM-MSCs, on T-cell proliferation (Fig. 6, P<0.001). No effect on cytotoxic degranulation was observed (data not shown). In conclusion, L-MSCs inhibit CD4⁺ and CD8⁺ T-cell degranulation partly by PD-L1/PD-1 interaction and inhibit T-cell proliferation partly by IDO.

FIG. 6.

Effects of blocking indoleamine 2,3-dioxygenase (IDO) on the inhibition of T-cell responses by L-MSCs and BM-MSCs. Scatter dot plots showing PFs of alloreactive CD4⁺ and CD8⁺ T cells in the MLR without MSCs (control MLR, dotted line) and in the presence of L-MSCs (n=12) versus BM-MSCs (n=6) in the presence and absence of IDO blocker, 1-methyl-dl-tryptophan (1MT). Data are depicted as mean±SEM. Differences between MLRs with and without 1MT were tested using Wilcoxon signed rank test. ***P<0.001.

Discussion

In this study, we showed that graft-derived L-MSCs, which are released during transplantation, suppress proliferation, cytotoxic degranulation, and IFN-γ production of alloreactive T cells. This suppressive capacity of L-MSCs was significantly better than that of BM-MSCs even though the phenotype and differentiation potential of L-MSCs and BM-MSCs were very similar. However, L-MSCs expressed higher levels of PD-L1 than BM-MSCs. PD-L1 is the ligand for the coinhibitory receptor, PD-1, on T cells. On L-MSCs, expression of PD-L1 was associated with inhibition of T-cell proliferation and cytotoxic degranulation in vitro. There was a strong correlation between levels of PD-L1 expression on MSCs and corresponding proliferation and cytotoxic degranulation of T cells in the cocultures. The upregulation of PD-L1 on graft-derived MSCs may be a consequence of in vivo priming by proinflammatory cytokines, which are elevated during the donation process [33,34].

Blocking of PD-L1 in vitro partly abrogated the suppressive effects of L-MSCs on CD4⁺ and CD8⁺ T-cell cytotoxic degranulation. However, blocking of PD-L1 did not abrogate the effects on proliferation and IFN-γ and TNF-α production. Apparently, the PD-L1 expression represents a surrogate marker of inhibition, but the higher PD-L1 expression on L-MSCs only partly explains that suppressive effects of L-MSCs on T-cell responses were superior to those of BM-MSCs. It is likely that other mechanisms are involved in this.

We showed that, besides cell–cell contact-dependent mechanisms, soluble factors produced by L-MSCs contribute to the superior suppression of alloreactive T-cell responses as well. Since soluble factors have been described to contribute to the immunosuppressive properties of MSCs [35], we further investigated the suppressive function of secreted factors of MSCs present in conditioned culture media on alloreactive T-cell responses. Our finding that CM from L-MSCs suppressed T-cell proliferation, but not T-cell degranulation and IFN-γ and TNF-α production, whereas CM from BM-MSCs exerted no suppressive effects suggests that L-MSCs produce T-cell-suppressive soluble factors that BM-MSCs do not produce. These soluble factors released by graft-derived L-MSCs may directly affect T cells or indirectly through their effect on antigen-presenting cells [12,36]. Soluble factors secreted by L-MSCs could inhibit the stimulator CD40-B cells in the cocultures, which thereby cause inhibition of the T-cell responses.

An interesting finding was that L-MSCs showed higher expression of PTGS1, a key enzyme involved in prostaglandin production, than BM-MSCs. Although prostaglandins are produced by MSCs and are able to inhibit T-cell responses [9], we did not find abrogative effects of blocking prostaglandins either by indomethacin or by prostaglandin receptor inhibitors. Clearly, these results show no role for PGE2 in the inhibition by MSCs.

More importantly, we found that blocking IDO partly restored T-cell proliferation that was inhibited by L-MSCs, while this had no effect on the conditions with BM-MSCs. IDO has previously been shown to inhibit T-cell responses by MSCs [25,26,37]. Our finding that blocking IDO only abrogated the inhibition by L-MSCs and not BM-MSCs supports the hypothesis that the immunomodulatory function of MSCs is highly plastic, as published previously [38], since L-MSCs have been exposed to an inflammatory environment.

One of the limitations of comparative studies of MSC populations is the lack of consensus to immunophenotypically define these cells [6]. The current study compared MSC populations from liver grafts and BM. MSCs found in liver graft perfusates adhered to plastic, were amenable to expansion, and expressed cell surface markers that are described to identify MSCs, similar to BM-MSCs [6]. The superior suppressive capacity of L-MSCs may be related to the tissue source. The phenotype of MSCs in the liver may reflect the hyporesponsiveness of the hepatic immune system. Another difference is that L-MSCs have been exposed to extensive ischemia, whereas cells from BM were not exposed to substantial ischemia. Last, a difference between BM-MSCs and L-MSCs is their mobilization. L-MSCs were mobilized and released from the liver parenchyma during the donation process, whereas BM-MSCs were obtained from the stroma without mobilization and release. Which of these differences is most likely to attribute to the T-cell-suppressive capacity of MSCs needs further evaluation.

For future clinical applications, our study provides insights in the superior immunosuppressive capacity of L-MSCs, which may lead to cell-based immunosuppressive therapies using liver graft-derived MSCs. For this therapeutic application, L-MSCs need to be isolated from the donor liver and expanded in vitro, as described previously [39], since L-MSCs are present in low numbers. After expansion, L-MSCs should be infused in patients and may have an important function in the setting after liver transplantation [39] or may be used as a bridging therapy before transplantation in patients with acute liver failure [40,41].

In conclusion, we showed that MSCs mobilized from livers grafts strongly suppress alloreactive T-cell responses. This suppression was stronger than nonmobilized and nonprimed MSCs derived from BM stroma. These differences may not only be related to the intragraft immune priming of L-MSCs in the organ donor or due to mobilization of MSCs from the graft but may also be related to the inherent immunological tolerogenicity of the liver. In the setting of liver transplantation, resident and/or mobilized L-MSCs may therefore regulate immune-mediated liver damage by suppressing pathogenic alloreactive T-cell responses, leading to better acceptance of liver grafts.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Friedenstein

, Chailakhyan

, Latsinik

, Panasyuk

and Keiliss-Borok

. (1974). Stromal cells responsible for transferring the microenvironment of the hemopoietic tissues. Cloning in vitro and retransplantation in vivo. Transplantation, 17:331–340.

Friedenstein

, Ivanov-Smolenski

, Chajlakjan

, Gorskaya

, Kuralesova

, Latzinik

and Gerasimow

. (1978). Origin of bone marrow stromal mechanocytes in radiochimeras and heterotopic transplants. Exp Hematol, 6:440–444.

Pittenger

, Mackay

, Beck

, Jaiswal

, Douglas

, Mosca

, Moorman

, Simonetti

, Craig

and Marshak

. (1999). Multilineage potential of adult human mesenchymal stem cells. Science, 284:143–147.

Crisan

, Yap

, Casteilla

, Chen

, Corselli

, Park

, Andriolo

, Sun

, Zheng

, et al. (2008). A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell, 3:301–313.

Murray

, West

, Hardy

, James

, Park

, Nguyen

, Tawonsawatruk

, Lazzari

, Soo

and Peault

. (2013). Natural history of mesenchymal stem cells, from vessel walls to culture vessels. Cell Mol Life Sci, 71:1353–1374.

Keating

. (2012). Mesenchymal stromal cells: new directions. Cell Stem Cell, 10:709–716.

Tolar

, Le Blanc

, Keating

and Blazar

. (2010). Concise review: hitting the right spot with mesenchymal stromal cells. Stem Cells, 28:1446–1455.

Uccelli

, Moretta

and Pistoia

. (2008). Mesenchymal stem cells in health and disease. Nat Rev Immunol, 8:726–736.

Aggarwal

and Pittenger

. (2005). Human mesenchymal stem cells modulate allogeneic immune cell responses. Blood, 105:1815–1822.

10.

Le Blanc

, Rasmusson

, Sundberg

, Gotherstrom

, Hassan

, Uzunel

and Ringden

. (2004). Treatment of severe acute graft-versus-host disease with third party haploidentical mesenchymal stem cells. Lancet, 363:1439–1441.

11.

Nauta

and Fibbe

. (2007). Immunomodulatory properties of mesenchymal stromal cells. Blood, 110:3499–3506.

12.

De Miguel

, Fuentes-Julian

, Blazquez-Martinez

, Pascual

, Aller

, Arias

and Arnalich-Montiel

. (2012). Immunosuppressive properties of mesenchymal stem cells: advances and applications. Curr Mol Med, 12:574–591.

13.

da Silva Meirelles

, Chagastelles

and Nardi

. (2006). Mesenchymal stem cells reside in virtually all post-natal organs and tissues. J Cell Sci, 119(Pt 11):2204–2213.

14.

Pan

, Fouraschen

, Kaya

, Verstegen

, Pescatori

, Stubbs

, van Ijcken

, van der Sloot

, Smits

, et al. (2011). Mobilization of hepatic mesenchymal stem cells from human liver grafts. Liver Transpl, 17:596–609.

15.

Bernardo

and Fibbe

. (2013). Mesenchymal stromal cells: sensors and switchers of inflammation. Cell Stem Cell, 13:392–402.

16.

Kaipe

, Erkers

, Sadeghi

and Ringden

. (2014). Stromal cells-are they really useful for GVHD?. Bone Marrow Transplant, 49:737–743.

17.

Calne

, Sells

, Pena

, Davis

, Millard

, Herbertson

, Binns

and Davies

. (1969). Induction of immunological tolerance by porcine liver allografts. Nature, 223:472–476.

18.

Diehl

, Schurich

, Grochtmann

, Hegenbarth

, Chen

and Knolle

. (2008). Tolerogenic maturation of liver sinusoidal endothelial cells promotes B7-homolog 1-dependent CD8+ T cell tolerance. Hepatology, 47:296–305.

19.

Thomson

and Knolle

. (2010). Antigen-presenting cell function in the tolerogenic liver environment. Nat Rev Immunol, 10:753–766.

20.

Clavien

, Harvey

and Strasberg

. (1992). Preservation and reperfusion injuries in liver allografts. An overview and synthesis of current studies. Transplantation, 53:957–978.

21.

Goode

, Webster

, Howdle

, Leek

, Lodge

, Sadek

and Walker

. (1994). Reperfusion injury, antioxidants and hemodynamics during orthotopic liver transplantation. Hepatology, 19:354–359.

22.

Ishida

, Agata

, Shibahara

and Honjo

. (1992). Induced expression of PD-1, a novel member of the immunoglobulin gene superfamily, upon programmed cell death. EMBO J, 11:3887–3895.

23.

Freeman

, Long

, Iwai

, Bourque

, Chernova

, Nishimura

, Fitz

, Malenkovich

, Okazaki

, et al. (2000). Engagement of the PD-1 immunoinhibitory receptor by a novel B7 family member leads to negative regulation of lymphocyte activation. J Exp Med, 192:1027–1034.

24.

LaGier

and Pober

. (2006). Immune accessory functions of human endothelial cells are modulated by overexpression of B7-H1 (PDL1). Hum Immunol, 67:568–578.

25.

Croitoru-Lamoury

, Lamoury

, Caristo

, Suzuki

, Walker

, Takikawa

, Taylor

and Brew

. (2011). Interferon-gamma regulates the proliferation and differentiation of mesenchymal stem cells via activation of indoleamine 2,3 dioxygenase (IDO). PLoS One, 6:e14698.

26.

Francois

, Romieu-Mourez

, Li

and Galipeau

. (2012). Human MSC suppression correlates with cytokine induction of indoleamine 2,3-dioxygenase and bystander M2 macrophage differentiation. Mol Ther, 20:187–195.

27.

Di Nicola

, Carlo-Stella

, Magni

, Milanesi

, Longoni

, Matteucci

, Grisanti

and Gianni

. (2002). Human bone marrow stromal cells suppress T-lymphocyte proliferation induced by cellular or nonspecific mitogenic stimuli. Blood, 99:3838–3843.

28.

Hall

, Jiang

, Leary

, Yavanian

, Eminli

, O'Neill

and Marasco

. (2013). Identification and isolation of small CD44-negative mesenchymal stem/progenitor cells from human bone marrow using elutriation and polychromatic flow cytometry. Stem Cells Transl Med, 2:567–578.

29.

Tapirdamaz

, Mancham

, van der Laan

, Kazemier

, Thielemans

, Metselaar

and Kwekkeboom

. (2010). Detailed kinetics of the direct allo-response in human liver transplant recipients: new insights from an optimized assay. PloS One, 5:e14452.

30.

Hoogduijn

, Verstegen

, Engela

, Korevaar

, Roemeling-van Rhijn

, Merino

, Franquesa

, de Jonge

, Ijzermans

, et al. (2014). No evidence for circulating mesenchymal stem cells in patients with organ injury. Stem Cells Dev, 23:2328–2335.

31.

Yang

, Riella

, Chock

, Liu

, Zhao

, Yuan

, Paterson

, Watanabe

, Vanguri

, et al. (2011). The novel costimulatory programmed death ligand 1/B7.1 pathway is functional in inhibiting alloimmune responses in vivo. J Immunol, 187:1113–1119.

32.

Bos

, Richel

, Ritsema

, Peppelenbosch

and Versteeg

. (2004). Prostanoids and prostanoid receptors in signal transduction. Int J Biochem Cell Biol, 36:1187–1205.

33.

Jassem

, Koo

, Cerundolo

, Rela

, Heaton

and Fuggle

. (2003). Leukocyte infiltration and inflammatory antigen expression in cadaveric and living-donor livers before transplant. Transplantation, 75:2001–2007.

34.

Jassem

, Koo

, Muiesan

, Cerundolo

, Rela

, Fuggle

and Heaton

. (2003). Non-heart-beating versus cadaveric and living-donor livers: differences in inflammatory markers before transplantation. Transplantation, 75:1386–1390.

35.

Yang

, Park

, Yoon

, Kim

, Hong

, Shin

, Nam

, Kim

and Park

. (2009). Soluble mediators from mesenchymal stem cells suppress T cell proliferation by inducing IL-10. Exp Mol Med, 41:315–324.

36.

Nauta

, Kruisselbrink

, Lurvink

, Willemze

and Fibbe

. (2006). Mesenchymal stem cells inhibit generation and function of both CD34+-derived and monocyte-derived dendritic cells. J Immunol, 177:2080–2087.

37.

Chinnadurai

, Copland

, Patel

and Galipeau

. (2014). IDO-independent suppression of T cell effector function by IFN-gamma-licensed human mesenchymal stromal cells. J Immunol, 192:1491–1501.

38.

Wang

, Chen

, Cao

and Shi

. (2014). Plasticity of mesenchymal stem cells in immunomodulation: pathological and therapeutic implications. Nat Immunol, 15:1009–1016.

39.

Le Blanc

and Ringden

. (2007). Immunomodulation by mesenchymal stem cells and clinical experience. J Intern Med, 262:509–525.

40.

Peng

, Xie

, Lin

, Liu

, Zhu

, Xie

, Zheng

and Gao

. (2011). Autologous bone marrow mesenchymal stem cell transplantation in liver failure patients caused by hepatitis B: short-term and long-term outcomes. Hepatology, 54:820–828.

41.

Shi

, Zhang

, Xu

, Lin

, Fu

, Zou

, Zhang

, Shi

, Chen

, et al. (2012). Human mesenchymal stem cell transfusion is safe and improves liver function in acute-on-chronic liver failure patients. Stem Cells Transl Med, 1:725–731.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.17 MB

Human Graft-Derived Mesenchymal Stromal Cells Potently Suppress Alloreactive T-Cell Responses

Abstract

Introduction

Materials and Methods

Ethics statement

Isolation and culture of MSCs

Immunophenotypic profile of MSCs in liver perfusates and BM-MSCs

Differentiation assays

Immune priming of MSCs with IFN-γ plus tumor necrosis factor alpha

MLRs in the presence or absence of MSCs

Polymerase chain reaction

Statistics

Results

Immunophenotypic and differentiation potential of MSC-like cells mobilized from liver grafts are similar to those of BM-MSCs

L-MSCs suppress alloreactive T cells better than BM-MSCs

Soluble factors secreted by L-MSCs suppress CD4+, but not CD8+, T-cell proliferation, and CD4+ and CD8+ cytotoxic degranulation

L-MSCs express higher levels of PD-L1 than BM-MSCs, which is associated with suppression of alloreactive T-cell responses

Immunosuppressive effects of MSCs partly involve IDO, but not PGE2

Discussion

Footnotes

Author Disclosure Statement

References

Supplementary Material

Soluble factors secreted by L-MSCs suppress CD4⁺, but not CD8⁺, T-cell proliferation, and CD4⁺ and CD8⁺ cytotoxic degranulation

Immunosuppressive effects of MSCs partly involve IDO, but not PGE₂